The aims of this project are to develop HeartCare, an Internet-based cardiac recovery resource, and evaluate its effects on nursing practice and patient outcomes. HeartCare is designed to help nurses with the in-hospital discharge planning and teaching responsibilities and post-hospital self-monitoring, motivation, and home management necessary for recovery among patients who have undergone coronary artery bypass graft surgery (CABG). HeartCare extends nursing practice by linking patients in their homes to computer communication pathways and information resources located world wide on the Internet. We will develop methods by which HeartCare can be personalized for each cardiac patient, and by which HeartCare's effectiveness can be tested, demonstrated, and replicated. We will examine the effects of computerized home care tools on the nursing care of persons following CABG and on selected patient outcomes.

It is timely to design and evaluate nursing informatics applications that consider patients as direct users of the technology and extend nurses' abilities to care for patients. Evidence exists that patients can and will use computer technologies, including computer networks, to obtain clinical services (Brennan, Moore & Smyth, 1995; Gustafson, et.al. 1994; Wald, et.al. 1995). Contemporary patient care informatics applications, however, exist in isolation from the general health care delivery system, and may not be accessible to the nurses who would find them useful. It is a key aim of this project to develop an informatics application that can be integrated into the nursing care of persons recovering from CABG surgery.

The project will involve 140 patients of a cardiac surgical service in a randomized field investigation designed to determine the effects of HeartCare on nursing and patient outcomes. During a patient's index hospitalization, nurses will use HeartCare to tailor home care resources for individual patients; patients will take a HeartCare computer home with them and use it to access special programs and information resources available locally on the computer or world-wide via the Internet. Outcomes will be addressed through the following questions:

(1) What are the nursing practice impacts of HeartCare?

(2) How does HeartCare contribute to patient outcomes?

a. How do patients use HeartCare?

b. Do patients recovering from CABG surgery who have access to HeartCare experience significantly (1) better physical function, (2) less symptom distress, (3) less psychological distress, and (4) better perceived family function, at 1 week, 1 month, 3 months and 6 months post discharge than those in a comparison group who receive an audio tape intervention and those in a control group who receive standard post-op care?

c. Do patients with access to HeartCare report significantly greater adherence to cardiac risk behavior modification at 3 and 6 months post discharge than those in the comparison group?

d. Do these effects of HeartCare remain after controlling for the typical post-CABG surgery trajectory, co-morbidities, age, social support, presence of a caregiver, post-surgical length of stay, pre-surgical cardiac status, and social service use?

These questions will be answered through a randomized clinical trial. Patients who undergo CABG surgery with standard post-operative nursing support will receive either the HeartCare intervention (n=50), an audiotaped homecare program (n=50), or Usual Care (n=40). The Usual Care group is included in the study monitor changes in patient outcomes that may occur during the study period related to changes in surgical techniques or medical practice, and to evaluate postoperative recovery unaffected by technology or data collection contacts with nurses.

An important aim of this project is to determine the effects on practice of introducing computerized tools into the post-surgical care of persons following CABG. Therefore, in addition to establishing the extent to which technologically based interventions aid in the cardiac recovery process, the project will examine the ability of nurses to integrate patient-focused computer tools into their clinical practice and the impact of these tools on their practice, through interviews and group assessment techniques. The current emphasis on early discharge and the movement of patient care from organized health service agencies to the home mandates the development of computer-based tools to augment services provided by nurses. While there will always be a role for nurses in helping to prepare patients for discharge and to manage their recovery, well designed and validated technologies can be of great assistance to nurses in providing home care support to patients.


The goal of the Program Announcement (PA 95-010) to which this proposal is addressed is improvement in the delivery of clinical nursing care through the use of information systems. We propose to capitalize on the Internet as a special type of information system with enormous potential to improve the delivery of nursing care to patients recovering from coronary artery bypass graft surgery. The innovative project proposed here integrates nursing informatics with work on cardiac recovery and the clinical practice of nursing.

Nursing Informatics

Nursing informatics is "a combination of computer science, information science, and nursing science designed to assist in the management and processing of nursing data, information, and knowledge to support the practice of nursing and delivery of nursing care" (Graves and Corcoran, 1989, p227). Contemporary research in nursing informatics has addressed "management and processing of nursing data" in terms of vocabulary adequacy (Grobe, 1991; Henry, Holzemer & Campbell, 1994; McCloskey & Bulechek, 1994) and the design of clinical decision support systems (Woolery & Grzymala-Busse, 1994; Petrucci et al., 1992; Zielstorff, 1995; Chang & Hirsch, 1994). Additional efforts have targeted the development and evaluation of information management systems "to support the practice of nursing" (Gassert, 1996; Zielstorff et al., 1993) and issues related to human-computer interface (Staggers, 1994). A focus on the application of information technologies directly in the "delivery of nursing care" to patients is slowly emerging. The HeartCare evaluation falls in this third category, and it builds on a growing body of knowledge regarding customizing computer technology for patient care.

Brennan and colleagues (Brennan, Moore, & Ripich, 1990; Brennan, Moore & Smyth, 1991, 1992, 1995; Brennan & Ripich, 1994) have investigated the application of computer networks to aid nurses in the challenges of providing home care. Like other investigations in nursing informatics (see, for example, Norris, Cuddigan, Font & Leak, 1989 and Henry et al., 1994), Brennan's work is based on the premise that nurses diagnose and treat human responses to health and illness and that nursing services can be extended through appropriate application of computer technology. However, Brennan's work stands apart from other investigations in nursing informatics in that it views the patient as the primary user of the informatics application. Brennan's work is similar to that of investigators in the areas of health promotion (Bosworth & Gustafson, 1991; Gustafson et al.,1994), patient education and self help (Clark & Ellis, 1982) , and patient participation in clinical practice (Sands, Safran, Slack & Bleich, 1993; Slack, Safran, Kowalsof, Pearce & Delbanco, 1993).

A key criticism of most informatics applications for direct patient use, including Brennan's initial work, is that these systems exist in parallel, but not integrated with, the health care delivery system. Patients self-select into informatics research and demonstration projects, and their clinical providers become aware of the informatics applications in a happenstance manner. The informatics applications, then, exist almost independent from the clinical care of the patient. (Beth Israel's Patient Record System [Wald et al, 1995]) stands as an exception to this; however, that work primarily targets direct entry of patient data into the clinical record and does not provide any in-home resources for self care.) For informatics applications to be effective in nursing practice, it is necessary to discover how to integrate these applications into the routine clinical care of persons with complex health problems.

The project proposed here represents a first step in using a nursing informatics application to facilitate the delivery of nursing care to a large and important clinical population, persons recovering from coronary artery bypass graft surgery. The HeartCare project can serve as a prototype of the application of hypermedia technologies to nursing practice challenges. It will enable us to examine the feasibility of introducing technological home care supports during the in-hospital care of persons following CABG surgery. Further, it will make effective use of existing Internet-based resources, allowing nurses to develop only those patient resources that are explicitly tied to local practice. It will provide evidence of the gains in outcomes likely through efficient use of technology in practice.

Cardiac Recovery

Each year 350,000 patients who survive CABG surgery undergo a recovery period associated with adverse psychological and physical functioning, lasting for as long as 6 months following surgery (Allen, Becker & Swank, 1990; Dracup, 1982). Although preparatory discharge teaching for home recovery is a standard part of care for CABG patients, the trend toward aggressively early discharge of CABG patients (4 to 5 days following surgery) and the rising number of women and elders undergoing CABG (Pashkow, 1993) have resulted in a great diversity of home recovery needs that must be planned for in an exceedingly short period of time. Recovery goals include symptom management, physical and psychological functional recovery, and life–style changes for reduction of cardiac risk. Current CABG discharge teaching efforts, which use standardized approaches to patients and allow little tailoring to specific home care goals for a particular patient, fall short of adequately supporting patients in their efforts to achieve these varied goals (Cosmoss, 1988; Steele & Ruzicki,1987). Thus there is a need for discharge planning programs that more efficiently match recovery resources with individual patient goals. HeartCare, the cardiac surgery care resource to be developed and tested in the proposed investigation, is designed to provide patients with tailored home care support to facilitate CABG recovery.

During the first week of home recovery, patients' recovery goals typically center on: (1) detection and prevention of life–threatening complications, including arrhythmias, congestive heart failure, bleeding, and pneumonia; 2) promotion of physical function, including symptom management, and resumption of personal care activities; and (3) reduction of psychological distress (Finkelmeier, 1995). Near the end of the first month of recovery' patients' goals include building activity tolerance and independence in performing ADL (Gortner&Jenkins, 1990).Recovery goals during the second and third months of recovery focus on: (1) return to work and household maintenance activities; (2) resumption of sexual and social/recreational activities; and (3) reduction of emotional distress, including depression, anxiety, frustration, boredom and hopelessness (Tack & Gillis,1990; Allen, Becker & Swank, 1990; Stanton, Jenkins, Savageau, & Thurer, 1984). Recovery goals during the fourth to sixth month after CABG are: (1) return to maximal activity level; (2) adherence to risk factor modification behaviors, such as regular exercise, heart–healthy diet, weight control, BP management, smoking cessation, and stress management; (3) reduction of emotional distress; and (4) family and life satisfaction (Oka, Froelicher, & Burke, 1995).

The first month of recovery from CABG surgery is particularly distressful for patients; they report anxiety, anger, depression, and mood swings (King, 1985; Moore, 1994; Gillis, 1983) as well as fatigue, sleep disturbances, changes in appetite, discomfort from incisions, shortness of breath and anginal pain (King & Parrinello, 1988; Dracup, 1982; Gillis, 1983; Moore, 1994). As a result of these symptoms, patients may curtail their activities, delaying their return to previous roles. The recovery period has been characterized by changes in the family's usual pattern of functioning (Stanley & Frantz, 1988; Gillis, 1983) and families may be at high risk for disruption following CABG (Gortner et al., 1988), Miller, Wikoff, McMahon, Garrett & Ringel, 1990). Dependence by the patient and changes in family roles affect marital adjustment and family functioning. HeartCare can potentially provide more peer support and professional monitoring than traditional approaches to recovery support. These neutral sources of monitoring (non-family members) may decrease the patient's experience of family conflicts.

Except for Gillis and associates (1993) and Gortner and colleagues (1988, 1990), few investigators have examined the effects of discharge teaching on functional outcomes after CABG. In a series of studies these researchers found that written homegoing instructions plus follow–up telephone calls enhanced patient physical activities but had no effect on psychological functioning. On the other hand, Beckie (1989) found that a home telephone intervention increased knowledge about recovery and coronary artery disease and risk factor modification and reduced anxiety during the first 8 weeks after discharge. However, no physical or social functional outcomes were addressed in this study. In each of these intervention studies, the authors noted the substantial use of nurses' time required for follow-up telephone calls. Methods of providing nursing support to patients recovering at home that require less direct nursing time, such as HeartCare, may be more cost effective.

Additionally, CABG discharge intervention studies have used subjects whose mean hospital length of stay was approximately 10 days, and findings from these studies may not be applicable to current CABG patients, whose average length of stay is 5 to 6 days. Interventions to aid recovery in the first week after discharge are just being formulated; they include both home visits and telephone monitoring (Whitman,1995). There have been no studies of the effectiveness of these interventions, however, and as noted above, home visits and telephone monitoring require intensive use of human resources and are expensive. Alternative, less human resource-intensive methods of home monitoring for CABG patients during the first week after discharge, such as HeartCare, are needed.

A large proportion of CABG discharge information focuses on long-term behavior changes associated with risk reduction. Several studies have shown that cardiac discharge teaching increases patient knowledge (Steele & Ruzicki, 1987; Marshal, Penckofer & Llewellyn, 1986), but studies are less conclusive about the ability of these educational programs to produce changes in patient risk-reduction behaviors (Penckofer & Llewellyn, 1989; Eyherabide & Yates, 1985). The timing of the interventions (at the point of discharge) may have been too early for patients to think about long term life-style changes given their immediate concerns about symptom management and resumption of usual daily activities. Formal cardiac rehabilitation programs (which occur later in recovery) are attended by only 15% of eligible persons and they traditionally focus on exercise for fitness training (Ades, Waldman, Polk & Coflesky, 1992); they have not been shown to be particularly effective for changing patient risk factor behaviors. More effective may be an educational approach to risk-factor modification that includes specific, sequenced tailoring of content for each individual, like HeartCare.

Nursing Practice in the Post-CABG Surgery Period

Following coronary artery bypass graft surgery nurses teach and motivate patients for home management. Current pressures for cost containment have led to a reduction in the number of staff nurses available discharge preparation at the very time when the patient's expected hospital stay has been cut almost in half. Thus, nurses must accomplish the goals of discharge preparation in an shorter time period with fewer peers to rely on. Traditionally nurses have relied on tools such as videotapes, family counseling sessions and group education to prepare post-surgical patients and families for discharge (Gortner et al., 1988; Ruzicki, 1989; Barbarowitz, Nelson, DeBusk & Haskell, 1980). Printed brochures, prepared by the clinical staff or by national organizations (American Heart Association, 1995), are important resources to help the patient transition to home, and to remind the patient once at home of the important teaching in the hospital. These resources have been shown to have some effect in improving patient compliance (Eyerbide & Yates, 1985; Penckofer & Llewellyn, 1989) but they often are boring and do not organize information in a manner that is likely timely for use or to be recalled.

Our intervention, HeartCare, is similar to paper based resources, in that it will provide content relevant to the post-cardiac recovery period. However, HeartCare offers significant advantages over paper-based resources in that it is enriched in presentation format, offers a greater variety of materials, and, most importantly, provides a way to organize information related to recovery from CABG surgery in a manner that can be tailored to individuals. Because the computer network will be available when the patient chooses to use it, HeartCare offers timely access to a richer and better organized set of information resources than could ever be gained by providing a set of brochures. Further, Internet-based resources can be more easily updated than brochures. It is timely and essential to systematically evaluate a resources that patients are already using to determine how to integrate it into practice.

Computers and Home-based Cardiac Care One of the first computerized home care interventions, established in 1958, provided in-home monitoring for persons recovering from cardiac surgery. In a demonstration project, the Massachusetts General Hospital Departments of Surgery and Cardiology set up a pulse monitoring system into which patients telephoned their pulse rates each morning. The system also provided oversight, initiating a call to those patients who had not called in by a certain time. Findings indicated that patients were capable of monitoring their own pulse rates and calling the report into a central office. Although this early study was not a randomized trial, it provides evidence that patients are able to comprehend the role of technology in recovery and the importance of technology in symptom management. HeartCare will offer an e-mail based monitoring system.

Recent applications of computers in cardiac home care have included risk factor assessment programs (Gustafson et al., 1987) and cardiac treatment choice programs (Kasper, Mulley, & Wennberg, 1992). All documented programs are free-standing systems; that is, they require a local computer on which to store and run special programs. Program updating requires transfer of new code to the local computer. These systems have their merits but they are limited by an inability to tailor content in an efficiently and by a lack of support for communication or automatic updating. In contrast, the Internet provides a vast array of information that, properly harnessed and indexed, could be of help to persons recovering from cardiac surgery.

The Internet is an electronic communication pathway for information resources across the world. Access to the Internet requires that one have an account; these are publicly available and cost about $30/month. The World Wide Web (WWW) is a newly developed strategy (ca 1992) for accessing and organizing the text, picture, video and sound on the Internet. WWW servers, specialized computers linked via the Internet, function as repositories for health related information; persons with access to the Internet and special local programs called Browsers can use the Web servers. Examples of successful, clinically relevant Web servers include Oncolink, providing cancer information and advice, and the Virtual Hospital of the University of Iowa (D'Allseandro et al, 1994). Patients can explore information in their homes at the pace they want, as often as they want and at the level of detail they desire.

The quantity and variety of helpful information offered through the Internet cannot be provided through traditional health care channels. Nor are resources available through traditional channels to offer patients the level of education, guidance, knowledge and coping strategies that the network can provide. The traditional discharge support approach provides paper brochures to patients and requires them to adapt to meeting held at a time and location not necessarily convenient for them in order to gain information. In contrast, a discharge support approach that uses the Internet as a base can provide the information patients need at their fingertips 24 hours a day. HeartCare may therefore be a superior and more cost-effective method to help patients recover from CABG surgery than current strategies, whose success in supporting patients and families in their achievement of recovery goals has not been outstanding (Beckie, 1989; Gortner et al, 1988,1990; Marshall,Penckofer,&Llewellyn, 1986).

However, in their current state the resources available on the Internet are overwhelming and not easily accessible to patients who may lack technical skill and information access strategies. In order for this source of information to be truly useful, it must be easy to use and there must be selection of resources available to meet the patient's preferences and needs. With his/her own designed "Home Page" (a WWW convention that supports tailored linkages to Internet resources and personalized organization of knowledge), every patient will have an individualized recovery program which includes the resources he/she has identified as valuable and which can be adjusted according to changing needs during the different stages of recovery. We will also devise and implement a strategy for patients to evaluate WWW resources they discover on their own.

There are many questions emerging about the implications of using Internet and WWW-based resources in nursing practice. What resources exist? What is the quality of WWW-based information resources? How does one match available resources with a patient's set of information requirements? Appropriate integration of these resources and other, more expansive patient information resources into clinical practice is the subject of some debate (L. Buhle, Personal Communication, May 19, 1995). To some, direct patient access to clinical information represents a threat to clinician dominance and challenges clinical authority. To others, patient clinical information resources are essential for partnerships in care. Appropriate integration of computer resources into the clinical practice of nursing stands to make efficient use of scarce nursing resources and benefit patients directly. However, while computerized resources may be of great use to patients, they remain out of the reach of overworked nurses because (1) nurses lack familiarity with these resources; (2) where such computerized resources exist, there is no strategy for integrating them into practice; and (3) evidence is needed demonstrating when expensive and complex approaches improves patient outcomes. Integration of computerized resources requires monitoring the effects of introducing these new technologies into nursing care and exploration of the manner in which technologies like HeartCare increase the demand or relieve the burden for nursing resources while improving patient care. This study will provide important, needed information to determine when, and under what circumstances, the expense and complexity of Internet resources make a valid contribution to the nursing care of patients in the CABG-period.

Conceptual Framework

This study will be guided by the work of Litwak (1985) which identified the complementary nature of informal and formal groups in the accomplishment of patient care. Informal groups consist of the patient's family, friends, and neighbors. Formal groups include agencies that are organized around expertise (hospitals, physicians' offices, cardiac rehab centers). Currently, patients rely heavily on formal care groups during the early recovery phase, including nurses in hospitals. With the change to early discharge, however, many of the tasks once accomplished with the aid of the formal groups must be accomplished with the informal group. Nurses, part of a formal group, can employ computer technology such as the HeartCare to enable patients to access needed resources from formal and informal groups. Thus, it allows us to extend the expertise of formal groups to informal groups.

We consider HeartCare a boundary-spanning concept, permitting the formal care system of nursing to achieve patient care goals through more efficient support of patients. The proposal is submitted in response to the Program Announcement on Nursing Informatics. This randomized clinical trial evaluating HeartCare addresses the challenge raised in the Priority Expert Panel's report on Nursing Informatics (NINR, 1993) to "investigate new technologies to create tools for patient care" (p. 12) and to "measure, if possible, the contributions of information systems to patient outcomes" (p. 83). Furthermore, the HeartCare intervention, as an adjunct to professional nursing care, has the potential to ensure high quality nursing services delivered to patients in their homes.



HeartCare Pilot. As a result of a HeartCare pilot funded by an internal grant from CWRU we have developed a prototype of HeartCare web page, implemented the patient information needs assessment and home implementation protocols, and received initial review of the screens, content, and data collection protocol by clinical staff. Ten patients (8 men and 2 women; age range = 50 to 74; mean hospital length of stay = 5.6 days) were approached on the third postoperative day for all assessment and HeartCare instructional activities in the hospital. Four patients received full information needs assessment interviews from which data were applied to mock tailoring activities (Appendix F of this proposal contains examples of this tailoring process for two of the pilot subjects). Six patients were given demonstrations of HeartCare and asked to use it during an instructional session in the hospital. The instructional/demonstration sessions were done individually at the bedside of each patient by Dr. Moore. Patients were alert and comfortable enough to learn how to use HeartCare during a 20-30 minute training session in the hospital as early as the third day after surgery. A portable computer with the HeartCare prototype was sent home with 2 patients who were given access to WWW-based cardiac resources. These 2 participants were visited at their home and given additional instruction on the use of the HeartCare computer resources and use of the Internet. Of the 6 patients who had access to HeartCare, one had extensive experience with computers, two had some experience, and three had no prior experience with computers.

Our experiences with the patients in this pilot study have provided the following information: (1) although we implemented selected segments of the HeartCare interface and server using a laptop, we plan to stay with the Net Computer because it is cheaper and simpler to use; (2) HeartCare is acceptable to most patients and they would be willing to try it at home; (3) further assessment of WWW-based resources is necessary to assist patients to access information that is understandable to lay persons (perhaps a topical index of URL's that other patients have found particularly helpful); and (4) a range of features may be needed to engage the varied interests of individuals, i.e., some people are primarily interested in e-mail features to ask questions of professionals, others what to be able to send personal health data about recovery to a nurse, others want information about different aspects of recovery.

HeartCare was also demonstrated to the staff of the cardiac care recovery unit. Staff were enthusiastic about the capabilities of HeartCare and recommended the following improvements: (1) increase the size of the print on the screen (which was done prior to demonstration and use with patients); (2) add more recovery symptoms to the patient self report health assessment; and (3) add all the current information that they teach patients in the hospital so patients could review it first on HeartCare, to be followed by reinforcement by the nurse. Interest also was expressed in the ability to respond to patients from the nurses' homes; staff members were anxious to develop new ways to extend their presence with patients given the short hospital length following this extensive surgery.

The proposed project builds on several, ongoing programs of research investigating the application of computers and computer networks to the challenges faced by home care patients and their informal caregivers. Drs. Brennan and Moore collaborated on the original ComputerLink projects and established that persons naive to computers can and will use innovative technologies to obtain information, decision assistance and peer support. Dr. Moore's recent success in developing interventions for persons recovering from CABG procedures. With the addition of Dr. Gustafson as Co-Investigator, we have an opportunity to study in more depth the engineering and social science dimensions of Internet-based Intervention.

ComputerLink. From 1988-92, Dr. Brennan directed the ComputerLink projects, two field experiments examining computer home care networks for AIDS patients and Alzheimer's Disease caregivers in the Greater Cleveland area. ComputerLink used the information resources of the Internet through FreeNet, the free, public access computer network based at Case Western Reserve University in Cleveland. The effects of ComputerLink on persons living with AIDS (NR2001) (Brennan, Ripich, & Moore, 1991) and caregivers of Alzheimer's Disease patients (AG 8614) (Brennan, Moore, & Smyth, 1992; Brennan, Moore & Smyth, 1995) were evaluated in field experiments involving 60 AIDS patients and 102 Alzheimer's Disease caregivers, respectively. Each group was randomly halved, with one group receiving typical home care and the other using ComputerLink.

ComputerLink provided three services to users: 1) The Electronic Encyclopedia consisted of information screens designed to enhance self-management, promote effective home-based treatment of patients and promote patient/caregiver understanding of illness-specific issues. Information needs were derived through a series of interviews with potential users and through analysis of existing home health care services delivered to AIDS patients and Alzheimer's caregivers. Electronic texts, including authorized reprints, were constructed or obtained to meet patients'/caregivers' stated needs. Users browsed through over 200 electronic pages. The Electronic Encyclopedias were expensive to build (6-12 person months) and maintain. The level of sophistication in terms of content and presentation was understandably low --the project team targeted an eighth-grade reading level and a universal interface. Criticisms from study participants focused on lack of depth in the material (1-3 screens at most) and the uninteresting presentation. In the HeartCare intervention we will make use of a vast and varied array of information resources on the Internet, and capitalize on the now-available, sophisticated interfaces.

The second service, Communication, included several public/private options:(1) an unrestricted public bulletin board, which allowed users to post by name or anonymously anything that was on their minds for open, on-going discussion; (2) private electronic mail, through which users could send and receive their own private electronic mail, including messages from the Project director/ Nurse Monitor answering their personal health care inquiries; (3) a question/answer area, in which answers to questions posed anonymously by users were posted by the team.

The third service, Decision Support, helped ComputerLink users make choices about which personal decisions were necessary, how they could best express these decisions, and how they could best generate insights for such decisions. This area was least used in both experiments (89 times by the AD Caregivers; 140 times by the persons with AIDS). However, because it lacked broad appeal and because the decisions faced by recovering CABG patients are more elusive and less well-defined, we will not include this in the first version of HeartCare.

Each of the ComputerLinks had a Project Director who served as the nurse moderator. The Nurse Moderator paid an initial in-home visit to each participant to assess his/her needs and provide training in the use of the ComputerLink. The Nurse Moderator then scanned the public areas of the ComputerLink daily, updating information, posting announcements and trouble-shooting. She monitored users' public and private accesses, responded via electronic mail to users' questions, and facilitated discussions on the public forum. Dr. Moore was the nurse moderator for the AD Caregivers ComputerLink. HeartCare will also make use of a NurseModerator.

During ComputerLink's 4-year period of operation, 15,000 user accesses averaging 10-13 minutes per log-on were recorded. The Persons Living With AIDS (PLWA) ComputerLink system remained active for a total of 325 days (47 weeks) to permit all participants a full 6 months of participation. All participants accessed the ComputerLink (mean access count per subject 297; median 129); generally participants accessed the system twice a day. Over the entire experimental period, PLWA accessed the system a total of 8449 times. A typical PLWA encounter with the ComputerLink lasted 12.5 minutes. PLWA used the ComputerLink at all hours of the day and night. Most encounters occurred between 10 PM and 3 AM and included use of at least two ComputerLink functions. The AD caregivers' ComputerLink remained active for 19 months, permitting the users 12 consecutive months of access. There were a total of 3875 accesses to the ComputerLink. AD caregivers accessed the ComputerLink a mean of 83 times (s.d.101.86). The mean length of encounter was about 13 minutes; most accesses involved the use of two or more features. In both ComputerLink studies, users accessed the communications area most often (in at least 85% of all encounters) and spent most time in this area (average of 10 minutes in a 12-13 minute encounter). The enthusiasm for "connecting to others in the same circumstance," as reported by one user and found by other, similar projects (e.g. Gustafson et al, 1994), leads to the inclusion of communications area within HeartCare.

Qualitative inquiry revealed that ComputerLink served as a "support system without walls" . Modest effects of ComputerLink on decision making and social isolation were discovered. Among PLWA, use of ComputerLink led to a reduction in social isolation once the effects of depression were controlled (R2 .39; F 16.24; df 2, 48; p <.01). Among AD caregivers, ComputerLink led to improved confidence in decision making ( net change ComputerLink Group +1.53; Control Group -0.20; F = 9.73; df 1,93; p < 0.001). Guided by the Andersen model of predictors of service use (Andersen & Newman, 1973; Andersen & Aday, 1978), the team established in a secondary analysis that use of ComputerLink varied across time, and that caregiver gender, years of school attended and emotional strain experienced accounted for 18 % of the variance in ComputerLink use. Caregivers' health state per se was not a barrier to ComputerLink use and gives us confidence that newly discharged persons will also be able to make use of supportive resources provided in HeartCare. For nonspouse AD caregivers to persons, access to information, moreso than communication, reduce the negative health consequences of Caregiving (b = -.2.52; Bass et al; See Appendix B). Results also indicate lagged effect of ComputerLink use on subsequent use of formal services (McClendon, Bass & Brennan, 1995); we will monitor service use among all participants (Appendix I).

ComputerLink was the first project ever to demonstrate that existing computer networks can be customized to address the home health care needs of chronically ill patients and their caregivers; it featured two innovations: 1) each of the two ComputerLink systems was designed to meet the clinical needs of a specific population; 2) the services were designed to be usable by all persons, not only ones with technical skill. ComputerLink thus has major long-term implications for effective, low- cost, indirect clinical service delivery and it serves as a prototype for HeartCare. It also had several limitations. First, the interface was primitive and did not support interesting excursions through various information resources. Therefore, in the present project we will use a more sophisticated computer for the home-based unit. Second, the positive benefits of even a primitive information service (the electronic encyclopedia) argue strongly for providing enhanced information resources in a manner that is timely and convenient. Finally, and most importantly, ComputerLink existed outside of the formal health care delivery system. For technology to be truly useful to nurses and patients, it must be integrated into the formal health care system. Therefore, we will build HeartCare as an adjuvant intervention, well embedded in the formal nursing and health care system.

The HeartCare project builds on and extends the ComputerLink projects in several important ways. First, HeartCare will deliver text, video, still images and sound into the home through a more sophisticated and engaging interface than that provided by ComputerLink. Second, the informational component of HeartCare will be enriched and more varied, resolving the complaints that the Electronic Encyclopedia in the ComputerLink was primitive and without depth. In addition, because we will draw on resources available on the Internet, information management and updating will be done externally and will thus not tax grant resources. The role of project personnel will be to locate and evaluate resources, not develop them. Finally, HeartCare will extend the concept of technologically mediated home care to an acute population.

Other work by Dr. Brennan demonstrates her ability to develop and evaluate innovations to assist nurses in practice. Collaborative Care Unit (CCU). Through a study funded by the Robert Wood Johnson Foundation/Alpha Center, Dr. Brennan evaluated whether collaborative care managed by nurse practitioners reduces costs, improves functional status, reduces morbidity, and provides a habitable work environment for nurses and physicians (Brennan et al., 1991; Brennan, Emmons, & Silvers, 1991; Genet et al., 1995; Brennan, Daly, Landefeld, Phleps & Rosenthal, see Appendix B). The CCU project has illuminated the time and clinical demands faced by staff nurses in acute care hospitals; this information has helped us to establish a protocol for HeartCare introduction that is mindful of the practice challenges faced by acute care nurses. Through this project Dr. Brennan has also gained skill in monitoring the impact of practice innovations on the scope of nursing practice and time demands faced by hospital nurses.

Nursing Practice Models. A key characteristic of hospital-based nursing practice is that it is a group practice model. Traditional investigations of nursing practice, however, have focused on interventions directed to the individual client by a specific nurse. Dr. Brennan developed and directed evaluations of indices to measure practice models in nursing homes (OAIC/ [AG10418] and in acute care hospitals (NR 5538). These projects are methodological in nature, designed to establish an index of the multidimensional phenomenon known as "nursing practice model" (Zurakowski, Brennan, & Kahana, under review; see Appendix B). The practice model projects will support and inform the activities of the HeartCare project by explicating the nursing practice environment into which HeartCare will be introduced and provide a measure of the impact of HeartCare on practice.

Computer Tools for Patient Care In collaboration with Rehabilitation Clinical Services, Metro Health Medical Center, Dr. Brennan is currently developing a CD-ROM that will provide instructions and guidance to spinal cord injured patients in the strategies necessary to prevent pressure ulcers. This project formalizes many nursing discharge teaching strategies into a medium that is entertaining, informative and portable. Prototypes are being developed using JAVA and the html environment that will also be used here. Social marketing theory provides direction for developing patient-motivating informational tools.

Drs. Brennan and Sreenath demonstrated the use of hand-held computer (PDA) by physicians and nurse practitioners in long term care (AG 10418). The goal was to develop easy access to patient-specific and Internet-based information resources. The screen layout challenges and information resources assessment tasks of the PDA project have increased our skill in interface design and information appraisal, two key tasks for HeartCare. We developed the Usability index in the PDA project which we will modify for HeartCare (Appendix I)

Finally, for several years Dr. Brennan has collaborated with a pediatric hospital to design ComputerLink-like services to aid in the discharge and post-hospital support of children with complex medical problems. Although addressing a different population, the pediatric project's functional aspects (discharge and post-hospital support) and its character as a service integrated within the formal health care system are directly relevant to the HeartCare intervention. Dr. Brennan has developed strategies to assess information needs of the patient at home (Youngblut, Brennan, & Sweigert, 1994) and to assess stakeholders' views of computerized discharge support tools (Payton, Brennan, & Ginzberg, 1995).

Cardiac Recovery Studies.

Dr. Shirley Moore has combined qualitative and quantitative research methods in a series of studies focusing on the development of interventions to facilitate recovery following acute cardiac events. In three studies Dr. Moore has developed and refined a discharge information intervention for CABG patients to increase the congruence between expected and actual home recovery experiences. In her first study, using qualitative methods, descriptions of CABG recovery experiences were elicited from patients as a basis for developing preparatory information for home recovery (See Appendix B for a description of this study). Results included a detailed description of commonly occurring symptoms that occurred in each of the first 4 weeks of recovery, in the patients' own words. Using this information, a 15-minute audiotape was professionally produced to "walk" patients through the first month of home recovery and suggest ways to manage common experiences. This program has been named the Cardiac Home Information Program (CHIP).

In a second study CHIP was tested on 82 (68 men and 14 women) CABG patients. Using a quasi-experimental design, recovery outcomes were compared between two groups: those receiving CHIP and those receiving a standard cardiac discharge information program. Subjects ranged in age from 34-84 years (mean = 65); 93% were European American and 6% were African American. Patients receiving CHIP had better physical functioning than those in the control group (experimental group mean=8.5,SD=5.5; control group mean=13.2,SD=9.9)(F = 9.42, df =1,80 , p = .001) at 1 month following discharge. No effect was found on psychological functioning or perception of family functioning. Patient acceptability and utilization of CHIP were evidenced by the fact that patients listened to the audio tape an average of four times (SD = 6.43) each, with a range from 1 to as many as 30 times, mode of 3. CHIP was used most frequently in the hospital just before discharge and during the first few days at home. Patients reported that CHIP accurately described their experiences during recovery. All subjects stated that CHIP gave them helpful suggestions for dealing with recovery. The lack of an effect on perception of family functioning may have been due to a weak dose effect in that CHIP contained descriptions of family problems that might occur during recovery, but few gave suggestions for managing them.

A third study was undertaken to further explore CABG recovery experiences specific to women. Twenty women (mean age=67) were interviewed at three points during recovery to obtain descriptions of their experiences. These women had unique physical experiences, such as breast pain, and several (N = 4) were discharged on oxygen therapy. The women reported being very disturbed by the chest incision disfigurement. Women's social circumstances also differed from those of men and influenced their recovery patterns. Furthermore, only 40% of the women lived with a spouse. The women were worried about who would take care of them during the recovery period; many were not able to recover in their own homes. Thus, in the proposed study, the presence of a caregiver and social support will be monitored as possible intervening variables affecting recovery. Next, further refinement of the CHIP intervention was undertaken in a study of the early home recovery experiences of fast track (discharge on postoperative day 4-6) CABG patients. Thirty-two individuals (15 women and 17 men) (mean age = 63 years) were interviewed at discharge and again in their homes 24 hours later. Recovery needs described by patients included information about incision care, pain management, energy conservation, and sleeping and eating difficulties. This study has provided data about the feasibility of interviewing CABG patients on the 4th and 5th postoperative day. Patients were alert, relatively comfortable, able to be engaged in a 40 minute discussion, and able to describe their concerns and information needs about home recovery. The CHIP intervention has been revised to include information gained from these CABG studies of women and early home recovery and is currently being tested in a study recently funded by the American Heart Association.

These studies demonstrate that even a simple, technically limited CABG recovery intervention can have a positive impact on outcomes. However, there were several limitations to the audio taped CHIP intervention. Although the tape is useful for the first month of recovery, it cannot be updated or changed as the patient's status changes. Also, the content of CHIP is fixed (standardized) and cannot be tailored by the nurse to each patient's specific recovery situation. The proposed study will expand the use of technology to support home recovery by developing and testing a program that is both more tailored to the individual and more comprehensive.

Focus Group Interviews. Dr. Moore recently conducted a series of focus group interviews with individuals in cardiac rehabilitation programs. A convenience sample of 40 subjects (20 women and 20 men; 75% had CABG surgery, 40% had a myocardial infarction) have participated in six focus groups. Results indicated that between 3 and 6 months following an acute cardiac event, subjects desire more information about diet, weight control, stress management, and cardiac medications (including antihypertensives). Both men and women stated that making life-style changes to reduce cardiac risk was much harder than they had expected and felt that they needed assistance from both peers and professionals. Information gained from this study supports the need to address risk factor modification goals of patients during the first 6 months of recovery. Dr. Moore's experience in both conducting focus groups and analyzing data from focus groups will be useful in the proposed study, in which focus group interviews will be used to examine both patients' and nurses' experiences with HeartCare. Additionally, Dr. Moore's experience teaching health professionals techniques for continuous quality improvement (CQI) will be valuable in setting up the interfaces between the quality monitoring systems desired by the clinical personnel and the HeartCare project.

Dr. Moore's experience in recruitment, implementation of an intervention and data collection in field studies on the clinical unit where we will recruit CABG patients provides an important entree to the practice. Her experience with frequently encountered research problems associated with this population, such as attrition due to relapse, complicating medical problems, and mortality, will be valuable in the execution of this project. Further, several measures of study variables (physical functioning, psychological distress, and family functioning) to be used in this project have been used by Dr. Moore in her studies and were found to be valid and reliable for post CABG patients.

Computer Home Care Support. Additionally, Dr. Moore was the Project Director on the AD Caregivers ComputerLink project and gained experience in the installation of home computers for caregivers of these patients. Dr. Moore trained the caregivers in home computer use and served as the nurse moderator for the network, providing information, emotional support and group facilitation. Dr. Moore's experience and knowledge of nurses' roles in the delivery of caregiver interventions using computers, as well as her experience in assisting naive users to use computer technology, make her a great asset to the project.

Other Team Members: An Industrial Engineer noted for his work in computers and health promotion (David Gustafson) , a cardiac surgeon (Dr. Geha), a systems engineering with expertise in interface design (Sree Sreenath), and a well-respected statistician (Ralph O'Brien) complement our team. Dr. Geha will provide clinical advice and medical oversight to the project. Dr. O'Brien's work in power analysis for clinical trials facilitated our study design. In addition, he devised a scheme for testing explicit contrasts and summarizing results of multiple hypothesis tests that will make excellent use of the small sample size dictated in part by feasibility considerations of this project. The project director, Gail Casper, brings extensive experience in assessing patients' own desires and preferences. The consultant team brings expertise in acute nursing care of patients following cardiac surgery (Dunbar, Artinian), and contemporary home health care (Stricklin). A Clinical Advisory Panel comprised of nurses will serve as a reaction panel to help evaluate the impact of HeartCare in nursing practice.



The specific aims of this project are to develop HeartCare, an Internet-based cardiac recovery resource, and to evaluate its effects on nursing practice and patient outcomes. We propose to achieve this through an intensive 6-month development phase and a 2-year experimental period. We will develop a HeartCare World Wide Web server, by searching and linking to Internet-based WWW resources as well as developing our own. We will devise programs that patients will use on computer terminals (net computers) at home that are tailored to an individual patient's information needs . We will conduct another, more extensive pilot test of the entire HeartCare intervention during Year 1, with particular focus on identifying and managing the implementation challenges experienced by patients in the home.

From months 8 through 22 we will conduct a randomized field experiment to answer the research questions. We will answer Research Question 1, What are the nursing practice impacts of HeartCare? through documentation, interviewing, and group assessment methods. To answer research question 2, How does HeartCare contribute to patient outcomes?, we will conduct a three-group randomized experiment involving 140 recovering CABG patients. One group, "Usual Care", will receive standard post surgical care and no technological intervention. The second group, "CHIP", will receive standard post-surgical care and the CHIP intervention. The third group, "HeartCare" will receive standard post surgical care and the HeartCare intervention. Activities to answer the two research questions will occur concurrently.

Research Question 1: What are the nursing practice impacts of HeartCare?

Assessing the impact of nursing informatics applications on nursing practice is well recognized as both important (NINR, 1992) and difficult (Kjerulff & Moray, 1988), and benefiting from both quantitative and qualitative strategies. Because of the preliminary nature of the proposed work, it is important to project effects on practice under controlled circumstances before proceeding to integrate innovations like HeartCare into clinical nursing practice. We will employ a model of assessment derived from Rogers (1983) and used in our earlier work (Brennan, Moore, & Smyth, 1991) to estimate the effects of introducing HeartCare into nursing practice. We will use non-invasive strategies to document and forecast the nursing practice effects of HeartCare. For example, we will 'benchmark" the time and skill necessary to set up a HeartCare computer and conduct the patient information needs assessment by observing project research staff (not active, clinical staff). We will also use assessment techniques such as focus groups involving our Clinical Advisory Panel and staff nurse volunteers to explore the impact of HeartCare on nursing practice without disrupting the existing practice within our clinical setting. We recognize that having nurses employed for the HeartCare study interview patients and train them in the use of HeartCare, limits our ability to generalize about how exactly hospital-based nurses currently in practice might integrate tools like HeartCare into their routine post-operative and discharge practice. However, our strategy will allow us to obtain indications of the impacts of HeartCare on nursing practice, and its effect on patient outcome.

We will use the following questions as a guide for the project staff to benchmark the time and skill necessary to set up a HeartCare computer and conduct the information needs assessment for every patient: (1) What length of sessions were required to evaluate the patient's information needs? (2) What information did clinicians request about the system? (3) What percent of the information resources determined necessary for the patient were available on HeartCare? (4) Did patients interactions with clinicians change during their assessment and introduction to HeartCare (e.g., more questions posed to clinicians, display of greater anxiety, etc)? (5) Was special equipment needed to conduct the information needs assessment, demonstrate HeartCare, or set it up?

For an innovation to be accepted its intended users, it must have relative advantage, compatibility, complexity, divisibility, and communicability (Rogers, 1983) . We will obtain evidence of the adoption of the innovation by monitoring all aspects of the HeartCare procedure, obtaining estimates of the time required to complete the information needs assessment, patient training, and home set up, the reaction of patients and their clinical providers to the project, the equipment, and any subsequent references to HeartCare by the patient. The project director and all data collectors will keep a field diary. The field diary will include a log of e-mail messages from participants and a record of the time needed to answer them. Quarterly we will review logged information, summarizing time estimates, enumerating special events, and describing barriers to the implementation of HeartCare. Dr. Moore will be responsible for summarizing this information. Consistent with contemporary methods of continuous quality improvement techniques, data from the quarterly assessments will be analyzed and used to make appropriate changes, such as locating new Web sites.

To monitor and describe the changes in nursing practice brought about by HeartCare we will impanel a team of six cardiac care nurses to meet semi-annually with the research team (see Appendix A for letters of support and willingness to participate). The purpose of the meetings will be to review the development and implementation of HeartCare, inform us of changes in clinical practice due to organizational, clinical, or legislative reasons, and construct scenarios anchored in our actual experiences of the anticipated experiences of other nurses. We will employ the critical success factors technique to obtain perspectives on implementation challenges. This technique involves brainstorming and feedback strategies that illuminate factors necessary for practice. Dr. Brennan participated in this strategy in an earlier project addressing discharge-and-home-care computer technology (Payton, Brennan, & Ginzberg, 1995; see Appendix B). Once a year, Mary Lou Strickland (consultant), President of the Visiting Nurses' Association of Cleveland, will participate in the meeting to ensure that the discussion appropriately handles home care issues. Additionally, each year, a series of focus group interviews (two groups of 6-8 members each) will be conducted with nurses from the inpatient cardiac unit to obtain their perspectives on the impact of HeartCare on nursing practice.

Research Question 2: How does HeartCare contribute to patient outcomes?

A field experiment will involve 140 patients in a 6-month trial; all participants will receive usual post-CABG care; in addition, each patient will be randomly assigned to one of three groups: HeartCare, CHIP (an audio tape intervention) or no Usual Care Only (no technological intervention). Assessment of patient outcomes will be driven by the research questions to evaluate how patients use HeartCare and to compare the HeartCare group and the CHIP intervention group at key time points in terms of physical function, symptom distress, psychological distress, perceived family function and adherence to cardiac risk behavior reduction. Group by time differences are hypothesized for each of five variables, employing a strategy of explicit contrasts to evaluate differences between HeartCare and CHIP patients at 1 week, 1 month, 3 months and 6 months. Usual care only patients will complete data collection forms in the hospital (baseline) and at their standard 6-week follow-up visit. We will model the change over time for the Usual Care Only group describe the CABG recovery experiences unaffected by experimental attention or intervention.

Setting and Target Population

Participants will be recruited from University Hospitals of Cleveland (UH), Ohio, a non-profit, private hospital with over 600 beds. It has a cardiac intensive care unit and step-down unit for CABG patients. This large urban, teaching hospital has a total of 335 CABG (non-valve) patients annually, approximately one third of whom are women. Approximately two-thirds of the patients will meet the study criteria, resulting in 211 possible participants for recruitment annually. This represents a total of 70 women and 141 men for recruitment yearly. Approximately 11% of the patients in the sample are expected to be African American. The average length of stay post surgery for CABG patients in 1994 was 9 days (mode = 6 days). Fifty percent of the patients were discharged between the 5th and 8th postoperative day. A Phase I cardiac rehabilitation and discharge teaching program is in place. Assurance of access to patients has been received from UH (Appendix A, see letter of Elizabeth Mairano).


The sample will consist of 140 CABG patients who meet the study criteria and provide consent to participate. Criteria for inclusion are: (a) stable and transferred out of surgical intensive care (about 24 -36 hours after first CABG surgery); (b) able to speak, read and write English; (c) living within a 100-mile radius of Cleveland, and, (d) because of the nature of the experimental intervention and the data collection protocol, having a personal telephone number. Patients experiencing major complications from surgery, such as myocardial infarction during or since surgery, pulmonary emboli, hemorrhage, or cerebral vascular accident, will be excluded because each of these complications has been associated with more adverse recovery outcomes. The number of African Americans and women in the proposed study will profile the percent of patients in Cleveland having CABG surgery, which is 10% African American and 28% women, which exceeds the national rate of 6% and 27% respectively. To control for previous experience with recovery, only first-time CABG patients will be included. Only those patients living within a 100-mile radius will be included in order to ensure feasibility of home visits. The sampling plan allows for an uneven cell size to support more power in the comparisons between the two experimental conditions. Power computations were provided by Dr. O'Brien; the sample size provides power ranging 0.83-0.99 to detect group differences over time at a 0.05; effective a 0.01. ( Appendix C).

Rationale for Sample

The decision to study patients recovering from CABG surgery represents our view that innovative hospital-home transition strategies are best studied in a group with (1) sufficient numbers of individuals are in need of assistance; (2) the challenges faced by these persons in the transition period are similar to those faced by patients recovering from other serious illnesses and procedures; (3) the potential gains from effective home care, in terms of recovery, prevention of readmission, and patient well-being, and (4) the course of recovery is generally well known. Once the success of our intervention is demonstrated with this well understood group of patients, it will be possible to extend the strategy to groups with less well understood recovery courses or more complicated treatment protocols.

The sample size takes into account subject mortality and other circumstances. With a one-year cardiac survival rate of 92 to 95% (National Cardiac Surgery Database, 1993), we may lose about 5-8% of our sample. It is not the intent of this study to determine whether the use of HeartCare reduces post-surgical mortality due to more prompt access to treatment. Therefore, we will monitor, but not evaluate differential death rate. A total sample of 140 participants (50 per experimental group; 40 controls) is appropriate to ensure resolution of hypotheses. Given the exploratory nature of our research, and the chance that participants may use HeartCare idiosyncratically, the power analysis provides simply a guide for estimating the likelihood of group differences over time.

Sampling Procedure

A list of women and men meeting the study criteria will be prepared daily by the clinical nurse specialist on the cardiac step-down unit. All consecutive patients meeting the study criteria will be approached to be in the study. Sampling of African Americans and women will continue until they make up 10% and 28% of the sample, respectively (or enrolling until at least 97 Caucasian and 86 male participants are recruited) to ensure that the representation of racial and gender minorities is acceptable. Using random numbers obtained from a table and printed on a sealed envelope, participants will be randomly assigned to the HeartCare (n=50), CHIP (n=50), or Usual Care (n=40). Participants' assignment will not be known until baseline data is collected.


The independent variable in this study is a program of post-CABG home care support. There will be three groups -- a custom-computerized home care support system (HeartCare), an audio tape (CHIP) home support program, and usual care only. Participants will be randomized to one of these three conditions, two of which represent the experimental interventions (HeartCare and CHIP) and one control (usual care) that is being used to monitor changes in CABG concurrent with the experimental interventions that could alter interpretation of the results.

Heart Care

HeartCare is an integrated, post-cardiac surgery recovery intervention built on the World Wide Web. The intervention consists of hardware, special programs resident on Net Computers (provided by project funds) in the participants' homes, and cardiac recovery resources available on the HeartCare Server as well as anywhere along the Internet. In addition to providing information resources, we will capitalize on the communication utilities of the Internet to support peer and professional communication. We will construct a Cardiac Recovery Use Group and an electronic mail service, complete with local aliases (efficient addressing), for all participants.

The patient's view of HeartCare will be through the home computer. The home Net Computer will be configured with two types of programs. One set of programs will mange the patient's access to and interaction with the World Wide Web and other aspects of the Internet -- these programs include Netscape 3.0 , the patient's individually tailored home pages, data collection routines, and macros to manage telephone connections to the HeartCare Web Server. The second set of programs will include a health diary, a set of clinical standards targeted to the day of recovery, special resources and the several programs designed to manage access to the HeartCare routine. Patients will access HeartCare by turning on the Net Computer in their homes and will see:


Good Morning, Mary Jane!

It is Day 5 after Discharge.

Usual activities for Day 5 include:

Taking and recording your pulse

Examining your leg and chest incision for discoloration or drainage.

Walking inside your home.

Select the icon below to:

Chat with See a Record Learn about

Other Healthy Heart Diet your Cardiac Care

Recoverers video Health Appraisal Recovery

***** %%%%% #### @@@

What is behind the HeartCare Screen? The various points on the HeartCare screen will automatically transfer the patient to the sections on their local computer, the HeartCare server, or other systems on the Internet. Each screen will be tailored with local information relevant to the individual patient's status and the options open to the patient at that point. For example, selection of "Record your health appraisal" will lead to the next screen:

Day 5: Health Appraisal

Pulse at your wrist (# beats in 60 seconds): _____

Color of your incisions

Chest:yellow-brown white green pink red

Leg:yellow-brown white green pink red

Unusual Odors? Yes No

How long did you sleep uninterrupted last night? ____minutes/hours

What kind of pain medication are you taking? __________________

How often? __________________ Does it work? __________

Any special events, feelings or reactions?_______________________

Would you like to save the information

save, then send to the Cardiac Recovery Nurse

erase and start over


Selecting the "Diet Video" would result in a link via WWW to the NIH where a set of patient education reside. Opting to report the health appraisal data to the project director will rely on the FORMS capability of Netscape.

HeartCare involves three components: construction of the HeartCare Web server (including all Internet features), in-hospital assessment of patient needs and tailoring of access to Internet resources, and in-home access via a customized interface to WWW resources that promote self monitoring, home management, and motivation.

HeartCare Component 1: HeartCare Web Server.

The HeartCare Web server will be actually be two identical (UltraSPARC) mirrored at University of Wisconsin-Madison and at CWRU. A Web Server is a repository of computer programs, information documents, and pathways (hypertext transfer protocol [http] links) to Internet-based information resources. The key feature of the HeartCare Web Server will be pathways to specific Internet resources relevant to the patient recovering from CABG surgery.

A large number and variety of information resources are available on the World Wide Web on the Internet that could be very valuable during patients recovery after CABG surgery. For example, searching under the key words: "heart disease" returns 165 resources, and 75 resources are found under "cardiology." Each resource has an additional number of sub-options which together provide a wide range of patient information. (See Appendix D for a summary of WWW hypermedia resources related to cardiac recovery). An important advantage of the WWW approach to developing cardiac recovery programs is that we may capitalize on programs already present on the Internet. We can make efficient linkages (computerized Internet addressing pathways) to these programs and information resources. Examples of these resources are now described below.

Although nurses teach patients basic self-monitoring skills in the hospital (e.g., taking a radial pulse), patients may require or desire information to refresh or remind themselves about self monitoring. One WWW site demonstrates how to interpret laboratory results; another helps patients learn how to monitor pulse and blood pressure. To provide anticipatory training to deal with serious complications, information is available at a Web site about how to act in emergencies, including instructions in First Aid and how to detect and monitor arrhythmias. Some patients may be motivated by understanding the basic physiology and pathology of their conditions. WWW-based resources could help a patient who wants to gain better knowledge of the heart's physiology. He/she could choose between pictures or textual explanations of the structure and function of the heart. Such information is available at many levels of sophistication. Patients may want a better understanding of heart disease and its treatment. They could choose a simulated trip through the cardiovascular system, showing how plaque is formed and what happens when a coronary artery is plugged or watch a video on CABG surgery.

Other resources on the Web, such as programs for smoking cessation or weight reduction, can help CABG patients adjust their life styles and prevent secondary heart disease. A variety of choices between different types of exercises are provided, as well as information about cardiac rehabilitation from the recently released clinical practice guidelines for cardiac rehabilitation (Wenger, et al, 1995). A number of healthy eating instructions with heart smart menus and references for heart healthy cook books are available. Being able to efficiently access education materials through the tailored WWW interface will enhance nurses' abilities to facilitate patient recovery.

A major challenge in using cardiac recovery resources on the WWW is evaluating the quality of the information presented and its relevance to an individual person, a challenge found also when printed materials are used for cardiac recovery. The clinical quality of these WWW resources and their congruence with local practice standards must be assessed before they can be fully integrated into the patient's care. We propose to establish and implement a strategy for review during the one-year start-up for HeartCare. To select appropriate resources from the Internet for access via the HeartCare Web Server, we will employ a strategy for evaluating Internet cardiac resources that includes educational evaluation according to the criteria proposed by Rankin & Stallings (1990). Assessment of the available resources on the WWW-Internet will include attention to system requirements for patient health information as well as assessment of the content of the health information resource, including accuracy, scope, authenticity, extent to which the resource reflects current knowledge, level of difficulty and coverage, readability, grammar, clarity of communication, capacity to hold interest, and technical quality (Appendix E contains evaluation criteria for each WWW site).

Technical considerations A mirrored pair of computers (ULTRASparc) will serve as the HeartCare Web Server. A Web Server is capable of storing all of the local routines necessary for supporting external linkage to an Internet connection (e.g. verification, use monitoring, etc.). A centralized set of programs will reside on the Web Servers and be physically linked to the Internet node and through there to all Internet Computers. The Web Server will be physically secure (i.e.. in a locked room in the Dr. Brennan's research area with the computer running the mirrored site located in Dr. Sreenath's lab at CWRU). Because institution staff provide 24-hour a day network support they will be responsible for maintenance of the server (See Appendix A for letter of support). This support will assure speedy response in the event that a system failure occurs, and will permit us to take advantage of new developments in Internet access, Netscape updates, and computer network security. An internet service provider (ISP) will manage the initial contact with the servers, and provide updates as needed. The ISP provides an access point through which participants' net computers will link with the HeartCare servers. No research information will be stored there.

Design of user interface would be iterative and respect the fundamentals such as ergonomics - commonly needed information/utilities should be accessible with a single click, the screen layouts/ visual design consistency, etc. A basic character of the HeartCare project is that the user interface has to be designed primarily for information access by the user, and less for information capture. The content seen by the user throughout the HeartCare project would have been screened and further customized to the patient needs. Anytime a user needs to go beyond the screened contents of HeartCare, we will provide visual warnings (such as a pop-up window). We have selected a hierarchical design as opposed to a flat design. The hierarchy is determined by content with the content more refined and specific as one approaches the bottom layers. Capability for bubbling-up of important information such as announcements/warnings/ news etc., will be provided. Appropriateness of using a context sensitive ticker line will be explored. The user interface would take into account the need to access bookmarked information such as commonly used utilities (e.g. e-mail).

The system design will be based on an object oriented analysis, see Horslen, et al. (1992), and the design approach as in Yourdon et al. (1995). After an information needs assessment, prototype design mock-ups of user

interface will be made in consultation with the core team. Information about hierarchy of access will be determined before designing this stage. A specifications document will be produced at this stage with details on data elements to be displayed, screen access sequence, use of sub-windows, pop-up menus, scrolling vs. paging,

fixed vs. dynamic palettes, etc., Poon, Fagan, and Shortliffe (1995). Part of the specifications document will be acceptance criteria signed off by the project core team. These requirements address specifically what

the software will do when completed. Changes to this document is expected to be made only in the most extenuating of circumstances. Implementation will be using the SUN JAVA technology since this will allow us flexibility in layout changes without scraping entire design at each stage. An iterative approach is most applicable through out the entire software design process. Usability testing will be carried out with core team members and with representative individuals from the set of potential users. This phase would consolidate the design, but possibility of future tune-ups is to be expected.

Maintaining Currency of the Web Server. The stability of all links will be evaluated weekly using a program such as MOM Spider or Verify Link EIT. The list of possible WWW-based cardiac resources will be reevaluated and updated each quarter, using a check program for monitoring link stability and availability. A patient feedback system about new or timely WWW resources will also provide data about needed changes, which will then be made in subsequent patient instructional materials. The Net Computer and modem will allow us to remotely update each patient's home HeartCare system and remotely reconstruct broken links.

A limitation of the WWW approach to home care clinical support is that the integrity of the linkages depends in part on the source of the document, not the state of the person trying to access it. For this reason we will maintain a list of all active links on home net computers, verify the stability of the linkages weekly, and notify users via electronic mail if links change. We will also need to build various capabilities into HeartCare to manage some of the realities of dealing with the WWW and the Internet. For example, we may need to develop a strategy for delayed downloading full motion video, so that a patient does not spend 8-10 minutes waiting for the images to move over telephone lines to a net computer. It may be useful to set up an image management system whereby a patient could opt to have a large file transmitted overnight, and viewed at some later time.

Patients may leave the prescribed set of WWW sites selected by the research team and explore the Internet independently. Any time a patient explores the Web beyond the screened contents of HeartCare, we will provide visual warnings (such as a pop-up window) . An important part of the HeartCare intervention is to provide patients with skills to evaluate the quality, relevance, and applicability of web sites they encounter. Simple announcements, such as disclaimers and recommendations to "Always consult your health care provider" will appear on most HeartCare screens. In addition, based on suggestions from Consumer Health Information Services handbook (Rees, 1991) we will develop a guide for "Evaluation Web Resources" (Appendix E).

HeartCare Component 2: In-Hospital Assessment and Tailoring

The key HeartCare intervention rests on a custom tailored interface (a "home page") that will provide each patient with a unique view of the HeartCare Web Server and other available resources. We developed an in-hospital information needs assessment strategy to identify needs unique to an individual, a tailoring strategy to match WWW-based resources to the individual patient, and construction of individual "home pages" (9 per person), the time-sequenced WWW access pathway for each patient. Appendix F contains two examples from the pilot study.

The project director, who is a nurse and a member of the research team, will manage the interface between the research and the patient's clinical team, will visit each patient on admission and as soon as possible after recovery from anesthesia to initiate the information needs assessment. The in-hospital assessment will include a life-style review, exploration of the topical areas of most interest to the patient, and a review with the patient's primary nurse and physician of any special needs of the patient. The elements for information needs assessment follow the guidance of the Informed Decisions Research Group (Jimison, 1995 and personal communication, May 15, 1995) modified for the needs of the immediate post-surgical recovery period (See Appendix F for the information needs assessment protocol). We envision the in-hospital portion to require no more than 90 minutes, divided into 3-4 encounters according to the patient's needs and preferences. During these encounters the project director will conduct the needs assessment and review options for the HeartCare screens.

The project director will use standard assessment techniques, such as probing questions, and will solicit patient assent to specific components of HeartCare. The project director will explore the patient's understanding of cardiac recovery, probing for thoughts, ideas and fears and integrate the patient's profile with the discharge planning activities coordinated by the patient's primary nurse and primary physician. Ms. Casper is uniquely qualified for this activity, based on her emerging work in the assessment of patient preferences for treatment (see, for example, Casper & Brennan, 1994). Questions such as "Describe for me a typical day you can imagine during your first week at home" will be used to uncover patient expectations. The basic HeartCare program content will be reviewed with the patient, and the patient will be given a list of information content and asked to check those of interest (Appendix F). In the pilot study all patients were able to complete these tasks during a 30 minute visit.

The information needs assessment will be designed to determine a parsimonious set of information requirements specific to an individual patient. These requirements will obtained by the research nurse, summarized and transmitted via e-mail to Drs. Brennan and Moore and the systems manager at CWRU, who will provide a 24 hour response time (we anticipate enrolling 2-3 patients per week for the HearCare group; thus this work load is reasonable). The resources available will be matched to the patient's information requirements using a format suggested by Jillings (1990) which considers individual variables, sociocultural variables, illness-related variables, situation variables, education-related variables, and variables related to cardiac status. The project director will present the details of the patient's individual needs to the research team programmer, who will modify the basic Heart Care Screens to include pointers to information resources specific to the individual's needs (See Appendix F). The research team's programmer will configure the Net computer and prepare it for home delivery. The programmer will in turn develop, by identifying and encoding, all the necessary parameters and WWW links. The entire process, from recruitment to implementation of a particular patient's HeartCare system, must be completed within 3-4 days so that everything is ready to give to the patient before discharge. Next we describe HeartCare in the home.

HeartCare Component 3: In-Home Computer Support

Each HeartCare patient will receive a net computer with a personalized home page before leaving the hospital. The project director will demonstrate the HeartCare tailored system and provide patients with printed instructions on setting the system up at home. Patients will provide a return demonstration, and the project director will answer any questions posed by the patient. Patients will be instructed to take the net computer home with them and establish the connections. Dr. Brennan's experience indicates that persons naive to computers were able to learn computer set up, return home, and connect the system properly. Where necessary the research nurse will delivery and set up the Net Computer during her first visit.

Patients will be encouraged to use HeartCare as often as they desire once they arrive at home. We will provide printed materials with instructions for accessing the system and basic troubleshooting, and we will provide telephone number for emergency system problems. Challenges of the recovery period, including stress and fatigue, require that we provide research team nurse to visit the patient 24 hours after discharge to confirm the set-up of the HeartCare system. To help manage barriers and frustrations with HeartCare use, we will include in the training process (both during the training sessions and in writing with materials sent home with the person) a list of undesirable but not unexpected problems (e.g. slow response of the system during mid-day, unstable links that may lead to closed sites or inaccessible WWW resources, electronic glitches like disconnected plugs or interrupted phone lines). We will build the HeartCare system to minimize these, and attempt to normalize these problems and generate workable solutions to them (eg Use HeartCare in the morning when response time is best.)

A manual will be prepared and distributed to help HeartCare participants get familiar with the computer, know whom to call in case of problems, etc. The project director and research nurse will be trained in the basics of Net Computer usage and diagnostics to distinguish hardware from software (e.g. monitor does not turn on -- is the power on?) A patient encountering problems with the computer as well as those with questions about HeartCare content would first call the project director. Project staff will be accessible to the patient. They will intervene in technical problems only; clinical issues will be referred to the patient's primary provider. A log will be maintained of such calls, statistics generated, and examined by the project team. This will aid in recognizing patterns of problems for redesign.

Utilization of HeartCare will be measured using two strategies. First, there will be a passive monitoring system encoded into the Net Computer to be kept at the patient's home. Dr. Sreenath, the UW programmer and the CWRU systems manager will establish an automatic data upload procedure. Second, at each data collection point, we will explore patients' reasons for accessing HeartCare, and ask for a description of a typical time they used HeartCare, including the presence of others, their reactions to the system and their level of satisfaction.

The role of the Project Director and the Research Nurse The project director will make the initial contact with patients, serve as liaison with the clinical care staff, and monitor all public communication on the HeartCare communication resource. She will enter the system daily to respond to clinical questions and concerns that arise. An important function of the project director will be to transmit all messages about cardiac status appraisal to the clinical practice staff at University Hospitals and to Dr. Geha. The project director will also establish utilities that will allow patients to change the HeartCare interface over the course of the experiment, adding or deleting resources as needed. The research nurse will work under the direction of the project director, and manage in home support, home visits, and locating new WWW resources.

CHIP-The Comparison Intervention

It is plausible that any home care intervention in which patients have pre-discharge contact with a specific nurse and return home with a device and the knowledge that a nurse will be calling them is sufficiently powerful to influence cardiac recovery outcomes. Indeed, Dr. Moore's early work demonstrated that an audio taped intervention was effective in improving physical functioning at 1 month following discharge. To allow us to determine the differential effect of HeartCare, we propose to contrast its use with a less expensive, easier to use intervention - the Cardiac Home Information Program (CHIP). CHIP consists of a 15-minute audio taped message that describes the typical recovery experiences of CABG patients, with particular attention to symptoms they may experience and coping behaviors they may find helpful. Based on the self-regulation theory of Leventhal & Johnson (1983), this intervention describes recovery symptoms and concerns of patients using the words commonly expressed by patients. It includes the cause of the symptom experience if known, and suggestions for managing the recovery problem. The audiotaped recovery information addresses commonly experienced problems that occur during the firs couple of days at home, and the 2nd, 3rd, and 4th weeks of recovery. The audiotaped information will be a revision of earlier versions (Appendix G). The audio tape will include information specific to the early recovery period at home and symptoms unique to women's experiences (Appendix H). The audio tape has recently been produced in a recording studio using a professional female voice.

CHIP will be given to patients on the 4th or 5th postoperative day. Participants will be given both a tape player with headsets and the audio taped program and will be encouraged to listen to it for assistance with recovery problems. They will be told that they may take the audio tape with them at discharge. To assure the dose effect of the intervention, participants will listen to the tape once in the hospital, under the direction of the research nurse. In an attempt to standardize information about cardiac risk factors and recovery, the patients in the CHIP group also will receive the usual cardiac teaching and homegoing instructions provided on this unit to all persons following CABG. Utilization of the CHIP program will be recorded in a use log kept by the patient.

Usual Care

To examine for history and Hawthorne effects, we propose to have a third, non-intervention group whose treatment we label Usual Care. Standard post-surgical care for CABG surgery is discharge between Days 5-10 post-op and follow-up visit with the cardiac surgeon at 6 weeks. The typical discharge teaching protocol consists of one-to-one counseling by the primary nurse, a visit by the dietitian, and viewing a videotape on home recovery. 40 participants will receive Usual Care Only to provide us with the ability to monitor changes in clinical protocols that may occur which may affect interpretation of outcomes. We will contact the 40 participants in this condition twice, at baseline and at the routine 6 week post-surgical visit. We will administer instruments to these participants. This strategy provides non-equivalent control, in that the Usual Care group receives fewer contacts than those proposed for the two intervention groups. It is designed to provide a benchmark of recovery unaffected by experimental conditions or interventions. Furthermore, it avoids the problem of unanticipated intervention evident in our earlier work when data collection contacts by a nurse were perceived by the patient as a very powerful support. This strategy will protect us against the potential confounding that may occur.

Dependent Variables

Physical functioning will be measured using two instruments. The Sickness Impact Profile (SIP) contains eight scales (ambulation, mobility, body care/movement, rest/sleep, eating, recreation/pastime, home maintenance, and work) (Bergner et al., 1988). The SIP assesses the impact of illness on daily activities and behaviors (Appendix I). Participants respond to items describing illness-related behavioral dysfunctions by indicating those items that describe their health on that day. Weighted sums of the items are created for each of the subscales and for a total physical functioning score. Scores can range from 0 to 100. Higher scores indicate greater dysfunction. The validity and reliability of the SIP have been established for measuring functional outcomes at 1 week 1 month, 3 months, six months and one year following CABG (Gortner et al, 1988; Moore, 1992; Moore & Song, under review). The Symptom Inventory (Artinian, 1993), a 20-item instrument, provides a second measure of symptoms specific to CABG recovery (Appendix I). Respondent are asked to indicate on a scale from 1 (not at all) to 7 (always) the frequency of a given symptom during the previous week. The total score is calculated by summing items; the scores range from 20 to 140. Cronbach's Alpha in a sample of cardiac patients at 1, 3, and 6 weeks following CABG was 0.83 (n = 183) (Artinian, 1993).

Psychological distress will be measured using the Profile of Mood States (POMS) (see Appendix I). The POMS (McNair, Lorr, & Droppleman, 1981) assesses mood states indicative of a person's psychological and emotional response to stressors. Adjective checklists constitute six independent mood states: the depression/dejection (16 items), anger/hostility (12 items), tension/anxiety (9 items), confusion/bewilderment (7 items), vigor/activity (8 items) and fatigue/inertia scales (7 items). Participants respond on a 5-point Likert type scale that describes the degree of their feelings for the past week, ranging from 0 (not at all) to 4 (extremely). Scale scores are obtained by summing the responses to the items (with vigor scores reversed). The higher the score, the greater the mood disturbance. In the present study, the fatigue score will not be used to compile the total score for patients at the 1-week and 1 month measurement points because fatigue in the post-CABG period may be due to physical causes, not psychological distress. The possible range of scores is -32 to 168. Test-retest reliability was originally examined on 100 psychiatric patients at a 20-day interval, revealing reliability ranges from .70 to .74. Internal consistencies were calculated on 82 coronary artery bypass patients 1 month after discharge and ranged from .81 to .93 (Moore, 1993). Measures of internal consistency will be calculated for this study. Construct and predictive validity have been demonstrated through significant correlation of scores with the outcomes of short-term psychotherapy, outpatient drug trials and studies of emotion-inducing conditions.

Family Functioning will be measured using the Family Adaptability and Cohesion Evaluation Scale (FACES III)(Olson, Portner, & Lavee, 1985). FACES III measures two major dimensions of family functioning: family cohesion (10 items) and family adaptability (10 adaptability items). Parallel forms of the instrument are used for families with and without children. Scale scores can range from 10 to 50. Scoring for this instrument is addressed in Appendix I. The lower the score, the better the family functioning. Internal consistency, using Cronbach's alpha, was computed in a cardiac surgical sample and reliability estimates of .60 and .72 were obtained for the adaptability and cohesion scores, respectively (Moore, 1993). Test-retest reliability of the FACES, done at 4 to 5 weeks, was .83 for cohesion and .80 for adaptability (Olson, Porter & Lavee, 1985).

Adherence to cardiac risk modification behavior will be measured by the Health Behavior Scale (HBS)(Miller, Wikoff, McMahon, Garrett & Johnson, 1982). The HBS assesses the individual's performance of prescribed actions of the cardiac risk-reduction regimen (Appendix I). The HBS is a 20-item scale to which participants respond on a 5-point Likert scale from "unlikely follows" (1) to "likely follows" (5) the degree of their adherence to diet, medication, physical activity, smoking and stress management recommendations . Cronbach's alphas for the HBS have ranged from .81 to .99 (Miller et al., 1988; Miller et al., 1990) .

Mediating Variables

These variables were selected based on evidence that they mediate the effect of computerized interventions on clinical outcomes (Brennan, Moore & Smyth,1995) or are salient factors in cardiac recovery (Allen,Becker&Swank, 1990).

Service Use will be measured using the 21-item Government Accounting Office's service inventory (see Appendix I) at each point in time (1 week, 1 month, 3 months, 6 months). We will add three additional services to this inventory: emergency room visits, nursing homes, and cardiac rehabilitation services. The data will be collected using a time referent of the number of weeks since the last data collection period (i.e.. 1, 4, 12, and 26). Our experience with this inventory suggests that it is not possible to construct a summative index (as a measure of intensity of service use), but it is possible to count the number of different services used in each period of time (extensiveness of service use). We will add items to precisely assess use of cardiac rehabilitation services.

Co-morbidity will be measured using a weighted sum of secondary diagnoses (co-morbid conditions) that reflects the functional burden of illness conditions (Charlson et al., 1987).(see Appendix I for diagnosis list). IN addition, each participant will be asked if he or she has any illnesses other than heart disease at the 1-, 3- and 6-month data collection points, and these will be added to the co-morbidity list for that subject and new co-morbidity ratings calculated. Validity of the Charlson scale has been supported through correlation of scores with 6-week hospital mortality, length of stay and hospital charges (Deyo, Cherkin & Ciol,1992). Cardiac Functional Status comprises another category of comorbidity and will be measured using the New York Heart Association (NYHA) Classification (The Classification Committee of the New York Heart Association, 1973). A trained rater assigns participants to one of four ordinal cardiac functional classes (I -IV) based on participants' responses to questions about the amount of fatigue, dyspnea or pain they experience at varying levels of physical activity (Appendix I).

Social Support will be measured using the Instrumental-Expressive Social Support Scale (IESS) (Ensel & Woefel,1986)(see Appendix I) This 28-item instrument measures the perceived adequacy of access to and use of social resources for the preservation of one's well being. The IESS has the ability to differentiate between strong and weak social ties and taps both the functions (instrumental) and the emotional (expressive) context of the relations imbedded in a social network. Item scores are summed for the two subscales (Instrumental and Expressive) and the total scale. Psychometrics of the IESS were examined using participants across all age and socioeconomic groups (Ensel & Woefel,1986). Cronbach's alpha coefficients of .89 to .93 were obtained.

The following demographic variables will be collected: age, marital status, number of dependents at home, presence of a caregiver, household income, years of education, and race/ethnicity. Information about surgical procedure, in-hospital complications and length of stay will also be gathered from the Cardiac Surgery Database (Appendix I). We will employ the Short Portable Mental Status (Pfeiffer,1975) questionnaire during all in-hospital visits to ensure that patient sensorium remains intact, and will reschedule visits if necessary.

HeartCare response will be measured via open-ended questions at each data collection point. We will evaluate usability and satisfaction by enumerating the various features and asking whether patient used them, their expectations and their level of satisfaction. We will also use an adjective check list employed in our earlier studies to assess reactions to HeartCare. We will conduct a focus group of participants at various stages.


Development of HeartCare

We will undertake a 1-year development effort with HeartCare. During that time we will seek and evaluate Internet resources, and develop resources as indicated. We will employ our panel of experts (Dunbar, Artinian, Strickland) to propose an enumeration of all possible cardiac recovery themes and content during Month 2 of the project. We will spend Month 3-4 searching for those WWW-resources that fit these themes and content , and evaluating their quality, following a process defined above. Months 3--4 will be spent developing the locally resident programs (including data monitoring) for the home computers and constructing the necessary HeartCare resources not found on the Internet. Each of these components will be tested prior to use.

Second Pilot Study of HeartCare

We will conduct a full pilot test of the basic Home Page prototype for all patients. We will also conduct a 12-week evaluation (month 5-7) of HeartCare with five pilot participants from the University Hospitals (UH) site. The purpose of the pilot will be to test the feasibility and stability of HeartCare, to obtain patients' reactions to the intervention and to the manner in which the intervention is brought to them, and to glean suggestions of missing or irrelevant material. It will allow us to monitor whether patients find actual use of HeartCare frustrating or demanding, and provide insights into making necessary changes. During this time we also propose to demonstrate HeartCare at SCAMC to obtain feedback from peers in the medical informatics community and to demonstrate it to cardiac care clinicians (nurses and physicians from outside the UH system) to gain clinical responses. During this time we will evaluate the integrity of the links in the system using existing link management programs.

Experimental Period

During the clinical trial participants will be enrolled continuously. Each day the hospital-based cardiac nurse will contact the project director and provide her with names of eligible participants who have agreed to talk to the project director and who have been out of ICU at least 12 hours(See letter of support, Appendix A). The project director will go to UH, approach the patient between 10 AM and 8 PM and explain the project. Once the patient consents, the project director will obtain written permission to review the patient record, speak with the patient's nurse and physician, and will schedule an appointment for the next day (in hospital) (15 minutes).

On the second visit (approximately 30 minutes; to be rescheduled if patient sensorium necessitates it), the project director will obtain baseline data and initiate the in-hospital information needs assessment. At the end of this meeting the project director will open an envelop containing the patient's random assignment to one of three conditions, and inform the patient if the patient desires. Usual Care patients will receive only this visit; HeartCare and CHIP patients will receive at least one more. Between the second and third visits the project director will make ready the intervention for the patient. For patients in the HeartCare group, the project director and project programmer will tailor the home page to the patient's information needs.They will also set the clocks and activate the appropriate functions on the computer to be sent home with the patient. For CHIP group members, we will obtain the tape, tape player, headphones and batteries, and deliver them to the patient.

On the third research visit (expected to occur at least 96 hours after leaving the ICU, and to last approximately 35-45 minutes), the project director will present the intervention to the patient. If the subject is assigned to the HeartCare intervention, the project director will show a demonstration of HeartCare to the patient, providing an overview and continue exploration of the patient's information needs. For patients in the comparison group (the CHIP intervention), the project director will explain the tape recording and leave the tape with the patient. The project director will plan one more return visit. Patients will be encouraged to test out their equipment while hospitalized. On the fourth and last research visit (20 minutes), the day of discharge, the project director will deliver equipment and review the procedure. Patients will take their equipment home.

The research nurse will visit the patient 24 hours after discharge, set up the computer if necessary (or check the functioning of the tape) and provide printed telephone resource numbers. This will serve as an intervention check.

The HeartCare server will be active for 24 months to allow all participants a 6 month participation. Gustafson's group found that the benefits of interventions stabilize in 5-6 months (Personal Communication, May 17, 1995)

Termination of the Experiment

The experimental period for each patient will last 6 months. It is important not to withdraw an intervention considered beneficial by the patients. We will, therefore, offer to leave the computers in place indefinitely on loan. We anticipate that 15 to 20 patients may request to keep the computers for 3-6 months at the end of the study period. We will make a final contact by phone to obtain a measure of interest in maintaining the system. For patients who wish to return the HeartCare or CHIP devices, we will go to their home and remove the equipment. (Past experience is that 50% of the participants will want to keep the equipment). After data collection is complete, we will also offer HeartCare access to the CHIP and Usual Care groups to equalize benefits.

Data Collection

Interview data for patients in the HeartCare and CHIP groups will be collected at five points: at baseline (in-hospital), 1 week, 1 month, 3 months, and 6 months. All data following the baseline in-hospital data collection will be done via mail interview surveys. Patients will be sent home from the hospital with their first 1-week interview packet labeled with the date on which the patient is to complete and return the instruments. In an earlier study, Dr. Moore administered these instruments in a non-face-to-face fashion (Moore, 1994) and found that the instruments performed as well as in a face-to-face administration. It is desirable to limit the face-to-face contact during an intervention such as HeartCare because of the disproportionate impact of nursing care perceived by the patient during personal encounters. Participants will be sent interview packets containing the questionnaires and self-addressed, stamped return envelopes at 4 days prior to the 1, 3 and 6-month collection points. Reminder telephone calls will be made to participants who do not return the interview packets within 4 days of the required mailing date. HeartCare utilization data will be collected passively by computer data capture and downloaded via modem call back at each data collection time. Usual care participants will complete the assessment instruments twice, once during the in-hospital period and once at their regularly scheduled 6-week post-surgical visit to the surgeon. The focus group will involve HeartCare participants. All participants will receive a $50 payment.



Week 1


6 Week

(Usual Care Only)



Phys Function









Symptom Management







Psychological Distress









Family Function









Risk Behavior Modification

























Social Support














Use Variables


Intervention Use




HeartCare Response







Focus Group





Monitoring HeartCare Use. There are two levels of HeartCare use that must be monitored: individual user behavior and access to the HeartCare Web Server. We will employ a passive monitor on the local computer that will increment records each time a patient accesses the HeartCare computer at the home. If we do not receive data from a specific patient for more than four weeks (i.e., if the patient has not contacted the HeartCare network), we will contact the person via telephone, explore for problems and troubleshooting, and schedule a specific contact.

We will devise a cache management scheme, similar to UnMozify (c), to allow us, with the participant's permission, to sample trajectories through HeartCare and the WWW that will allow us to characterize the nature of the HeartCare interaction. The Web Server statistics will be monitored using available server monitoring routines; we anticipate being able to determine how often the centralized HeartCare web server was accessed and how access occurred.

Monitoring CHIP Use. Patients will be instructed to keep a log of each time they listen to the CHIP tape (see Appendix G. Dr. Moore has successfully employed use logs in previous studies of patient behaviors (Moore, 1996). Patients will be asked to record the duration of use each time they listened to the tape on a use log card that is attached to the back of the audio tape player. At each data collection point, patients will receive a new log card to place on the tape player and will be instructed to send the log of the previous interval back with the interview packet materials. We will also ask for descriptive information about motivation and reactions to the tape.

Statistical Considerations

The statistical considerations will focus on the question, What does HeartCare contribute to patient outcomes?

Data Management

Data management will be the responsibility of the project director and overseen by Drs. Brennan and Moore. Our previous successful studies have prepared us for managing longitudinal data, including establishing tracking schemes, using color-coding forms anchored to collection dates, and developing and adhering to protocols for coding and cleaning. In our prior ComputerLink projects we developed and implemented schemes for automatic logging of computer network use, and we employed strategies to reduce the large volume of data associated with passive monitoring (see Brennan, 1995; copy in Appendix B). All data obtained during face-to-face or phone interviews will be examined for completeness and consistency, prior to double-coding and double-entry into a database developed in Paradox. Further checking on data quality will be done during the data analyses.

Based on strategies used in the ComputerLink experiments (Brennan, 1995; Brennan and Ripich, 1994), we will construct profiles of HeartCare use patterns and compare these profiles with within-subject trends on major dependent variables. We propose to characterize the independent variable as dichotomous (two levels: HeartCare, CHIP); should our early explorations indicate that a more refined characterization is appropriate we will do so.

For the Usual Care Only group we will use graphical techniques and modelling approaches to examine the usual course of patients over time in the recovery period. We will be able to make comparisons with the HeartCare and CHIP groups at time one to determine the success of the randomization process, and evaluate the baseline-6 week differences in the measures. Using heuristic strategies, we will be able to evaluate whether changes found in the HeartCare and CHIP groups are similar to or different from the Usual Care Only group. This will allow us to construct, if necessary, specific modifications in the analysis strategies used to contrast Heart Care and CHIP.

Data Analysis

The primary analyses will focus on the comparison of HeartCare vs. CHIP patients, who will have been randomly assigned to their treatments after baseline assessment, and then assessed at 1 wk, 1 mo, 3 mo, and 6 mo following discharge. All analyses will be done on an intention-to-treat basis. The primarily outcome measures will be physical functioning (SIP), symptom inventory (SI), psychological distress (POMS), perceived family function (FACES), and adherence to cardiac risk behavior modification (HBS). Mediating variables measured will include cardiac functional status, the general level of co-morbidity, service use, and social support. Power evidence of the comparison between the HeartCare and CHIP groups across time are presented below and in Appendix C. Standard descriptive statistics and modern graphic methods (Cleveland, 1993) will help us summarize and visualize the general characteristics of the data, assess the appropriateness of critical distributional assumptions for inferential tests, suggest appropriate transformations, and identify outliers and assess their potential impact on estimation and inference. One-sided tests will be not be used because such hypotheses are not clearly indicated in this study. Adjusted p-values (Wright, 1992) and similar methods will be used to control overall error rates if multiple tests are required to assess a single general research hypothesis.

Hypotheses Tests

We hypothesize differences between the HeartCare and CHIP intervention groups on the five major outcome variables: physical function, psychological distress, perceived family function, symptom management and adherence to risk behavior modification. Analyzing the primary outcome variables separately gives us a relatively straightforward two-group, longitudinal (repeated measures) situation. Unfortunately, these patients are subject to one-year mortality rates of 5-8%. In addition, drop-outs will occur among survivors, some at random and some probably related to their recovery process. We conjecture that about 10% of the patients will supply some missing data. This informed censoring is a knotty problem in longitudinal data analysis. Diggle, Liang, and Zeger (1994, p. 210) recently concluded that they know of no well- developed method for dealing with informative censoring in longitudinal designs. However, Mark Schluchter, at the Cleveland Clinic, is more optimistic (Schluchter, 1992) and can be easily consulted. We can develop random coefficient models having parameters that correspond one-to-one with our specific hypotheses. These parameters can be estimated and tested using unweighted averages of the least squares estimates obtained from patients' individual scores. This method is unbiased, but lacks efficiency (power) in some cases. What tact we take will depend on the amount and pattern of missingness. We think that meaningful differences will occur as quickly as one week after discharge, when little data will be missing.

Even though we are studying various constructs, patients recovering well should show this over a number of measures. We plan to address this easily yet effectively by adapting O'Brien's (1984) idea, as follows. Within each time, assign ranks to the measures such that a rank of 1 is the "healthiest" score across all patients in the two groups. Tied ranks are not problematic. Then find the average rank for each patient (again, within each time). Missing data will be assigned the worst of the average ranks at that time. We will have an average rank for each patient at each time, which should be amenable to a MANOVA-based repeated measures analysis. Models using the mediating and demographic variables to predict the outcomes will be exploratory, but will still be guided by the following structure. At first, such exploration will NOT include a dummy variable for the treatment group. If such models predict a given outcome measure using mediating and demographic variables, then the Group dummy variable will be forced in and interactions will also be examined.

To assess statistical power, Drs. Brennan and Moore first made conjectures for the expected means for HeartCare and CHIP groups at each time, the groups' common standard deviation at each time, and the "base" correlation (explained below) for the five measures central to the specific aims and major hypotheses. In general, the numbers were estimated from the pilot work reported by Dr. Moore and from reports of other cardiac recovery studies. In particular, the conjectured values for the physical function variable (to be measured by the Sickness Impact Profile) and psychological distress (to be measured by the Profile of Mood States) were obtained from reports of four studies testing the effects of interventions on cardiac recovery outcomes over various points in time using these instruments (Gortner et al., 1988, 1990; Gillis et al., 1993; Moore, 1996).

To counter the clinicians' natural tendency to be too optimistic in estimating treatment effects--especially on behavioral measures -- Dr. O'Brien, the collaborating biostatistician on this project, conservatized the conjectured means for the HeartCare group by moving them closer to the CHIP means. In addition, upper estimates of the standard deviations were employed. Dr. O'Brien adopted a correlational structure which assumed that (1) the unequal spacing of the measurement times nevertheless represented equal spacing of the recovery process; (2) neighboring pairs (no intervening measures) were correlated r; pairs with one intervening measure were correlated r1.5 and all other pairs were correlated r2.

The base correlations, r, are conjectured to be so low that using the baseline measurement in a repeated measures analysis adds more "noise" variation than it offers in terms of variance reduction. In other words, these patients are likely to serve poorly "as their own controls." Accordingly, powers are provided for custom, planned group-by-time effects in that effects are handled through special sets of "contrast variables" (O'Brien & Kaiser, 1985; O'Brien & Muller, 1993). Let the notation

Time[0 1 -1 0 0, 0 0 1 -.5 -.5]

represent a 2 degree-of-freedom contrast comparing the 1 wk, the 1 mo, and the average of the 3 mo and 6 mo measurements. Except for the HBS (not measured at 1 wk and 1 mo), the special group x time tests can be characterized:

(#1) Using baseline & averaging 3 mo & 6 mo measures (3 df):

Group x Time[1 -1 0 0 0, 0 1 -1 0 0, 0 0 1 -.5 -.5]

(#2) Using baseline & averaging 1 mo, 3 mo, & 6 mo measures (2 df):

Group x Time[1 -1 0 0 0, 0 1 -1/3 -1/3 -1/3]

(#3) Ignoring baseline & averaging 3 mo & 6 mo measures (3 df):

Group x Time[0 1 0 0 0, 0 1 -1 0 0, 0 0 1 -.5 -.5]

(#4) Ignoring baseline & averaging 1 mo, 3 mo, & 6 mo measures (2 df):

Group x Time[0 1 0 0 0, 0 1 -1/3 -1/3 -1/3]

These are tighter effects than the traditional (overall) group x time effect, which has 4 degrees of freedom. They were designed because only small Group x Time effects are conjectured at times, so that adding another degree of freedom for the Group x Time effect (Time[0, 0, 0, 1, -1]) reduces power. Such hypothesized weak "subeffects" will still be examined separately as secondary questions. Ignoring the baseline measurement generally increases the power, especially so for measures that are likely to have their largest variance at baseline but little relationship with later measures. Only by ignoring the baseline measure is the power acceptable. We favor form #3 above, which averages only the last two measurements. With 50 in each group, it has .05-based power of .91 for the SIP, .99 for SI, .96 for POMS, and .99 for FACES. The Group x Time test (ignoring baseline) for HBS is conjectured to have a power of .99 (Appendix C). Powers could be a little lower due to the informed censoring; additional power (and improved models) is expected by using baseline measures as covariates for the Time contrast variables (not as part of repeated-measures profile).

The use of O'Brien's (1984) tactic to produce a single summative measure across the 5 endpoints should produce a summary test of the primary hypothesis that has power at least as good as the weakest individual measure. There is no reason to adjust the p-level on this summative measure. When considering the individual measurements (SIP, SI, POMS, FACES, HBS) the p levels can be adjusted simply but somewhat conservatively using the Bonferroni's method, that is, using alpha = .05/5 = .01. In this case, powers for contrast #3 are .76 for the SIP, .99 for SI, .83 for POMS, and .98 for FACES. The Group x Time test for HBS is estimated with power of .99.

TimeLine Yr1 Yr2 Yr3

1 2 3 4 1 2 3 4 1 2 3 4

Recruit & Train Project Staff x

Semi-annual Clinical Advisory Meetings x x x x x x

Construct HeartCare Web Server x x

Establish In-hospital Assessment Scheme x

Pilot Test of HeartCare with 5 participants x

Q1: Nursing Outcomes Monitor and evaluate benchmarks x x x x x x x

Stricklin & Clinical Advisory Panel Mtg x x x

Focus groups with Staff nurses x x x

Q2:Patient Outcomes Enroll first subject x

Ongoing enrollment x x x x x

Open access to all participants x x x

Data Analysis and Reporting x x xxx


It is plausible that introducing new technological innovations during a period of high stress may prove infeasible. Our earlier research indicated that patients would prefer to have available resources presented to them as close to the clinical event as possible, and integrated as fully as possible into their post-operative routine. Therefore, we propose to initiate the intervention during the index hospitalization. We considered enrolling patients during a pre-hospitalization physician office visit. However, because almost 60% of CABG surgeries are conducted under urgent or emergent conditions, our sample would be biased toward persons in sufficiently stable states to allow for a delay between the discovery of a clinically relevant lesion and surgery. If we discover during our pilot study that introducing a HeartCare into the routine cardiac recovery period is infeasible, we will alter our sampling plan and begin recruiting patients from physician office visits. Dr. Geha's support will be invaluable in this decision.

Patients may find use of WWW-based resources frustrating. Novices and experts may experience delays, incomplete error messages, and unmet expectations as annoyances or insurmountable problems. We propose to (1) build a system with high integrity; (2) update broken links discovered through link-checks and other monitoring devices remotely as quickly as possible; (3) set realistic expectations on response time, lack of availability of emergency help, and instability of some web sites and (4) provide participants with back-up phone numbers and troubleshooting. We recognize that the introduction of a new technology into the health care delivery scheme will require a learning period. Our goal is to provide patients with the tools to help integrate the new technology, not to provide patients with a perfect, ever-accessible technology.

The proposed evaluation will be conducted on a relatively small number of participants. The early success of in-home computer resources for health care (Brennan, 1991; Bosworth & Gustafson, 1990) suggests that a larger sample is warranted for more precise testing of hypotheses. The sample these represents a compromise between the feasibility of implementing the study and the precision needed for hypothesis testing. The diffusion of home computers, 30%, is too low to assure that a sample of CABG patients with computers would be unbiased.

Examination of costs and policy implications is important to this project. Through interaction at national meetings and with policy groups Dr. Brennan shares experiences with others involved in community health informatics. She advises policy groups, including the National Research Council, HCFA, and the American Nurses Association aspects of cost and social desirability of consumers direct use of the Internet. This project will provide details of the feasibility of the implementation of nursing informatics interventions, and will serve as a basis upon which to design more explicit examination of costs. At present, the cost of the intervention is artificially high, a result of our attempts to design prototype but clinically relevant computer network applications in anticipation future resources.


The proposed study will involve the participation of 140 individuals who have experienced a first CABG surgery. Following Human Subjects' Approval at University Hospitals of Cleveland and Case Western Reserve University, individuals meeting the following criteria will be recruited into the study: (1) stable and transferred out of Surgical Intensive Care, (2) ability to read, speak and write English, (3) live within a 100-mile radius of Cleveland and (4) have a personal telephone number. A list of patients meeting this inclusion criteria will be identified by the cardiac nurse specialist on the cardiac care unit and referred to the project nurse research assistants via a prepared list. Approximately 11% of the eligible patients are expected to be African American. (Nationally, only 5-6% of CABG patients are African Americans {Ayanian et al., 1993}and in the Greater Cleveland area approximately 10% of the CABG patients are African American.) Oversampling of African Americans will be done by enrolling all of the African American subjects who otherwise meet study criteria. Similarly, women will be oversampled by enrolling all of the women who meet study criteria during (Nationally and locally, women comprise 27% CABG patients {Ayanian et al., 1993}). Sampling of African Americans and women will be continued until they make up 18% and 28%, respectively, (or until 126 Caucasians and 100 men are recruited).

After permission from the medical and nursing staff is obtained, Subjects will be approached by the Project nurse to be in the study. Following a verbal and written description of the study, consenting patients will be asked to sign an informed consent. Following subject consent to participate, the project nurse will obtain written permission to review the patient record, speak with the patient's nurse and physician, and will schedule an appointment for the next day to explain the intervention. Baseline data will be collected in an interview in the hospital; mail surveys will be used to collect data at 1 week, 1, 3 and 6 months.

There are several unique considerations regarding human subjects and technologically based innovations like HeartCare. First, patients must recognize that Net Computer Systems (and HeartCare) do not provide emergency response they need to be able to know what to do when emergency care is needed. We will (1) inform patients on enrollment; (2) place alert messages throughout the HeartCare interface and where possible, within the programs instructing patients that their physician or some emergency service should be contacted for advice on urgent problems such as chest pain and sudden bleeding; (3) repeat these alerts on any paper materials provided to patients; and (4) monitor each patient's experience with our research team, including Dr. Geha and consultants, to maintain the clinical appropriateness of the intervention.

Computer networks, particularly those involving Internet transmission, are subject to intrusion and "snooping". We cannot avoid transmission risks to confidentiality, which for the most part are limited to the instance of copying and later using a particular patient's password. We will identify the risks of transmissions being observed, monitored, or known to others.

Patients may come to rely on HeartCare for monitoring, symptom management and motivational support. We will allow patients who choose to do so to keep the HeartCare home computer on indefinite loan at the end of the project. Additionally, at the end of the experimental phase we will offer access to the HeartCare Net Computers and server to those subjects from other groups to equalize access to what we believe will be a beneficial intervention.

Patients may become frustrated with the mechanical demands of hooking up the net computer. We will provide telephone backup and go out to patients' homes to assist when necessary. We will record each additional visit, to be able to evaluate whether additional attention altered the outcome. Patients may also experiences delays or interruptions in service due to Internet traffic or failure of the Internet Service Provider equipment. We recognize these events as important, 'real world' challenges of using the Internet in the manner we have proposed, but do not wish to unduly stress patients. Therefore, we will, in our training experience, employ a problem-solving approach, and describe common problems and solutions to them (i.e., if ISP is not responding, try in 10 minutes; if still no response, contact research staff). If the ISP proves unreliable, we will change services. We will also document subjects' responses to determine any systematic impact on outcomes. Our experiences in the ComputerLink project, which began in the late 80s, is that naive persons were able to adjust their expectations to the technology, even providing each other hints for handling common problems.

Patient specific data will be only as secure as the technical platform can support. We will build the HeartCare web server using a computer system and server software that permits security and access checks. We will implement several measures to promote the maximum amount of security(e.g. providing each patient with a unique password, requiring password access to enter the HeartCare Web Server, using encryption algorithms and software such as Pretty Good Privacy). We will instruct members of the intervention group not to identify fellow participants, and we will encourage patients to retain the ID and passwords in a secure fashion. If others in the patient's home would like to access the HeartCare system, we will provide them with different ID's and passwords that permit access only to the information sections of HeartCare.

There are no risks from any of these procedures although at times subjects may find them inconvenient or time-consuming. Every effort will be made to prevent this inconvenience. A possible benefit to subjects in this study is an increase in knowledge about cardiac recovery. Subjects will be assured that their participation is voluntary and refusal to participate will not affect any other care or treatment that they receive. Subjects will be informed that they can withdraw from the study at any time with no effects upon the care they receive. Subjects will be asked to keep the use logs and to fill out questionnaires to be returned by mail. There are no costs to subjects for participation in the study.

From a research perspective, confidentiality will be assured in the following manner. While patient names will appear on the addresses of envelopes needed to send materials to the patient, only ID codes, not names will be put on the data collection forms. Data collection forms will be identified by sequential subject research numbers. Results will be reported in aggregate form. An index of subjects' names, medical records numbers, and research identification numbers will be kept in locked project files during the data collection phase. Data will be kept on disks or papers stored in Dr. Moore's research office. No patient names will appear, even on cross-reference lists, in Dr. Brennan's research site in Wisconsin. Only personnel directly related to the project will have access to the data. At the conclusion of the study, the list linking the subject name and medical record number with corresponding sequential number will be destroyed. All data sets and diskettes will be transferred to Dr. Brennan to be stored for a period of seven years.





Drs. Sandra Dunbar and Nancy Artinian, and Ms. Mary Lou Strickland, will serve as consultants for the project. Please see their biosketches and letters of support.

Sandra Dunbar, RN, DNS, is Associate Professor and Coordinator of Critical Care Nursing at the School of Nursing, Emory University and Clinical Director of Critical Care Nursing, Emory University Hospital, Atlanta, GA. Dr. Dunbar has considerable experience in designing and testing patient education programs for cardiac patients, and currently is conducting an NINR-funded study addressing the home adaptation of patients to implantable cardioverter defibrillators. As a member of the Expert Panel, Dr. Dunbar's knowledge of the cardiac patient education process will be important for the expert assessment of the HeartCare resource being developed and tested. Dr. Dunbar will also advise the research team about appropriate clinical considerations for home support of cardiac surgical patients.

Nancy Artinian, RN, PhD, is Associate Professor of Nursing, Wayne State University, College of Nursing, Detroit, MI. Dr. Artinian's research and clinical practice has focused on patient and family adjustment during the first year following CABG surgery. She will provide consultancy about measurement of perception of family functioning during CABG recovery and advisement about patient educational resources.

Mary Lou Strickland, RN, MSN, MBA, is president and chief executive of the Cleveland area Visiting Nurses Association. Ms. Strickland, Dr. Brennan, and Dr. Moore have collaborated on several projects over the past 8 years. Ms. Strickland is also very active in the national VNA. Her participation on the project will be to insure that the home care dimensions of HeartCare meet acceptable standards of practice, and that home care issues are dealt with in an appropriate manner.












Sandra Dunbar, PhD, RN, FAAN

Nancy Artinian, PhD, FAAN

MaryLou Strickland, RN, MSN, MBA




















Ades, P. A., Waldmann, M. L., Polk, D. M., & Coflesky, J. T. (1992). Referral patterns and exercise response in the rehabilitation of female coronary patients aged > 62 years. The American Journal of Cardiology, 69, 1422-1425.

Allen, J.K., Becker, D.M. & Swank, R.T. (1990). Factors related to functional status after coronary artery bypass surgery. Heart & Lung, 19, 337-343.

American Heart Association (1995). 1989 Heart Facts. Dallas, TX: American Heart Association.

Artinian N. T. (1991). Stress experience of patients having coronary artery bypass during hospitalization and 6 weeks after discharg. Hear and Lung, 20(1), 52-59.

Artinian, N.T. (1993). Age differences in patient recovery patterns following cornary artery bypass surgery. Ameican Journal of Critical Care, 2(6):453-460.

Ayanian, J. Z., Udvarhelyi, I. S., Gatsonis, C. A., Pashos, C. L. , & Epstein, A. M. (1993). Racial differences in the use of revascularization procedures after coronary angiography. Journal of American Medical Association, 269, 2642-2646.

Barbarowicz, P., Nelson, M., De Busk, R. F., & Haskell, W. L. (1980). A comparison of in-hospital education approaches for coronary bypass patients. Heart & Lung, 9, 127-132.

Beckie, T. (1989). A supportive-educative telephone program: Impact of knowledge and anxiety after coronary artery bypass graft surgery. Heart & Lung, 18(1), 46-55.

Bergner, M., Bobbitt, R.A., Carter, W.B., Gilson, B.S. (1981). The Sickness Impact Profile: Development and final revision of a health status measure. Medical Care, 19, 787-805.

Bosworth, K., & Gustafson, D. (1991). CHESS: Providing decision support for reducing health risk behavior & improving access to health services. Interfaces, 21, 93-104.

Brennan, P.F., Emmons, H., & Silvers, J. (1991). Estimating longevity of hospital nurses. Journal of the Society of Health Systems, 3(2), 103-111.

Brennan, P.F. & Moore, S.M. (1994). Networks for home care support: The ComputerLink project. Caring, 13(8), 64-70.

Brennan, P.F., & Ripich, S. (1994). Use of a home-care computer network by persons with AIDS. International Journal of Technology Assessment in Health Care, 10(2): 258-272.

Brennan, P.F., Emmons, H., Taft, S., Zurakowski, T., Silvers, J.B., Kahana, E., Jones, P., & Bolek, B. (1991). Strengthening Hospital Nursing: Dimensions of Success. Journal of Society of Health Systems, 3(2), 5-21.

Brennan, P.F., Moore, S.M. & Smyth, K.A. (1991). ComputerLink: Electronic support for the home caregiver. Advances in Nursing Science, 13(4), 14-27.

Brennan, P.F., Moore, S.M. & Smyth, K.A. (1992). Alzheimer's disease caregivers' uses of a computer network. Western Journal of Nursing Research, 14(5),662-673.

Brennan,P.F., Moore,S.M. & Smyth,K.A. (1995). The effects of a special computer network on AD caregivers. Nursing Research, 44(3), 166-172.

Chang, B.L., & Hirsch, M. (1994, Jan. - March). Nursing diagnosis research: Computer-aided research in nursing. Nursing Diagnosis, 5(1), 6-13.

Charlson, ME, Pompei, P., Ales, KL. & Mackenzie, CR. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Disability, 40, 373-383.

Clark, K., & Ellis, L. (1982). Balance: An interactive program for the design of balance study diets. In B. Blum, ED. SCAMC Proceedings. Los Angeles, CA: IEEE Computer Society.

Cleveland WS (1984) Visualizing data. Summit, NJ: Hobart Press.

Cosmoss, P.M. (1988). Nursing strategies to improve compliance with life-style changes in a cardiac rehabilitation population. Journal of Cardiovascular Nursing, 2(3):23-36.

D'Alessandro, M.P., Galvin, J.R., Erkonen, W.E., Lacey, D.L., Curry, D.S., Holtum, E.W., & Santer, D. (1994). The Virtual Hospital: Creating and organizing an ubiquitous health sciences organzation on the Internet. JAMIA, Supplement 1, 1061.

Deyo, R.A., Cherkin, D.C. & Ciol, M.A. (1992). Adopting a clinical morbidity index for use with ICD-9-CM administrative databases. Journal of Clinical Epidemiology, 45(6), 613-619.

Diggle, PJ, Liang, KY & Zeger, SL. (1994). Analysis of Longitudinal Data. Oxford, UK: Claredon Press.

Dillman, D.A. (1978). Mail and Telephone Surveys: The Total Design Method. New York:John Wiley & Sons.

Dracup, K.A. (1982). The effect of a role supplementation program for cardiac patients and spouses on mastery of the at-risk role. Dissertation. University of California at San Francisco. University Microfilms International, Ann Arbor, MI.

Ellis, L.B.M., Huan, P.C.K. & Buzzard, I.M. (1986). Touchscreen versus keyboard for menu-based food coding. In. R. Salamon, B. Blum, M. Jorgensen (Eds.). MEDINFO 86 (Part II), 999-1003. North Holland: Elsevier Science Publishers.

Ensel, W., & Woelfel, M. (1986). Measuring the instrumental and expressive functions of social support. In N. Lin, A. Dean & W. Ensel (Eds). Social Support, Life Events and Depression. Orlando, FL: Academic Press, Inc.

Eyherabide, A. & Yates, C. (1985). The effects of cardiac rehabilitation on compliance in the coronary artery bypass surgery patient. Cardiovascular Nursing, 21, 31-35.

Finkelmeir, B.A. (1995). Cardiothracic Surgical Nursing. J.B. Lippincott Company: Philadelphia.

Flynn M.K. & Frantz R. (1987). Coronary artery bypass surgery:quality of life during early convalescence. Heart Lung 16,159-167.

Genet, C., Brennan, P.F., Ibbotson-Wolff, S., Phelps, C., Rosenthal, G., Landefeld, C.S., & Daly, B. (1995). Nurse pracitioners in a teaching hosptial setting. Nurse Practitioner, November.

Gillis, C.L. (1983). Identification of factors contributing to family functioning following coronary artery bypass surgery. Dissertation, University of California at San Francisco. University Microfilms International, Ann Arbor, MI.

Gillis, C.L., Gortner, S.R., Shinn, J.A., & Tompkins, C. (1993). A randomized clinical trial of nursing care for recovery from cardiac surgery. Heart and Lung, 22(2),125-133.

Gortner, S.R., Gillis, C.L., Shinn, J.A., Sparacino, P.A., Rankin, S., Leavitt, M., Price, M., & Hudes, M. (1988). Improving recovery following cardiac surgery: a randomized clinical trial. Journal of Advanced Nursing, 13, 649-661.

Gortner, S.R., & Jenkins, L.S. (1990). Self-efficacy and activity level following cardiac surgery. Journal of Advanced Nursing, 15, 1132-1138.

Graves. J.R., & Corcoran, S. (1989). The study of nursing informatics. Image-the Journal of Nursing Scholarship, Winter, 21(4), 227-31.

Grobe, S. (1990). Nursing intervention lexicon and taxonomy study: Language and classification methods. Advances in Nursing Science, 13(2), 22-33.

Gumpertz , M., &, Pantula, SG (1989) A simple approach to inference in random coefficient models. American Statistician, 43, 203-210

Gustafson, D.H., Bosworth, K., Chewning, B.J., & Hawkins, R. (1987). Computer-based health promotion: Combining technological advance & problem solving technique to effect successful health behavior changes. Annual Review of Public Health, 8, 387-415.

Gustafson, D.H., Bosworth,K., Hawkins, R.P., Boberg, E.W., & Bricker, E. (1992). CHESS: A computer-based system for providing information, referrals, decision support and social support to people facing medical and other health-related crises. Proceedings of the Sixteenth Annual Symposium on Computer Applications in Medical Care: 161-5.

Gustafson, D.H., Hawkins, R.P., Boberg, E.W., Bricker, E., Pingree, S., & Chan, C.L. (1994). The use and impact of a computer-based support system for people living with AIDS and HIV infection. Proceedings of the Eighteenth Annual Symposium on Computer Applications in Medical Care: 604-8.

Hanish, P. (1993). Informational needs and preferred time to receive information for phase II cardiac rehabilitation patients: What CE instructors need to know. Journal of Continuing Education in Nursing, 24(2), 82-89.

Hanushek, E.A., & Jackson, J.E. (1977). Statistical methods for social scientists. New York: Academic Press.

Henry, S.B., Holzemer, W.L., Reilly, C.A., & Campbell, K.G. (1994). Terms used by nurses to describe patient problems: Can SNOMED III represent nursing concepts for the patient record? Journal of the American Medical Informatics Association, 1, 61-74.

Jenkins, C.D., Stanton, B.A. Savageau, J.A., Denlinger, P. & Klein, M.D. (1983). Coronary artery bypass surgery. Journal of American Medical Association, 12(6), 499-504.

Jillings, C.R. (1990). Cardiac Rehabilitation Nursing, pp. 189-193, Aspen Publication.

Kasper, J.F., Mulley, A.G., & Wennberg, J.E. (1992). Developing shared decision amking programs to improve the quality of health care. QRB, 18(6): 183-190.

King, K.B. (1985). Measurement of coping strategies, concerns, and emotional responses in patients undergoing coronary artery bypass grafting. Heart & Lung, 14(6), 579-586.

King, K.B., & Parrinello, K. A. (1988). Patient perceptions of recovery from coronary artery bypass grafting after discharge from the hospital. Heart & Lung,17(6),708-715.

Kjerulff, K.H. & Moray, L.R. (1988). Measuring the impacts of a hopsital information system on the nursing process. In Daly, N.,& Hannah, KJ (EDs). Proceedings of Nursing and Computers Third International Symposium on Nursing Use of Computers and Infomration Science. St. Louis, MO: CV Mosby.

LeDoux, D. & Shinn, J.A. (1995). Cardiac surgery. In Cardiac Nursing, S. L. Woods, E.S. Sivarajan Froelicher, C. J. Halpenny & S.Underhill Motzer (eds). J.B. Lippincott Company: Philadelphia.

Leventhal, H. & Johnson, J.E. (1983). Laboratory and field experimentation development of a theory of self-regulation. Dr. M. Schmitt, J. Skipper & R. Leonard (Eds.), Behavioral Science and Nursing Theory (pp.189-262). St. Louise: C.U. Mosby Co.

Litwack. E. (1985). Helping the elderly. The complementary roles of informal and formal systems. New York: Guilford.

Marshal, J. Peckofer, S., & Llewellyn, J. (1986). Structured postoperative teaching and knowledge and compliance of patients who had coronary artery bypass surgery. Heart & Lung, 15(1), 76-81.

McClendon, M.J., Bass, D.M., & Brennan, P.F. (1995). The effect of a computer network for Alzheimer's Disease caregivers on formal service use. Paper to be presented at the Gerontological Society of America.

McCloskey, J.C., & Bulecheck, G.M., (1994, March-April). Standardizing the language for nursing treatments: an overview of issues. Nursing Outlook, 42(2), 56-63.

McNair, D.M., Lorr, M. & Droppleman, L.F. (1981). Manual of Profile of Mood States, San Diego: Educational and Industrial Testing Service.

Miller, P., Wikoff, R., McMahon, M., Garrett, M.J. & Johnson, N. (1982). Development of a health attitude scale. Nursing Research, 31(3):132-136.

Miller, S.P., Wikoff, R., McMahon, M. Garrett, M.J., Ringel, K. (1990). Marital functioning after cardiac surgery. Heart and Lung, 19(1), 55-61.

Moore, SM and Song, R. (Under Review) Quality of life of individuals one year following coronary artery bypass surgery.

Moore, S.M. (1993). Effect of an information intervention on recovery outcomes of patients and spouses following coronary artery bypass surgery. Doctoral dissertation. Case Western Reserve University, Cleveland.

Moore, S.M. (1994). Development of a discharge information program for recovery following coronary artery bypass surgery. Applied Nursing Research. 7(3).

Moore, S.M. (In Press). Gender differences in symptoms during home recovery following coronary artery surgery. Heart & Lung.

Moore, S.M. (In press). Women's views of cardiac rehabilitation programs. Journal of Cardiopulmonary Rehabilitation.

National Cardiac Surgical Database. (1993). The Society of Thracic Surgeons

NINR (1993). Nursing Informatics: Enhancing Patient Care. A report of the priority expert panel on Nursing Informatics. Bethesda, MD.

Norris, J., Cuddigan, J., Foyt, M., & Leak, G. (1989). Decision support and outcomes of nurses' care planning. Proceedings of the Thirteenth Annual Computer Applications in Medical Care: 827-833.

O'Brien, P. (1984). Procedures for comparing samples with multiple endpoints. Biometrics, 40, 1079-1987.

O'Brien, R.G. (1986). Using the SAS system to perform power analyses for log-linear models. Pp. 778-784 in Proceedings of the Eleventh Annual SAS Users Group Conference. Cary, NC: SAS Institute, Inc.

Oka, R.K., Froelicher, E.S.S. (1995). Health promotion and behavioral interventions. In Cardiac Nursing, S. L. Woods, E.S. Sivarajan Froelicher, C. J. Halpenny & S.Underhill Motzer (eds). J.B. Lippincott Company: Philadelphia, PA.

Olson, D.H., Portner, L. & Lavee. (1985). FACES III. Family Social Science, St. Paul Minnesota, University of Minnesota.

Pashkow, F.J. (1993). Issues in contemporary cardiac rehabilitation: A historical perspective. Journal of the American College of Cardiology, 21, 822-834.

Payton, F.C., Brennan, P.F., & Silvers, J.B. (1995). Cost justification of a community health information network: The ComputerLink for AD caregivers. Proceedings of the ninteenth annual symposium on computer applications in medical care.

Penckofer, S. & Llewellyn, J. (1989). Adherence to risk-factor instructions one year following coronary artery byapss surgery. Journal of Cardiovascular Nursing, 3(3):10-24.

Petrucci, K., Jacox, A., McCormick, K., Parks, P., Kjerulff, K., Baldwin, B., & Petrucci, P. (1992). Evaluating the appropriate of a Nurse expert system's patient assessment. Computers in Nursing, 10(6), 243-249.

Pfeiffer, E. (1975). A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. Journal of the American Geriatrics Society, 23(10):433-41.

Pingree, S, Hawkins, RP, Gustafson, DH, Boberg, EW, Bricker, E., Wise, M., Tillotson, T. (1994). Will HIV-Positive people use an interactive computer simulation for information and support? A study of CHESS in two communities. Journal of the American Medical Informatics Association, Supplement 1, p22-26.

Rankin, SH & Stallings, KD (1990). Patient Education: Issues, Principles, Practice. Phila: Lippincott.

Ripich, S., Moore, S.M. & Brennan,P.F. (1992). A new nursing Medium: Computer networks for group intervention. Journal of Psychosocial Nursing, 30(7),15-20.

Rogers, E. (1983). Diffusion of Innovations. New York: Free Press.

Ruzicki, D.A. (1989). Realisticaly meeting the educational needs of hospitalized acute and short-stay patients. Nursing Clinics of North America, 24(3):639-637.

Safran C., Rind, D.M., Currier, J., Ives, D., Sands, D.Z., Slack, W.V., Makandon, H., & Cotton, D. (1993). An electronic medical record that helps care patients with HIV infection. Proceedings of the Seventeenth Annual Symposium for Computer Application in Medical Care: 224-228.

Schluchter, MO. (1992). Methods for the analysis of informatively censored longitudinal data. Statistics in Medicine, 11, 1861-1870. (Disc: 1881-1895).

Shinn, J.A. (1994). Rapid recovery form CABG surgery. Progress in Cardiovascular Surgery, 9(3): 47-48.

Slack W.V., Safran, C., Kowaloff, H.B., Pearce, J., & Delbanco, T.L. (1993). Be well! a computer-based health vare interview for hospital personnel. Proceedings of the Seventeenth Annual Symposium for Computer Application in Medical Care: 12-16.

Slack, W.V. (1984). A history of computerized medica interviews. M.D. Computing, 1, 52-59.

Sparks, K.E., Shaw, D.K., Eddy, D., Hanigosky, P. & Vantrese, J. (1993). Alternatives for cardiac rehabilitation patients unable to return to a hospital-based program. Heart & Lung, 22, 298-303.

Staggers, N., & Mills, ME. (1994). Nurse-computer interaction and stgaff perfomance outcome. Nursing Research, 43(3), 144-150.

Stanley, M.J., & Frantz, R.A.(1988). Adjustment problems of spouses of patients undergoing coronary artery bypass graft sugery during convalescence. Heart and Lung, 17(6), 677-682.

Stanton, B.A, Jenkins, C.D., Savageau, J.A. & Thurer, R.L. (1984). Functional benefits following coronary artery bypass graft surgery. Annals of Thoracic Surgery, 37(4):286-290.

Steele, J.M. & Ruzicki, D. (1987). An evaluation of tahe effectiveness of caridac teaching during hospitalization. Heart & Lung, 16(5):306-317.

Tack, B.B. & Gilliss, C.L. (1990). Nurse-monitored cardiac recovery: A description of the first 8 weeks. Heart & Lung, 19(5), 491-499.

The Criteria Committee of the New York Heart Association. (1973). Nomenclature and Criteria for Diagnosis of Diseases of the Heart and Great Vessels. Boston: Little Brown and Company.

Waitkoff, B., & Imburgia, D. (1990). Patient education and continuous improvement in a phase 1 cardiac rehabilitation program. Journal of Nursing Quality Assurance, 5(1), 38-48.

Wald, J.S., Rind, D, Safran, C.S., Kowaloff, H., Barker, R, & Slack, W. (1995). Patient entries in the electronic medical record: An interactive interview used in primary care. Journal of the American Medical Informatics Association, Supplement, 2 .

Whitman, G. (1995). Fast-tracking cardiac surgical patients. Paper presented at the Fifteenth Annual Dimensions in Cariac Critical Care Nursing Symposium, April 10, 1995, Cleveland, OH.

Woolery, L., Grzymala-Busse, J., Summers, S., & Budihardjo, A. (1991). The use of machine learning program LERS-LB 2.5 in knowledge acquisition for expert system development. Computers in Nursing, 9(6), 227-34.

Wright, S.P. (1992). Adjusted P-values for simultaneous inference. Biometrics, 48, 1005-1013.

Youngblut, J.M., Brennan, P.F., & Swegart, L.A. (1994). Families with medically fragile children: An exploratory study. Pediatric Nursing, 20(5): 463-8.

Zielstorff, R.D. (1995). Capturing and using clinical outcome data: Implications for information systems design. Journal of the American Medical Informatics Association, 2(3): 191-196.

Zielstorff, R.D., Barnett, G.O., Fitzmaurice, J.B., Oliver, D. E., Ford-Carleton, P., Thompson, B.T., Estey, G., Eccles, R., Martin, M., & Jenders, R. (1993). Troubleshooting pulmonary artery catheter waveforms. Proceedings of the Seventeenth Annual Symposium on Computer Applications in Medical Care: 351-355.

Zurakowski, T., Brennan, P.F., & Kahana, E. (Under Review). Elements of professional practice models in nursing as viewed by nursing service innovators.

















Appendix C:

Power Tables



Note: This Appendix is completely revised, reflecting the incraased sample sixe.


Scenarios and results for power analyses:


Measure: SIP (physical functioning)

Time 0 1 wk 1 mo 3 mo 6 mo


HeartCare 26 12 6 5 5

CHIP 26 15 9 7 7


Common StdDevs 10 7 5 4 4



1.00 0.32 0.18 0.10 0.10

0.32 1.00 0.32 0.18 0.10

0.18 0.32 1.00 0.32 0.18

0.10 0.18 0.32 1.00 0.32

0.10 0.10 0.18 0.32 1.00

n = 50 per group

Custom, Planned Power for Group x Time Effect

a = 0.05 a= 0.05/5

Using baseline & averaging 3 & 6 months (3 df): .268 0.105

Using baseline & averaging 1, 3, & 6 months (2 df): .235 0.088

Ignoring baseline & averaging 3 & 6 months (3 df): .910 0.760

Ignoring baseline & averaging 1, 3, & 6 months (2 df): .941 0.822


Measure: SI (symptom inventory)

Time 0 1 wk 1 mo 3 mo 6 mo


HeartCare 45 38 30 26 24

CHIP 45 42 35 30 27


Common StdDevs 8 7 6 5 5


1.00 0.27 0.14 0.07 0.07

0.27 1.00 0.27 0.14 0.07

0.14 0.27 1.00 0.27 0.14

0.07 0.14 0.27 1.00 0.27

0.07 0.07 0.14 0.27 1.00

n = 50 per group


Custom, Planned Power for Group x Time Effect

a = 0.05 a= 0.05/5

Using baseline & averaging 3 & 6 months (3 df): .615 .367

Using baseline & averaging 1, 3, & 6 months (2 df): .598 .352

Ignoring baseline & averaging 3 & 6 months (3 df): .998 .990

Ignoring baseline & averaging 1, 3, & 6 months (2 df): .999 .994

Measure: POMS (psychological distress)

Time 0 1 wk 1 mo 3 mo 6 mo


HeartCare 18 19 6 4 4

CHIP 18 22 8 6 6


Common StdDevs 6 7 4 3 3



1.00 0.38 0.23 0.14 0.14

0.38 1.00 0.27 0.23 0.14

0.23 0.38 1.00 0.38 0.23

0.14 0.23 0.38 1.00 0.38

0.14 0.14 0.23 0.38 1.00

n = 50 per group

Custom, Planned Power for Group x Time Effect

a = 0.05 a= 0.05/5

Using baseline & averaging 3 & 6 months (3 df): .408 .194

Using baseline & averaging 1, 3, & 6 months (2 df): .470 .240

Ignoring baseline & averaging 3 & 6 months (3 df): .954 .825

Ignoring baseline & averaging 1, 3, & 6 months (2 df): .959 .866



Measure: FACES (perceived family function)

Time 0 1 wk 1 mo 3 mo 6 mo


HeartCare 11 11 16 18 18

CHIP 11 8 11 12 12


Common StdDevs 5 5 7 8 8



1.00 0.35 0.21 0.12 0.12

0.35 1.00 0.27 0.21 0.12

0.21 0.35 1.00 0.35 0.21

0.12 0.21 0.35 1.00 0.35

0.12 0.12 0.21 0.35 1.00

n = 50 per group


Custom, Planned Power for Group x Time Effect

a = 0.05 a= 0.05/5

Using baseline & averaging 3 & 6 months (3 df): .952 .850

Using baseline & averaging 1, 3, & 6 months (2 df): .971 .897

Ignoring baseline & averaging 3 & 6 months (3 df): .996 .976

Ignoring baseline & averaging 1, 3, & 6 months (2 df): .998 .987

Measure: HBS (adherence to cardiac risk behavior modification)

Time 0 1 wk 1 mo 3 mo 6 mo


HeartCare 8 X X 18 18

CHIP 8 X X 16 15


Common StdDevs 2 X X 3 3



1.00 X X 0.15 0.06



0.15 X X 1.00 0.15

0.06 X X 0.15 1.00

n = 50 per group


Custom, Planned Power for Group x Time Effect

a = 0.05 a= 0.05/5

Using baseline (2 df): .986 .940

Ignoring baseline (2 df): .999 .995