The relationship of cardiovascular and psychological impairments to the health status of patients enrolled in cardiac rehabilitation programs.

The World Health Organization (WHO) defines health as "a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity."(1) This definition suggests a model for conceptualizing health and health care outcomes. This model provides a framework for examining health-related problems, not only in terms of disease (pathology at the cellular level) and impairment (loss or abnormality at the organ-system level) but in terms of health-related behaviors at the level of the individual (functional limitation) and expectations at the societal level disability).(2,3) Health status is a term commonly used for the behavioral aspects of health. The categories of behaviors affected by health are generally agreed to be physical function, social function, emotional function, self-perception of health, and well-being. Categories are sometimes added to examine domains such as pain level, cognitive function, and social opportunities.

Measures of health status can be specific or generic. Disease-specific instruments contain items concerning the degree of limitation by symptoms directly related to the disease of interest. Generic instruments contain items that ask about limitations due to health in general, rather than about specific symptoms, and are useful across various conditions and populations. Criteria for including items in measuring health status are that they represent (1) aspects of life affected by the presence or absence of disease in general, or a specific disease, and (2) concepts to which individuals and society attach importance in contributing to the quality of life.(4) Comprehensive assessment of the array of behaviors affected by health is necessary for the valid definition and measurement of health status.

As the population ages, a higher prevalence of chronic diseases is likely, accompanied by higher health care costs for the treatment of persons with chronic diseases. For individuals with chronic diseases, decisions to use expensive health care resources and their providers' decisions to use certain tests and interventions are made many times over the course of the disease. These decisions may be driven by the effect or potential effect of disease on patients' function and their perception of their state of health, rather than on laboratory measures of impairment or disease.

Awareness of health status levels associated with particular pathologies, stages of disease, and impairment levels can assist health care planners and providers in anticipating societal health care needs. Knowledge of these relationships might help identify those groups of patients requiring the greatest degree of care or those most likely to experience disability. Moreover, determination of factors related to health status and knowledge of the relative influence of these factors may improve the selection of interventions for addressing low levels of health status among persons with chronic disease.

Few studies have examined the cross-sectional relationship of a comprehensive, generic measure of health status to level of impairment or to demographic characteristics in patients with cardiovascular disease.(5-8) Findings based on the Sickness Impact Profile  SIP),(5-7) a well-established, psychometrically tested, generic health status instrument, have been reported. The generalizability of these findings, however, is limited because of the use of univariate analyses and the restriction of the samples to very specific disease categories. Whereas multivariate analyses may account for the potential effects of intervening factors on the relationship between two variables, univariate analyses do not. In addition, one might question the usefulness of the SIP in individuals with cardiovascular disease due to skewed scores, floor effects, and the length of time required to complete the instrument. Neill and colleagues(8) also studied the relationship of health status to results from exercise tests. Although multivariate analyses were performed, the findings were limited as only men were included in the study. In addition, the instrument used to assess health status was derived by the investigators, and its validity was not described.

To our knowledge, no study has examined the multivariate relationship between demographic characteristics and impairments in a large sample of both men and women with cardiovascular disease using a well-validated, comprehensive, generic measure of health status. The purpose of our study was to examine the relationship between health status as measured by the Medical Outcomes Study (MOS) 36-item Short-form Health Survey (SF-36)(9) and the demographic characteristics and physiological and psychological impairments commonly associated with cardiovascular disease in a group of subjects enrolled in cardiac rehabilitation programs.

Method

Subjects Personnel at 17 hospital-based cardiac rehabilitation programs in the state of Massachusetts were asked to participate in a multicenter data-collection effort. Representatives of 13 of those sites agreed to participate. These programs were located in rural settings (n = 3), suburban settings (n = 5), and urban settings (n = 5). Data were collected in a standardized manner on consecutive patients enrolled in these cardiac rehabilitation programs. Data collection took place over a 13-month period between January 1990 and February 1991.

Patients were accepted to these programs based on referral from their physicians. Referral criteria vary somewhat from physician to physician; however, eligibility and exclusion criteria for all cardiac rehabilitation programs are consistently based on recommendations of the American College of Sports Medicine.(10)

All patients entering the 13 programs were given information about the study and asked for their informed consent. Information was collected by personnel at each site at initiation of the program and included patient demographics and information concerning physiological impairments. These data were obtained from the referring physician, an interview conducted by licensed program personnel, a physical examination, and laboratory test reports. A psychological profile and a health status profile were obtained through self-administered questionnaires. Data were obtained concerning each patient's gender, age, years of education, and living situation.

Seven hundred eighty-nine patients initially agreed to participate. Complete data for all baseline independent variables of interest were available on 262 subjects. The greatest loss of data was due to failure to date the psychological profile correctly, preventing the investigator from determining baseline data from follow-up data. Table 1 provides data concerning the characteristics, physiological impairments, psychological impairments, and health status for subjects with complete data on all variables of interest at entry.

Subjects completing data collection were essentially the same age as all subjects entering the study and had similar years of education. There were slightly more men in the sample with complete data than in the total sample, and slightly fewer subjects were living alone. More subjects in the limited sample had a diagnosis of myocardial infarction (MI) or coronary artery bypass graft (CABG) than did subjects in the larger sample, and more subjects had a history of hypertension. Fewer subjects in the limited sample smoked than in the larger sample. Subjects with complete data tended to have higher (better) scores on the SF-36, but total psychological distress scores were the same. The scores for the SF-36 were most notably higher in the energy/fatigue, general health perceptions, and mental health scales. Fewer of the subjects with complete data, however, had "high" scores in physical role limitation.

Data Collection

Measures of physiological impairment. We considered measures of physiological impairment to be the event or procedure precipitating referral, the presence of one of the major cardiovascular risk factors, the presence of either concomitant lung disease or diabetes, evidence of left ventricular hypertrophy (LVH), and peak oxygen consumption ([VO.sup.2]) obtained from exercise testing. Presence of hypercholesterolemia was defined as a total cholesterol level of 240 mg/dL or above. A history of hypertension was determined from the interview and referral data. The presence of comorbidity was determined from the interview, referral information, and medication data. Information concerning conditions other than lung disease or diabetes was not collected. The presence of LVH was determined through evidence provided by electrocardiography and ventriculography during exercise tolerance testing. Events or procedures precipitating referral for cardiac rehabilitation were determined from referral information and classified as recent MI, CABG, angioplasty, new onset of angina, or other reasons.

Symptom-limited exercise testing was performed under physician super-vision prior to patients entering rehabilitation programs. Tests were performed using a variety of standardized treadmill or cycle ergometer protocols.(10) Twelve-lead electrocardiograms (ECGs) and blood pressure were monitored continuously during exercise testing. Endpoints for the tests were based on symptoms, including patient reports of fatigue, angina, or changes in the ECG, that suggested ischemia. Exercise duration, in seconds, and peak exercise work load, in terms of speed and incline (on the treadmill) or watts (on the bicycle), were recorded and used to calculate peak [VO.sub.2] (in milliliters per kilogram per minute). Although [VO.sub.2] estimates for walking on a treadmill determined from speed and incline are fairly accurate for adult men and women, less accurate estimates are obtained when walking on a level surface than when walking up a grade.(10) Oxygen consumption is underestimated by 15% to 20% when walking on a level surface and by 5% to 8% when walking up a grade.(10) Rail holding increases the error in estimating. We therefore used a regression formula designed to estimate [VO.sub.2] when rail holding during the Bruce protocol.(11) We were unable to obtain similarly corrected calculations for other protocols used by our patients. For subjects completing other treadmill protocols or bicycle testing, [VO.sub.2] was calculated using standard formulas.(10)

Table 1.
Subject Characteristics (n = 262)
Age (y)
  X                                                     60.4
  SD                                                    10.0
  Range
35-88

Gender
  Male                                                   77%
  Female                                                 23%
Education  (y)
  X                                                     13.4
  SD                                                     2.9

  Range                                                 7-23

Living situation
  Alone                                                   20%
  With partner                                            80%

Event/procedure leading to referral
  Myocardial infarction                                   51%
  Status after coronary artery bypass graft               29%
  Status after angioplasty                                 7%
  New angina                                              10%
  Other                                                    3%

Comorbidity
  Yes                                                      20%
  No                                                       80%

Peak oxygen consumption (ml/kg/min)
  X                                                       22.7
  SD                                                       6.6
  Range                                               8.0-39.9

Left ventricular hypertrophy
  Yes                                                     13%
  No                                                      87%

Hypercholesterolemia
  Yes                                                     23%
  No                                                       77%

Hypertension
  Yes                                                     62%
  No                                                      62%

Smoking
  Yes                                                      3%
  No                                                      97%

Total Symptom Questionnaire score
  X                                                      19.5
  SD                                                       15
  Range                                                  1-84

Energy/fatique score
  X                                                      53.5
  SD                                                     20.5
  Range                                                  0-95

General health perceptions score
  X                                                      64.2
  SD                                                     18.3
  Range                                                10-100

Mental health score
  X                                                     72.4
  SD                                                    19.7
  Range                                                0-100

Physical functioning score
  X                                                     64.4
  SD                                                    24.4
  Range                                                5-100

Social functioning score
  X                                                     69.5
  SD                                                    26.7
  Range                                                0-100

Bodily pain score
  X                                                     71.3
  SD                                                    24.4
  Range                                                0-100

Role limitation
  Emotional score
     High                                              56%
     Low                                               44%
  Physical score
     High                                              13%
     Low                                               87%

Health status assessment. Assessment of health status was accomplished using the SF-36. The SF-36 is a generic instrument, measuring eight health concepts: physical functioning (10 items), social functioning (2 items), role limitations due to physical problems (4 items), role limitations due to emotional problems (3 items), mental health (5 items), energy/fatigue (4 items), bodily pain (2 items), and general health perception (5 items). A single item evaluates the patient's perception of change in health over the past year. Each of the eight health concepts is measured on a scale of 0 to 100, as a percentage of the total possible score for that scale. A higher score indicates better health in all categories. The survey can be completed in approximately 10 minutes. At the time of initiation of this study, the standard form of the SF-36 obtained from its developers contained questions about the patient's perception of health in the previous 4 weeks. Subjects in our study, therefore, were asked how they felt in the last 4 weeks for both the initial and follow-up evaluations.

The usefulness, validity, and reliability of the SF-36 have been reported.(12-14) In our study, intraclass correlation coefficients (ICC[2,1]) for test-retest reliability of the SF-36 for a small subsample (n=30) were greater than or equal to .70 for all scales except the scale for emotional role limitation (ICC=.37).

Psychological profile. A psychological profile of each subject was obtained using the Symptom Questionnaire (SQ) developed by Kellner.(15) The questionnaire was self-administered. It includes scales for depression, anxiety, anger-hostility, and somatic symptoms. Scores for each symptom scale range from 0 to 23, with higher scores indicating more distress. A total distress score is also obtained from the questionnaire by adding the scores of the four scales. The SQ consists of 92 items, which are words or short statements that indicate symptoms or level of well-being. Individuals responded "yes" or " no," or "true" or "false," to each word or statement. The standard form of the questionnaire was completed, asking subjects to indicate how they felt in the last week. The validity, reliability, and usefulness of the questionnaire have been reported in various studies of asymptomatic subjects(16) and subjects with physical diseases.(17) In a study of anxious patients, the test-retest correlation coefficients were .71 to .95.(18)

Data Analysis

To examine the relationship between each of the eight scales of the SF-36 and physiological and psychological impairments and demographic characteristics, eight sets of analyses were conducted. Baseline scores were maintained as continuous data for six of the SF-36 scales: energy/fatigue, general health perceptions, mental health, bodily pain, physical functioning, and social functioning. The scales for physical and emotional role limitation were categorized as high ([greater than]50) or low ([less than or equal to]950) because of the non-Gaussian distributions of scores.

In each set of analyses, the relationship of an SF-36 scale score and the following variables was examined: gender, age, education, living situation, event or procedure precipitating referral, hypercholesterolemia, hypertension, LVH, current smoking, comorbidity, peak [VO.sub.2], and total psychological distress score. The 12 variables were selected (1) based on evidence from the literature that they had potential relationships to health status or (2) based on a need for rehabilitation professionals to understand the relationship.

Age, years of education, peak [VO.sub.2], and psychological distress score were maintained as continuous variables. Individuals were classified as living alone or with a significant other, having or not having hypercholesterolemia, having or not having a history of hypertension, having or not having LVH, currently smoking or not smoking, and having or not having lung disease or diabetes.

To initially examine the relationship between health status and characteristics of the individual and physiological and psychological impairments, the univariate relationships of each variable with each SF-36 scale were determined. The relationships of variables to scores for energy/fatigue, general health perceptions, mental health, bodily pain, physical function, and social function scales were determined using general linear-regression models. Because scores on the scales for emotional and physical role limitation were not normally distributed, the analyses for these scales were conducted using logistic regression procedures.

A forward selection procedure(19) was then used to select independent variables to enter each subsequent model with the goal of producing parsimonious models that could describe the relationship of the variables of interest to health status. An alpha level of .01 was used as the criterion for a variable to be included in a final model.

For the purposes of providing a framework for the results and discussion, the SF-36 scales were categorized into two health dimensions, physical and psychological, as described by McHorney et al.(14) According to these authors, the physical health component is best measured by the scales for physical functioning, physical role limitation, general health perceptions, and bodily pain; the psychological health component is best measured by the scales for mental health, social functioning, and emotional role limitation; and the energy/fatigue scale is similarly related to both health components.

Results

Relationships of Variables With Health Status Scales Table 2 shows the statistically significant univariate associations of demographic characteristics and indicators of physiological and psychological impairments with health status scores. Table 3 shows the final models that describe the multivariate relationships of the independent variables to health status.

[TABULAR DATA OMITTED]

Scales measuring the physical dimension of health. Twenty, three percent of the variability in physical functioning was accounted for by a model that included peak [Vo.sub.2] and total psychological distress score. The physical functioning score was positively associated with [Vo.sub.2] and negatively associated with total psychological distress score. Oxygen consumption contributed to the variability of this scale to a greater extent than did the total psychological distress score. No other independent variables were associated with this scale. The physical role limitation score was related to the total psychological distress score and years of education. The odds of being in the "high" category in this scale were 0.94, given an increase of one point on the total psychological distress score. The odds of being in the "high" category were 1.20, given an increase of 1 year of education. The bodily pain score was negatively associated with the total psychological distress score. This variable accounted for 18% of the variability in the scale score. Eighteen percent of the variability in the general health perception score was accounted for by measures of physiological and psychological impairment. Peak was positively associated and the total psychological distress score was negatively associated with the general health perception score. The psychological distress score contributed to the variability in this scale to a greater extent than did [Vo.sub.2].

Scales measuring the psychological dimension of health. The total psychological distress score and the event or procedure precipitating referral for cardiac rehabilitation accounted for 57% of the variability in the mental health scale. The psychological distress score was negatively associated with the mental health score and accounted for a greater amount of the variability in this scale than did the precipitating event. Patients with CABG had higher mental health scores than did patients with other precipitating events. The total psychological distress score was associated with the emotional role limitation score. The odds of being in the "high" category in this scale decreased as psychological distress increased (0.95, given a one-point increase). Social functioning was negatively associated with psychological impairment. The model accounted for 16% of the variability in the social functioning score.

Scales measuring both the physical and psychological dimensions of health. Twenty-one percent of the variability in the energy/fatigue score was accounted for by a model that included the total psychological distress score. Total psychological distress was negatively associated with the score. this scale.

Discussion

In patients entering cardiac rehabilitation programs, high levels of psychological distress are related to poorer health in both the physical and psychological dimensions. Peak [Vo.sub.2], a measure commonly considered by physical therapists to be important to health in individuals with cardiovascular disease, is related to physical functioning of patients and their perception of their own health but does not explain variability in other health-related behaviors. This finding confirms the need for physical therapists involved in cardiac rehabilitation programs to comprehensively assess their patients' health status. We have previously reported that patients enrolled in cardiac rehabilitation programs have low health status in all dimensions of health.[20] The current findings show that clinical data routinely collected, including demographic information, do not provide a very thorough profile of a patient's health. Knowing that psychological impairment is related to physical health, physical therapists working to address their patient's physical health needs must consider collecting data, setting goals, and devising interventions that address psychological impairment.

Similar to our findings, Brooks et al2l found that in a group of patients admitted to the hospital for combined medical and psychological problems, 20% of the variance in the SIP physical subscale and 62% of the variance in the psychosocial subscale were accounted for by psychological variables. Anderson et al[22] found that in women seen in general medical practice, lower health status scores were associated with a mental health diagnosis. Twenty-five percent or more of the sample in our study exhibited some evidence of depression, hostility, or anxiety or had symptoms of somatization in addition to their cardiovascular disease. Similar levels of psychological impairment have been found in other studies of patients with cardiovascular disease.[6,23,24]

Peak [Vo.sub.2], a measure of exercise capacity and an indicator of physiological impairment, was related to physical functioning and general health perceptions, scales that are purported to be valid measures of the physical dimension of health.[14] Such relationships are, therefore, not unexpected. Persons with a higher peak [Vo.sub.2] would be more likely to be able to perform those actions measured by the physical functioning scale, such as running, climbing stairs, and walking a mile. Similarly, Rector et al[5] found that many of the SIP scales related to physical health, ambulation, sleep and rest, home management, mobility, eating, and body care and movement were associated with peak [Vo.sub.2] in 45 patients with congestive heart failure. Social interaction, alertness, emotional behavior, and communication scales were not associated with peak [Vo.sub.2]. Neill and colleagues[8] studied a sample of 98 men with a diagnosis of coronary heart disease without comorbidity or recent acute clinical event. They examined the relationship of the physical and social functioning dimensions of health status with exercise duration on a symptom-limited exercise test and other disease symptoms and demographic variables. They found that exercise duration was related to the patients' reports of maximum distance walked, but not to performance of household activities or social activities.

Interestingly, peak [Vo.sub.2] was not related to the physical role limitation scale of the SF-36, one of the scales measuring the physical dimension of health, in our study. The role limitation scales of the SF-36 might be considered as measurements of the level of disability attributed to either physical or emotional health, including limitations in work. The physical role limitation score was related to education level, suggesting that individuals with a higher level of education perform the type of work that can be accomplished with less physical health or routinely have help with work around the home. Neill et al[8] also found that education level was related to a total score for function.

None of the physiological impairments represented by presence of the three main risk factors in cardiovascular disease - smoking, hypercholesterolemia, and hypertension - were associated with any of the SF-36 scales. Pinsky et al[25] determined that hypertension and long-term smoking were related to disability in female subjects and that hypertension was related to disability in men who were free of diagnosed cardiovascular disease. Nelson et al[26] found that smoking was an independent predictor of poor physical functioning, as measured by the Duke Activity Scale Index (DASI), in individuals with cardiac disease. Their subjects, although showing evidence of the need for cardiac catheterization, had not had recent cardiac events or procedures. We speculate that the effect of cardiac risk factors on health status of patients who have had a recent cardiac event or procedure is largely overshadowed by psychological impairment or other factors related to the acuity of the health problem.

Comorbidity and disease severity were not related to scores in any of the SF-36 scales in our study. In contrast, Stewart et al[27] demonstrated an additive effect of comorbid conditions on health status in patients with a variety of chronic conditions, and Anderson et al[22] found that lower scores in all scales of the SF-36 were associated with the number of diagnoses. Nelson et al[26] found that the presence of heart failure, angina, and three-vessel disease accounted for variability in physical functioning in patients with cardiovascular disease. In our study, the patients had undergone recent medical interventions designed to reduce physiological impairment. These newly instituted interventions could have had a positive impact on the health status of the patients, reducing the differences between those with impairment and those without impairment. In addition, the psychological impairment associated with a recent acute event could have influences on health status scores that reduce the strength of the relationship of health status to comorbidity, and disease severity seen in other studies.

Several studies have provided evidence for the impact of certain demographic factors on various dimensions of health status in patients with cardiovascular disease. One commonly accepted predictor is age. Health status is believed to deteriorate as age increases.[12] In our study, age showed a univariate relationship to physical functioning but was not related to health status scores in any of the multivariate analyses. Similarly, Nelson et al[26] found that age was related to physical functioning in patients undergoing cardiac catheterization. Bergner et al[7] found that age was related to scores in 3 of the 12 SIP scales in survivors of cardiac arrest. Rector et al[5] did not find any relationship of age and the total SIP score in patients with congestive heart failure.

Although the association of gender with functional status in individuals with cardiac disease has been reported by several investigators,[6,26,28] gender was not related to health status in our study. Kos-Munson et al[6] found that in 92 subjects undergoing CABG, men scored better than women did on all but one SIP scale. Similarly, Bergner et al[7] found that ill 424 survivors of cardiac arrest, men scored better than women did on 6 of the 12 SIP scales. Stanton et al[29] also reported that men had higher levels of physical activity and were less disabled than were women following CABG. Nelson et al[26] found an effect of gender on the DASI scores when age, severity of illness, and comorbidity were entered into the analysis. Because women have lower maximum aerobic capacity than men have, a relationship between gender and DASI scores is not surprising. In addition, many of the items on the DASI ask about the ability to perform activities that are often not done by women. The univariate association of gender with health status using the 20-item MOS health survey has been reported by Stewart and colleagues[27] in a large sample of patients visiting their health care providers. In their sample, women scored lower than men did on all scales except the general health perception scale. Chirikos and Nickel[28] have hypothesized that any positive findings relating gender to health status have resulted from the confounding effects of socioeconomic status.

Study Limitations

One concern addressed by Brooks et al[2l] is that individuals with psychological impairment may respond to health status items in biased ways, potentially increasing the observed relationship to health status. That study showed that answers to items on the SIP, even in a sample of patients with acute mental health problems, were not answered in such a way as to suggest this bias. We feel confident, therefore, that our results are not biased by such an effect. The high degree of association between the mental health and SQ scores might be explained, however, by the similarity of each of the items in this SF-36 scale to some of the items in the SQ. Examination of the SF-36 items suggests that this scale may be measuring psychological impairment, rather than functional limitations in the psychological domain. Similarly, the energy/fatigue scale of the SF-36, although theoretically measuring both the physical and psychological dimensions of health, contains items that may be interpreted as measuring psychological impairment. In contrast, the SQ contains some items that seem to be measures of function such as sleep and appetite. The mixing of items that measure impairment and functional limitation is not uncommon in health status instruments, and the SF-36 was not, to our knowledge, developed based on the conceptual framework for health presented here. The confusion, however, limits the interpretation of the findings in this study.

A convenience sample is an additional limitation. Although decision criteria for referral to cardiac rehabilitation programs vary somewhat from physician to physician, there is no reason to expect that referral will be based on psychological impairment or indicators of health status, as physicians do not generally formally or objectively assess health status. Patients, however, may differentially request referral to rehabilitation programs, with those who perceive their health to be worse pressing their health care providers for treatment. In addition, not all patients referred to cardiac rehabilitation programs actually enroll.

Patients who do not follow through with physician recommendations to enroll in a rehabilitation program may differ from those who do in some important ways. Patients who do not have at least partial insurance coverage generally will not participate in cardiac rehabilitation programs. This observation suggests that patients who have higher socioeconomic status will be more likely to engage in rehabilitation than are those with lower socioeconomic status. The relatively high socioeconomic status of the sample in this study is evidenced by the high level of education (X = 13 years). This effect is probably intensified by the fact that patients must have enough flexibility and resources to get to a program three times per week, spending at least an hour at the program. Although most patients, regardless of type of job, have medical leave from work following an MI or cardiac procedure, those with white-collar jobs tend to have more flexibility in their daily schedules.

Other study limitations result from our data-collection methods. There were a considerable number of subjects for whom data were not complete related to an error in the data-collection form for the SQ although individuals for whom data were not complete were in many respects similar at entry to individuals for whom data were complete. The relationships of variables found in this study might have been different had more data been available from all subjects in the study.

One of the measures of physiological impairment, peak [Vo.sub.2], was based on an estimation rather than on actual collection of expired air. Estimation is typically used in cardiac stress testing. This estimate varies approximately 7% among individuals and is valid for steady-rate exercise.[10] Variability of the estimates used in this study may be greater due to the fact that steady-rate exercise was not guaranteed in each subject. In addition, the majority of subjects tested using the treadmill held on to the rail, but equations to account for rail-holding were available only for the standard Bruce protocol. In spite of the potential error, the estimated co, correlated most highly with the physical functioning score of the SF-36, providing some evidence of its validity as a physiological measure of exercise capacity.

Conclusion

In patients engaging in cardiac rehabilitation, psychological impairment is related to both the physical and psychological dimensions of health status. The findings of this study make it clear that measurement of impairment or collection of data concerning patient characteristics is not sufficient to fully assess the health of patients with cardiac disease. The relationship of impairment and individual characteristics to functional limitations and disability is complicated and multifaceted. This study presents a small piece of the puzzle in terms of assessing these relationships. Our findings indicate that health care providers in cardiac rehabilitation programs should assess their patients' health status in a comprehensive manner to gain a better understanding of the patients' health and the outcomes of treatment. Moreover, in a patient-centered environment, physical therapists should identify problems, goals, and interventions that apply to both the physical and psychological impairments of their patients in an effort to address the patients' health requirements.

Acknowledgments

We thank the contributing centers and the staff associated with the Massachusetts Association of Cardiovascular and Pulmonary Rehabilitation; Kathy McCool, RN, Alexander Leaf, MD, and Judy Scheer, RN, of the Cardiovascular Health Center, Massachusetts General Hospital; Gary Balady, MD, and Carol O'Malley, RN, of the Cardiac Rehabilitation Program of the University Hospital; Jeanne Colbath, RN, of Brighton Cardiac Rehabilitation at St Elizabeth's Hospital; Fern Cooper, RN, MSN, of Faulkner Hospital Cardiac Treatment Center; Martha Ethier, RN, of Franklin Medical Center; Sally Pratt of Greater Lowell Cardiac Rehabilitation Center; Cynthia Loiselle, RN, Owen Christianson, MD, and Patricia Balaunas of Henry Haywood Memorial Hospital; Deborah Sullivan, RN, Diane Koczat, RN, and Ann McMackin of Lahey Clinic Medical Center-cardiac Rehabilitation Program; Diane Lavin, RN, of Medical Center of Central Massachusetts-Memorial; Aggie Casey, RN, and Nancy Tomaso, RN, of Milford-Whitinsville Regional Hospital Cardiac Rehabilitation Center; Rosemary Bill, RN, of New England Memorial Hospital; Mary Wade, RN, MSN, Mary Travaso, RN, and Barton Heller, MD, of Norwood Hospital; and Lynn Gotti, RN, of Shaughnessy-Kaplan Rehabilitation Hospital. We give special thanks to Terry Michel, PT.

References

[1] World Health Organization: Constitution of the World Health Organization. In: Basic Documents. Geneva, Switzerland: World Health Organization; 1948. [2] Nagi S. Some conceptual issues in disability and rehabilitation. In: Sussman M, ed. Sociology and Rehabilitation. Washington, DC: American Sociological Association; 1965:100-113. [3] International Classification of Impairments, Disabilities, and Handicaps. Geneva, Switzerland: World Health Organization; 1980. [4] Ware JE, Brook RH, Davies AR, Lohr KN. Choosing measures of health status for individuals in general populations. Am J Public Health. 1981;71: 620-625. [5] Rector TS, Francis GS, Cohn JN. Patients' self-assessment of their congestive heart failure, part 1: patient-perceived dysfunction and its poor correlation with maximal exercise tests. Heart Failure. 1987;3:192-196. [6] Kos-Munson BA, Alexander LD, Hinthorn PAC, et al. Psychosocial predictors of optimal rehabilitation post-coronary artery bypass surgery. Sch Inq Nurs Pract. 1988;2:171-193. [7] Bergner L, Bergner M, Hallstrom AP, et al. Health status of survivors of out-of-hospital cardiac arrest six months later. Am J Public Health. 1984;74: 508-510. [8] Neill WA, Branch LG, DeJong G, et al. Cardiac disability: the impact of coronary heart disease on patients' daily activities. Arch Intern, Med. 1985; 145:1642-1647. [9] Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36), 1: conceptual framework and item selection. Med Care. 1992;30: 473-483. [10] American College of Sports Medicine. Guidelines for Graded Exercise Testing and Exercise Prescription. Philadelphia, Pa: Lea & Febiger; 1991. [11] McConnell TR, Clark BA. Prediction of maximal oxygen consumption during handrail-supported treadmill exercise. Journal of Cardiopulmonary Rehabilitation. 1987;7:324-331. [12] Stewart AL, Hays RD, Ware JE. The MOS short-form general health survey: reliability and validity in a patient population. Med Care. 1988;26: 724-732. [13] McHorney CA, Ware JE, Rogers W, et al. The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts. Med Care. 1992;30:MS25.3-MS265. [14] McHorney CA, Ware JE, Raczek AE. The MOS 36-item short-form health survey (SF-36), II: psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care. 1993;31:247-263. [15] Kellner R. A symptom questionnaire. J Clin Psychol. 1987;48:268. [16] Kellner R, Pathek D, Romaniik R, Winslow WW. Life events and hypochondrial concerns. Psychiatric Med. 1983;1:133-144. [17] Hashimoto F, Kapsner K, Kellner R. Upper respiratory infections increase self-rated hostility and distress. Int J Psychiatry, Med. 1987;17:41. [18] Kellner R. Improvement criteria in drug trials with neurotic patients: part 2. Psychol Med. 1972;2:73-80. [19] Kleinbaum DG, Kupper LL, Muller KE. Applied Regression Analysis and Multivatiable, Methods. Boston, Mass: PWS-Kent Publishing Co; 1988:314-328. [20] Jette DU, Downing J. Health status of individuals entering a cardiac rehabilitation program as measured by the Medical Outcomes Study 36-item Short-Form Survey (SF-36). Phys Ther. 1994;74:521-527. [21] Brooks WB, Jordan JS, Divine GW, et al. The impact of psychological factors on measurement of functional status. Med Care. 1990;28:793-804. [22] Anderson JSTC, Sullivan F, Usherwood TP. The medical outcomes study instrument (MOSI): use of a new health status measure in Britain. Fam Pract. 1990;7:205-218. [23] Schleifer SJ, Macari-Hinson MM, Coyle DA, et al. The nature and course of depression following myocardial infarction. Arch Intern Med. 1989;149: 1785-1789. [24] Magni G, Unger HP, Valfre C, et al. Psychosocial outcome one year after heart surgery. Arch Intern Med. 1987;147:473-477. [25] Pinsky JL, Branch LG, Jette AM, et al. Framingham disability study: relationship of disability to cardiovascular risk factors among persons free of diagnosed cardiovascular disease. Am J Epidemiol. 1985;122:644-656. [26] Nelson CL, Herndon JE, Mark DB, et al. Relation of clinical and angiographic factors to functional capacity as measured by the Duke activity status index. Am J Cardiol. 1991;68:973-975. [27] Stewart AL, Greenfield S, Hays RD, et al. Functional status and well-being of patients with chronic conditions. JAMA. 1989;262:907-913. [28] Chirikos TN, Nickel JT. Socioeconomic determinants of continuing functional disablement from chronic disease episodes. Soc Sci Med. 1986;22: 1329-1335. [29] Stanton BA, Jenkins CD, Savageau JA, Thurer RL. Functional benefits following coronary artery bypass graft surgery. Ann Thorac Surg. 1984;37:286-290.