Previous randomized trials and observational treatment comparisons [1, 2, 5, 6] have failed to show an effect of bypass surgery on return-to-work rates, contrary to initial expectations. Most of these studies were done during the 1970s; however, substantial evidence exists that bypass surgery in the 1980s was done more effectively with less morbidity and mortality [7]. To date, the effects of coronary angioplasty on employment have been less intensively investigated [8-12].

We compared return-to-work rates in patients treated with angioplasty, bypass surgery, and medical therapy between 1986 and 1990 while accounting for the effects of nonmedical factors on long-term employment rates [13].

Methods

Top Methods Results Discussion Author & Article Info References

Patients

We studied patients with suspected coronary artery disease who were referred to Duke University Medical Center for diagnostic cardiac catheterization between March 1986 and June 1990. Patients were eligible for this study if they were younger than 65 years at the time of catheterization, were referred for evaluation of chest pain or equivalent ischemic symptoms, did not require intensive cardiac care at the time of their catheterization procedure, and had no previous angioplasty or bypass surgery. Patients found at angiography to have normal coronary arteries or clinically insignificant coronary artery disease ( 50% stenosis) were excluded. As described previously [13], 2022 patients meeting these criteria were enrolled in a prospective study of economic and quality-of-life outcomes. Of these, 1252 were employed at the time of their catheterization, and they form the basis of the present investigation of employment outcomes. Treatment groups were defined according to the first revascularization procedure the patient received within 60 days of catheterization: angioplasty (n = 312), bypass surgery (n = 449), and medical treatment (n = 491). The median time from catheterization to revascularization was 1 day for the patients who had angioplasty (25th percentile to 75th percentile, 1 to 3 days) and 5 days for patients who had bypass surgery (3 to 9 days).

Data Collection and Information System

The computerized information system used in this study, the Duke Cardiovascular Disease Databank, has been described previously [14]. Demographic, clinical, and angiographic data were collected prospectively as part of the routine patient care process [15].

Baseline quality-of-life and job-related data were usually collected on the day before catheterization or, less frequently, after the catheterization but before referral for revascularization using a self-administered questionnaire, as previously described [13]. Functional status was assessed with the Duke Activity Status Index, a 12-item scale developed and validated against measured maximal exercise oxygen consumption. The Duke Activity Status Index assesses whether patients can do a spectrum of activities without difficulty, and in this study, it was scored from 0 to 12, with higher scores indicating better functional capacity [13, 16]. General health perceptions were assessed by a four-level ordinal question that asked about the effect of the patient's health on function. Psychological status was assessed with four visual analog scales graded from 0 to 10 (lower scores indicate better function) from the General Well-Being Schedule covering anxiety, depression, health concerns, and energy [17]. Socioeconomic status was assessed with questions on years of education, marital status, spouse occupation, number of dependents, and income. The patient's job was assigned to one of six job classes using standard criteria [18]: managerial and professional; technical and sales; service; farming, forestry, and fishing; precision production, craft, and repair; and operators, fabricators, and laborers. The physical demands of the job were categorized using U.S. Department of Labor criteria into one of four classes (sedentary, light, medium, and heavy) [19]. The psychological stress of the job and the decision latitude available to the patient at work were measured using scales designed by Karasek and colleagues [20, 21].

Follow-up

All patients were contacted by mailed questionnaire or telephone interview at 1 year and were asked to complete a battery of quality-of-life and job status questions similar to those administered at baseline. Follow-up was 98% complete. Because the date of return to work was not included in this follow-up questionnaire, we did additional follow-up in a subset of patients to collect this data. Seventy-two patients who had returned to work (25 receiving medical treatment; 24, bypass surgery; and 23, angioplasty) were randomly selected from the December 1989 to June 1990 cohort and were contacted by telephone to determine the date of return to work, so that the time interval from cardiac catheterization until initial return to work (full- or part-time) could be calculated.

Data Analysis

Discrete variables were summarized with percentages and continuous variables were summarized with medians and interquartile ranges (25th to 75th percentile). In this study, the binary logistic regression model [22] was used to test for differences among treatment groups in 1-year work status, unadjusted and adjusted for previously identified independent predictors of follow-up work status [13]. Patients who died before their 1-year follow-up (n = 27) were deleted from the final analysis. Including these patients (3 receiving angioplasty; 9, bypass surgery; and 15, medical treatment) and counting them as not working at 1 year did not alter the results.

Differences among the treatment groups in time to return to work were tested using the Kruskal-Wallis test. In order to estimate the economic value of those days lost from work and the potential impact of an earlier return to work, we multiplied each patient's weekly income (calculated from the midpoint of the patient's annual income range ascertained at the baseline pretreatment assessment) by the number of weeks lost from work by that patient. To eliminate the potential effect on these results of slight imbalances in income distribution among treatments, we repeated the calculations using the average weekly income for the entire 70 patient sample in place of the individual income values. Both time until initial return to work and the dollar value of that lost productivity were plotted graphically using Kaplan-Meier methods [23] to display the total distribution of those variables by treatment group.

Results

Top Methods Results Discussion Author & Article Info References

Baseline characteristics of the three treatment groups differed in medical factors and in several nonmedical factors shown to predict subsequent work outcomes Table 1, Table 2, and Table 3. Overall, the angioplasty group was significantly younger, had fewer previous myocardial infarctions, had less congestive heart failure, and had less extracardiac vascular disease but had a higher prevalence of unstable angina than the patients who were treated medically or had bypass surgery (Table 1). Coronary anatomy also differed substantially Table 2, with more one-vessel disease in the angioplasty group and more three-vessel and left-main disease in the bypass surgery group. In addition, patients who had angioplasty had the highest ejection fractions, and patients receiving medicine had the lowest.

Patients who had angioplasty had the highest functional status at the time of initial evaluation (see Table 3); patients who received medicine had intermediate levels; and patients who had bypass surgery had the lowest functional status [median of six activities done without difficulty compared with five activities and four activities, respectively]. Seventy-seven percent of patients who had angioplasty reported that their health caused no or only mild limitation of general functioning compared with 68% for patients who received medicine and 60% of patients who had bypass surgery.

The three groups were similar, however, in terms of the socioeconomic predictors of work status (see Table 3). The median education level for all three groups was completion of high school, and the percentages who were married and who had a working spouse were also similar. In addition, the number of dependents in the household and the income did not differ by treatment group. The three treatment groups were also essentially identical at baseline in psychological status, with similar degrees of anxiety, depression, energy levels, and health concerns. Further, the three treatment groups had similar types of employment and similar physical and psychological demands on the job (Table 4). Particularly, the percentage of each group with heavy physical job demands was the same, as were the scores on the job psychological-demands scale.

Overall, at 1 year, 84% of patients who had angioplasty were still employed, compared with 79% of patients who had bypass surgery and 76% of patients receiving medicine. Employment rates for patients who had angioplasty and patients receiving medicine were different (P = 0.008), but rates for angioplasty compared with bypass surgery (P = 0.13) and medicine compared with bypass surgery (P = 0.20) were not different. To determine whether the difference between angioplasty and medicine was due to treatment effect or was related to imbalances in baseline factors, we adjusted this comparison for all of the factors we have previously validated as independent determinants of return-to-work outcomes Appendix Table 1 [13]. With these factors included in a logistic model, 1-year employment rates in patients who had angioplasty and patients who received medical treatment were not different (P > 0.2).

Of the patients who returned to work, most (86% of the angioplasty group, 88% of the bypass surgery group, and 87% of the medically treated group) went back to the same jobs at the same hours. At 1 year, approximately 6% in each group were doing less strenuous work, 5% had cut down substantially on their hours, and 1% had done both. Follow-up income at 1 year was only collected on the last 370 patients in the study cohort. A decrease in income ( 1 category, see Table 3 occurred in 31% of patients who had bypass surgery, 26% of those who had angioplasty, and 23% of medically treated patients.

A random sample of 72 patients (23 receiving angioplasty; 24, bypass surgery; and 25, medical treatment) who returned to work was interviewed to determine retrospectively the interval between initial angiography and return to work (either full- or part-time). The patients who had angioplasty had a median return-to-work interval of 18 days (25th percentile to 75th percentile, 7 to 40 days; mean, 27 ±25 days) compared with 54 days (25th percentile to 75th percentile, 36 to 85 days; mean, 67 ±58 days) for those who had bypass surgery and 14 days (25th percentile to 75th percentile, 9 to 57 days; mean, 45 ±80 days) for those who received medical therapy (P = 0.002). Figure 1 shows the days from diagnostic cardiac catheterization until initial return to work by treatment group. The bypass surgery group took longer to return to work than either of the other two groups. The medically treated and angioplasty groups had a similar proportion of patients who returned to work in about 40 days, but the medically treated group had a greater proportion of patients with prolonged return-to-work times.

View larger version (20K): [in this window] [in a new window]   Figure 1. Kaplan-Meier plot of time to initial return to work by treatment group for 72 patients. The x-axis shows days from diagnostic cardiac catheterization until return to work. The y-axis shows the proportion of patients returning to work at each time interval. The 0.5 point indicates the median value; the area under each curve indicates the mean value for that group: angioplasty, 27 days; bypass surgery, 67 days; and medicine, 45 days. CABG = coronary artery bypass graft; PTCA = percutaneous transluminal coronary angioplasty.

Using the weekly income of each patient to assign an economic value to this time lost from work, the average loss for patients who had angioplasty was $2770, whereas that for patients who had bypass surgery was $8434 and for patients who received medicine was $4357 (Figure 2). The average difference in lost wages (productivity) between bypass surgery and angioplasty by these calculations was $5664 per patient, whereas the medicine-angioplasty difference was $1767 per patient. The economic savings of angioplasty compared with medicine were due to the fact that, although the median time to return to work was slightly longer with angioplasty, more patients receiving medicine took longer to return to work. To eliminate the effects of small imbalances in the income distribution among the three treatments, we repeated these calculations using the average income for all 72 patients. With this approximation, the average value of time lost for work by patients who had angioplasty was $2795 per patient compared with $6941 for bypass surgery and $4681 for medicine. The difference between bypass surgery and angioplasty was $4146 per patient, whereas that between medicine and angioplasty was $1886 per patient.

View larger version (20K): [in this window] [in a new window]   Figure 2. Kaplan-Meier plot of the dollar value of the initial time lost from work displayed in Figure 1 by treatment group. The average weekly income for all 72 patients was used in the calculations. The x-axis shows the dollar value of the initial time lost from work (weekly income multiplied by the number of weeks lost from work). The y-axis shows the proportion of patients at or above each x-axis value for each treatment group. The 0.5 point indicates the median value; the area under each curve indicates the mean value. CABG = coronary artery bypass graft; PTCA = percutaneous transluminal coronary angioplasty.

Discussion

Top Methods Results Discussion Author & Article Info References

Coronary artery disease exacts a considerable toll on functional capabilities and is the cause of a substantial amount of work disability [24], in addition to its well-known effects on morbidity and mortality. Coronary bypass surgery has had surprisingly little effect on employment rates in patients with coronary artery disease [25-28]. The psychological effect of major cardiac surgery, along with a prolonged convalescence, may contribute to the lack of efficacy of bypass surgery on increasing subsequent employment [29]. Because coronary angioplasty also decreases coronary obstructions and relieves myocardial ischemia and because patients have a shorter convalescence period, it was hoped that this form of coronary revascularization might improve employment rates in patients with coronary artery disease [8]. This study is the first to compare return-to-work rates after angioplasty with a concurrent cohort of both medically and surgically treated patients. Our major findings are as follows: 1) Patients who had angioplasty have an earlier return to work, which yields an important short-term economic advantage; and 2) patients who had angioplasty, bypass surgery, and medical therapy have indistinguishable rates of employment at 1 year of follow-up.

We have previously reported [30] that the direct short-term costs of angioplasty are approximately half those for bypass surgery. In the current study, we examined the short-term indirect costs associated with loss of productivity [31] during the recuperative phase after these procedures. Because patients who have angioplasty can return to work earlier than those who have bypass surgery, angioplasty has an additional short-term economic advantage compared with surgery of approximately $4000. The relative economic effects of these two procedures for longer periods remain to be determined.

In this study, patients who had angioplasty also had a small increased work productivity compared with medically treated patients, but a larger study is required to determine if the estimated average decrease in indirect costs ($1886) of angioplasty represents a reproducible finding. Any savings in indirect health costs will serve to partially offset the higher direct costs of angioplasty compared with medicine. However, these findings about indirect health costs should not be misinterpreted. We only examined the pattern of initial return to work and did not account for the potential adverse effects of subsequent restenosis with repeat revascularization and further loss of work time. We also did not distinguish between full- and part-time work status in the recovery period. Most patients (about 88%) who returned to work went back to the same job at the same hours. At 1 year, only a small percentage in each treatment group were doing less strenuous work or had substantially decreased their hours. A complete economic analysis of angioplasty compared with bypass surgery and medicine should also account for the time lost from work after the initial return; the costs of baseline and follow-up hospital admissions; and the need for repeat coronary angiography, angioplasty, and bypass surgery. A detailed and comprehensive economic analysis is planned for the multicenter Bypass Angioplasty Revascularization Investigation Study of Economics and Quality of Life.

Previous studies from this institution [13, 32] and others [33] have shown that employment rates in patients with coronary artery disease are influenced more by nonmedical than by medical factors. In fact, we found [13] that standard clinical and catheterization variables only provided about 20% of the total predictive information available about follow-up work outcomes, whereas measures of functional status (specifically the Duke Activity Status Index) provided 27% of the available information, and demographic (age) and socioeconomic factors (education level, race) provided 45% of the available information [13]. Although medical severity-of-illness factors differed substantially among the three treatment groups (see Table 1)and Table 2, most nonmedical factors were well balanced (see Table 3)and Table 4. The better baseline functional status of the patients who received angioplasty, coupled with their younger age and lower incidence of symptomatic heart failure, largely accounted for their higher unadjusted 1-year employment rates. After statistical adjustment for the effect of these differences in baseline characteristics, patients with coronary artery disease in this study who received angioplasty, bypass surgery, or medical therapy had indistinguishable rates of long-term employment.

An important aspect of this study is that it reflects data from recently treated patients. The efficacy of medical therapy and bypass surgery has improved significantly from the 1970s when most reports of return-to-work rates were published [7, 34]. For bypass surgery, in particular, technical improvements in the procedure, such as internal mammary grafting and cold potassium cardioplegia, are associated with decreased operative mortality and improved long-term survival [4]. Although such improvements should theoretically translate into better long-term functional outcomes for patients, we found no evidence for parallel improvements in 1-year employment rates for bypass surgery compared with medicine. Concomitant improvements in medical care could have decreased the evidence of such changes in our study. However, we have also shown that in this cohort, return-to-work behavior is affected more by nonmedical than by medical factors [13]. In particular, neither the amount of coronary artery disease nor the extent of left ventricular dysfunction was an independent predictor of subsequent work status. It is, therefore, not particularly surprising that revascularization with either angioplasty or bypass surgery did not significantly affect work rates. Thus, although correction of physical limitations due to ischemic symptoms is an important first step in vocational rehabilitation, by itself it is not sufficient.

The failure of revascularization to improve long-term return to work rates could be partly due to the low prevalence of severe disability and the small percentage of patients whose job involved heavy physical demands. In other words, these factors could be viewed as decreasing the opportunity for revascularization to have a large effect on work outcomes. On the other hand, the physical demands of the job were not even a univariate predictor of return to work. In an earlier study from our institution, the physical demands of the job were a significant univariate but not a multivariable predictor of return to work [32]. In addition, we have recently reported that the medical, angioplasty, and bypass surgery groups each improved significantly from baseline to 1 year in physical function, as measured by the Duke Activity Status Index and that both revascularization strategies were associated with significantly better improvement than was medical therapy [35]. The conjunction of differential improvement in physical function at 1 year with similar return-to-work rates indicates that medical factors, including the type of therapy selected, are not the major determinants of continued employment.

A few limitations of our study should be noted. First, our study primarily evaluated 1-year work outcomes, thereby designating the same positive outcome to a patient treated with bypass surgery who returned to work after 6 months of recuperation and a patient treated with angioplasty who returned after 1 week. Our major goal was to examine the longer-term effects of each treatment on employment status, so we did not analyze the variations in work behavior that are associated with the early recuperative period after revascularization. Second, patients enrolled in this study are a lower-risk subset of our entire coronary artery disease population, because in order to participate patients had to be stable enough to complete our self-administered baseline assessment. Third, these results are for a single institution; thus, other centers using other patient selection criteria might observe different results. Our results, however, are similar to those reported from several other centers. Fourth, we studied only patients who were initially employed, thus eliminating the possibility that one or more of the treatments might return a substantial number of long-term disabled patients back to work. However, few patients who are disabled for more than 6 months will return to work even if their physical health is fully restored [36]. Our approach, therefore, was to look at the effects of treatment on preservation of employment rather than on rehabilitation of the chronically disabled.

Our analyses describe outcomes of treatment strategies rather than the pure effects of any given therapy on employment outcomes. Like previous observational and randomized treatment comparisons, our study examined the strategies of starting with a revascularization procedure (in this case, either angioplasty or bypass surgery) or starting with medical therapy. However, appropriate patients can and will be crossed over to another therapy when deemed necessary by their physician. Thus, the initial treatment decision and all subsequent consequences of that decision are equally a part of the definition of a given treatment strategy for this study.

We have shown that neither angioplasty nor bypass surgery as an initial treatment strategy had any long-term advantages compared with medical therapy in preserving employment status out to 1 year for patients with coronary artery disease. However, among patients who do return to work, angioplasty is associated with a significantly shorter recuperation period than is bypass surgery and with lower indirect costs due to decreased productivity in the months after the procedure than either bypass surgery or medicine. The overall economic effect of angioplasty compared with alternative therapies needs to be judged in the larger context of its effects on a spectrum of relevant outcomes, including mortality, morbidity, symptom relief, and functional capacity. We are currently analyzing our database for these other outcomes. In addition, a more detailed prospective analysis will be done as part of the multicenter Bypass Angioplasty Revascularization Investigation Study of Economics and Quality of Life comparing patients with multivessel disease randomly assigned to either angioplasty or bypass surgery, although these data will not be available until 1995.