Can Hospital Selection Avoid Deaths?

abstract & commentary

Synopsis: A significant number of deaths can be avoided by referring appropriate surgical and medical patients to HVHs and by actively working to prohibit these procedures from being performed in low-volume facilities.

Source: Dudley RA, et al. JAMA 2000;283:1159-1166.

Each year, a large number of patients die following elective surgery; in fact, in the Medicare population alone, 17,000 patients died in 1995 after undergoing 10 types of elective surgical procedures. Many previously published studies have demonstrated that mortality rates for certain procedures and diagnoses are significantly lower in hospitals that perform a high volume of these procedures when compared to the mortality rate for the same procedures performed in low-volume hospitals (LVHs).1

Dudley and colleagues from the University of California performed an extensive review of the literature from Jan. 1, 1983, through Dec. 31, 1998, looking for the highest-quality studies assessing the mortality-volume relationship for many medical and surgical diagnoses and then calculated the odds ratios for in-hospital mortality comparing LVHs with high-volume hospitals (HVHs). The purpose of the current study was to identify procedures and diagnoses for which there is good evidence that a volume-outcome relationship exists and to estimate the annual number of deaths in California LVHs that can be attributed primarily to their low volume. Mortality was found to be significantly lower in HVHs for elective abdominal aortic aneurysm repairs, carotid endarterectomies, coronary angioplasties, heart transplantations, treatment of HIV syndromes, pediatric cardiac surgery, coronary artery bypass surgery, lower extremity arterial bypass surgery, esophageal cancer surgery, and cerebral aneurysm surgery. A total of 58,306 patients were admitted to LVHs in California for one or more of these procedures in 1997 and, for that one year, it was estimated that 602 deaths could be attributed simply to the fact that these patients were admitted to the LVHs.

Comment by Harold L. Karpman, MD, FACC, FACP

Hospital referral initiatives have been extremely difficult to recommend due to the significant difficulties that exist when attempting to obtain reliable data that is uninfluenced by random events and can have significant effects on the mortality rates observed to occur in many medical conditions.3,4 There are many reasons for the absence of appropriate referral initiatives; in fact, with rare exceptions, the lack of adequate data has influenced health plans and health care purchasers to avoid selectively referring patients to hospitals with low case-mix-adjusted mortality or to hospitals that perform high volumes of specific procedures.2

Dudley et al carefully evaluated the literature from multiple sources including MEDLINE, Current Contents, and First-Search Social Abstracts databases. They evaluated the percentage of LVH patients who were admitted through emergency departments and determined the additional distance that these patients would have had to travel to reach a HVH; this information permitted them to properly assess clinical and practical barriers that could influence the referral to the HVHs. They determined that a significant number of deaths could be avoided in California through referral of patients to the regional HVHs. Statistically significant relationships between hospital volume and mortality were identified for the 10 surgical procedures and the one medical condition. The results revealed that, for example, patients were 64% more likely to die following abdominal aortic aneurysm repair in LVHs than in HVHs; if we extrapolate these numbers nationally, more than 4000 deaths per year could be avoided by referring patients with this surgical abnormality to HVHs.

Many techniques can be used for steering high-risk surgical patients to HVHs; however, it should be recognized that appropriate arguments can be offered against these steerages such as interference with the local continuity of care, the adverse effects that might occur particularly in rural areas because of the logistic problems, travel burdens, and the possible reduction of access to health care for rural patients—all of which could result in financial instability of local hospitals. On the other hand, many LVHs are not located in sparsely populated rural areas and, in fact, they are more commonly located in hospital-dense metropolitan areas in relatively close proximity to high-volume referral centers.5 Therefore, fully 75% of California patients undergoing surgery at low-volume centers in 1997 would actually have had to travel fewer than 25 additional miles to reach the nearest HVH.

Despite the many problems and the need to acquire additional data, it would appear that the potential benefits achieved by referring patients to high-volume medical centers are too significant to simply ignore. The ability and willingness of patients to move to HVHs and the effects of such movement on local health care facilities should be carefully studied, however, despite all of the uncertainties of the data that have been reported thus far, payers such as employers, health plans, and government health care programs should immediately start to analyze data in great detail in order to determine if they should adapt policies regarding selective referral of patients for specific surgical and/or medical procedures to appropriate HVHs. In summary, Dudley et al have clearly demonstrated that a significant number of deaths can be avoided by referring appropriate surgical and medical patients to HVHs and by actively working to prohibit these procedures from being performed in low-volume facilities.


1. Hannan EL, et al. Med Care 1991;29:1094-1107.

2. Ronning PL, Meyer JW. J Cardiovasc Manag 1995;6:21-23.

3. Hofer TP, Hayward RA. Med Care 1996;34:737-753.

4. Zalkind DL, Eastaugh SR. Hosp Health Serv Adm 1997;42:3-15.

5. Grumback K, et al. JAMA 1995;274:1282-1288.

Patients subjected to abdominal aortic surgical repair in low-volume hospitals:

a. are 64% more likely to die.

b. leads to 4000 unnecessary deaths per day nationwide.

c. should be admitted electively only to high-volume hospitals.

d. All of the above