Enrollment in the Health Alliance Plan HMO Is Not an Independent Risk Factor for Coronary Artery Bypass Graft Surgery
Background Henry Ford Hospital is the sole provider of cardiac surgical services for the Health Alliance Plan, a health maintenance organization (HMO) that presently serves 450 000 enrollees.
Methods and Results To determine the effect of managed care referral patterns on the outcome of coronary artery bypass graft (CABG) surgery, we retrospectively reviewed two concurrent groups of patients, 569 HMO patients and 225 patients with fee-for-service (FFS) insurance, who had undergone isolated primary CABG surgery between January 1, 1990 and January 31, 1994. The 605 patients with Medicare operated on during the same time frame were excluded to obviate age bias. Age, sex, use of cardiac medications, history of prior percutaneous transluminal coronary angioplasty or thrombolytic therapy, history of recent and remote myocardial infarction, extent of coronary disease, presence of preexisting comorbid conditions, and incidence of unstable clinical syndromes and left ventricular dysfunction (ejection fraction<40%) were comparable for both groups. Inhospital mortality (HMO group, 1.9%; FFS group, 2.2%), mean ICU stay (HMO, 2.6±0.3 days; FFS, 2.3±0.3 days), and total hospital length of stay (HMO, 9.8±0.8 days; FFS, 8.6±0.6 days) were likewise similar.
Conclusions These data refute the notion that the gatekeeper mentality often associated with managed-care health insurance vehicles results in delayed referral of patients with coronary artery disease and results in suboptimal outcome.
Henry Ford Hospital is the sole provider of cardiac surgical services for the HAP, an HMO that at present has approximately 450 000 enrollees. In addition, the hospital and medical group provide service to members of the community at large who have other more traditional health insurance vehicles. This characteristic, we believe, affords the opportunity to compare characteristics and outcomes in concurrent patient groups who should theoretically differ only in the type of health insurance they have.
The present study was undertaken to determine whether enrollment in the managed-care program in and of itself had any impact on clinical characteristics and outcomes in patients with coronary artery disease referred for CABG surgery at a single institution. To the best of our knowledge, no such analysis has been previously reported.
Between January 1, 1990, and January 31, 1994, 1399 patients underwent isolated primary CABG surgery. Five hundred sixty-nine patients belonged to a managed-care program, HAP; 225 patients were insured by a variety of FFS vehicles. These two groups form the basis of this report. We excluded from analysis 605 patients who had Medicare insurance to obviate any age bias. Data for comparison of preoperative demographics, intraoperative parameters, and postoperative outcomes for the two study groups were obtained from a computerized cardiac surgical database.
The following criteria were applied to all patients: Surgery was considered emergent if the patient had been taken directly from the Catheterization Laboratory or Coronary Care Unit (CCU) after initial consultation. Patients in the CCU who received intravenous nitrates and/or heparin were classified as having urgent surgery. All others were categorized as having elective procedures.
Classification of anginal status was as follows: Patients admitted to the hospital electively for surgery with angina as their predominant cardiac symptom were classified as having had chronic stable angina. Patients who underwent cardiac catheterization for stable symptoms but who were admitted for other reasons (ie, positive stress test, threatening anatomy) were likewise classified as having chronic stable angina. Patients were considered to have had unstable angina if they had been admitted with new or worsening symptoms and underwent surgical revascularization before discharge. The subgroup of patients with postinfarction angina were those who were admitted either with a documented transmural or non–Q-wave infarction, who developed postinfarct pain, and who were subsequently operated on during the same admission. Any patient who was first discharged from the hospital and then readmitted was classified as having either chronic stable angina or unstable angina, independent of the time course relative to the infarct.
Patients were classified as having had renal disease if the preoperative serum creatinine level was >2.5 mg/dL or if they had been on dialysis. Chronic obstructive pulmonary disease (COPD) was recorded as being present if the patient required pharmacological therapy or carried the diagnosis by history. Left main coronary artery disease was identified as an estimated reduction in luminal diameter of ≥50%.
All procedures were performed with standard cardiopulmonary bypass techniques using ascending aortic and single venous cannulation, with moderate systemic hypothermia (28°C to 32°C). Myocardial protection was afforded by multidose antegrade cold blood potassium cardioplegia supplemented with topical iced slush. All proximal and distal anastomoses were performed during a single period of aortic cross clamping.
Wilcoxon two-sample rank sum tests were used to compare the provider groups (HAP, FFS) with respect to ordinal and continuous variables. For dichotomous variables, χ2 tests were used to compare the proportion within each provider group with specified traits or conditions. (For low cell frequencies, a Fisher’s exact test was used instead of the χ2 test.) An α-level of 0.05 was used to interpret the results. All data, where appropriate, are presented as mean±SEM.
Comparisons of preoperative demographics and comorbidity, preoperative medical therapy, prior cardiac history, indications for surgery, angiographic findings, and operative categories of the two groups are shown in Tables 1 through 6⇓⇓⇓⇓⇓⇓. One factor was significant at the P<.05 level. Preoperative use of intravenous heparin was found in 35.7% of HAP patients versus 26.7% of FFS patients (P=.015). One additional factor approached significance at this level. In the HAP population, 65.3% of patients were classified as having elective procedures versus 72.4% of those in the FFS group (P=.056).
Intraoperative data for the two groups were as follows: HAP patients received 2.61±0.04 grafts compared with 2.67±0.06 grafts for those with FFS insurance (P=.371). Cross-clamp times (HAP, 44.4±0.8 minutes; FFS, 46.6±1.2 minutes; P=.138) and bypass times (HAP, 80.0±1.2 minutes; FFS, 81.4±1.7 minutes; P=.197) were likewise comparable.
Hospital mortality was 1.9% for HAP patients and 2.2% in the FFS group (P=.794). Mean length of stay in the ICU was 2.6±0.3 days in the HAP group versus 2.3±0.3 days in the FFS group (P=.734). Total length of hospital stay was also similar for the two groups (HAP 9.8±0.8 days versus FFS 8.6±0.6; P=.911).
Enrollment in managed-care programs continues to grow, from 2% in 1970, to 7% in 1984, to 15.9% in 1991.1 In fact, recent estimates suggest that >70% of the insured population in metropolitan areas of California now have some form of managed-care insurance.2 Data are increasingly available to support the idea that HMOs can be effective in controlling costs. A randomized trial by the Rand Corporation reported a 25% decrease in expenditures for all health care provided to one HMO group compared with the FFS system despite the fact that care within the study protocol was free to patients in both systems.3 Savings were largely due to a 40% decrease in hospital admissions.
The more critical issue is whether limiting expenditures in this manner comes at the expense of quality care. More specifically, does the gatekeeper concept central to managed-care programs excessively limit patient access to tertiary care and ultimately affect patient outcomes? Lee-Feldstein et al4 reported that for patients with breast cancer diagnosed between 1984 and 1990, those treated at HMO hospitals were at significantly greater risk of dying than those treated at large community hospitals. Ware et al5 reported worse health outcomes among sick low income patients in the HMO group when compared with those in the FFS group. Most recently, Mark et al6 reported that US patients had better functional status and fewer cardiac symptoms 1 year after acute myocardial infarction than Canadian patients and noted that patients in Canada underwent fewer invasive cardiac procedures and had fewer visits to specialists.They suggested that the more aggressive care in the United States may have been responsible for the better outcomes.
The intent of this study was to determine whether enrollment in an HMO had any effect on the outcome of patients with a diagnosis of coronary artery disease referred for CABG surgery. Our hypothesis on initiation of this retrospective analysis was that patients within the HMO group would, because of their presumed restrictions to subspecialty care, present for surgical therapy at a later stage of their disease (ie, older age, more extensive anatomic disease, worse ventricular function). We believed that these patients would therefore constitute a higher risk cohort and have suboptimal outcomes when compared with those patients with traditional FFS insurance vehicles. Given that the treating cardiologists and surgeons were the same for both groups, we believed we could then theorize that this difference in outcomes was related to either delayed recognition or delayed referral by the primary care or gatekeeper component of the managed-care program.
The results of this analysis, we believe, dispute that hypothesis. The preoperative characteristics and postoperative outcomes are quite similar for both groups. This would, we believe, suggest that there was in fact no delay in diagnosis or referral of patients from within the HMO group. Furthermore, allowing for exclusion of the Medicare cohort, the relevant clinical characteristics of the two study groups are comparable with those of other series for factors such as sex, presence of preoperative comorbidity, extent of coronary disease, and previous cardiac event.7 8 9
Despite the great similarity in the patient cohorts, two factors warrant specific comment. More patients in the HAP group were receiving intravenous heparin preoperatively than in the FFS group (35.7% versus 26.7%, P=.015). In addition, 65.3% of patients in the HMO group underwent elective procedures compared with 72.4% of those with FFS insurance (P=.056). One therefore achieved significance, and the other closely approached it. These percentages suggest a proportionate increase in the HMO patients’ level of disease severity on presentation that could contradict our conclusion relative to the hypothesis. The clinical indications for surgery were closely comparable for the two groups (ie, existence of similar percentages in each group of stable and unstable syndromes). These differences then essentially reflect that a somewhat greater percentage of patients in the HMO group were receiving heparin and were in the CCU before surgery. We are concerned that these data may indicate some increase in overall acuteness for HMO patients; while not sufficient in our opinion to cause the hypothesis to be rejected, we believe that continued scrutiny of this issue is certainly warranted. Ultimately, the clinical significance of these differences did not affect the early outcomes for patients in the HMO group.
The data presented here are from a retrospective analysis of patients who had already undergone surgical revascularization. We therefore cannot account for patients with known coronary disease or for that matter with chest pain who have yet to be referred beyond the primary care level. We likewise have no way of knowing whether patients with more advanced disease or symptoms are being denied access to specialty care. Intuitively, it seems likely that if large numbers of HMO patients were being treated excessively at the primary care level, they would present with a higher incidence of previous myocardial infarction, more depressed left ventricular function, and increased prevalence of congestive heart failure; this was not the case. Unlike the HMO group, all of whom were treated at this institution, our FFS group represents some proportion of a population that as a whole can be, and is being, referred to other institutions in the community. Again, similarities between the two groups and those of other series would suggest that our FFS group is not in some way preselected but rather is quite representative of a typical cardiac surgical practice.
It should be noted, and in fact emphasized, that the results of this analysis are specific to a single HMO at a single institution. The HAP, like other HMOs, employs an internist or a general practitioner as the “Gatekeeper.” The patient cannot self-refer to tertiary levels of care but must be referred by this primary care provider. Unlike the general perception of HMO programs, and indeed the practice at some, there are no direct individual financial incentives within the program to discourage these primary care physicians from referring patients to subsequent levels of care when deemed appropriate. It is therefore neither possible nor appropriate to further extrapolate these data into a broad-based endorsement of managed-care health delivery in general. We have no means of comparing the specifics of the HAP program with those of other HMOs. We don’t know whether the lack of financial incentives at the primary care level fosters a more timely referral to tertiary care than would otherwise be expected if such were not the case. Furthermore, this study involves no cost analyses. Therefore, we wonder whether it is possible that outcomes were similar because overall cost expenditures were similar. If so, this would negate the major theoretical advantage of a managed-care program, which is in fact presumed to be effective cost savings.
These data document that enrollment in the HAP was not an independent risk factor for CABG surgery and refute the notion and our hypothesis that the gatekeeper mentality associated with managed-care programs would necessarily result in suboptimal outcomes. To go beyond this conclusion, however, further investigation, including prospective analyses across large demographic groups and intense and more-meaningful cost analyses to adjudicate risk-benefit relations, is needed.
Selected Abbreviations and Acronyms
|CABG||=||coronary artery bypass graft|
|HAP||=||Health Alliance Plan|
|HMO||=||health maintenance organization|
- Copyright © 1995 by American Heart Association
1. Marion Merrill Dow Managed Care Digest, HMO edition, 1992.
How can hospitals survive? Integrated Healthcare Report, July, 1994.
Mark DB, Naylor CD, Phil D, Hlatky MA, Califf RM, Topol EJ, Granger CB, Knight JD, Nelson CL, Lee KL, Clapp-Channing NE, Sutherland W, Pilote L, Armstrong PW. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med. 1994;331:1130-1135.
Khan SS, Kupfer JM, Matloff JM, Tsai TP, Nessim S. Interaction of age and preoperative risk factors in predicting operative mortality for coronary bypass surgery. Circulation. 1992;86[suppl II]:II-186-II-190.