Key financial ratios can foretell hospital closures

Healthcare Financial Management, Nov, 1993 by Monty L. Lynn, Paul Wertheim

PLANNING

An analysis of various financial ratios sampled from open and closed hospitals shows that certain leverage, liquidity, capital efficiency, and resource availability ratios can predict hospital closure up to two years in advance of the closure with an accuracy of nearly 75 percent.

Thirty-nine U.S. community hospitals closed in 1992, compared to 85 in 1988.(a) Despite this decline in hospital closures, hospitals continue to battle rising costs, declining reimbursements, and fierce competition. By the end of 1993, as many as 30 more hospitals may have been forced to close. Knowing which hospitals are most susceptible to closure (based on specified financial indicators) could be of great use to administrators and others interested in hospital financial performance.

Several studies have attempted to predict closure, although most have focused on operational and environmental characteristics rather than financial measures. And many of the financial studies, unfortunately, are limited by methodological flaws that restrict their ability to be applied in all cases. Some, for example, draw their conclusions from data of hospitals in default or from hospitals with budget deficits, rather than actual closures.(b) Others are based only on a small number of closed hospitals or do not compare data between open and closed facilities.(c) Even Cleverley's Financial Flexibility Index, perhaps the best known measure of hospital financial distress, has yet to be adequately validated.(d)

Outside the healthcare environment, several reliable bankruptcy prediction models have been developed, but a model generated in one industry is not always predictive in another. And many academically generated models are too cumbersome to be of practical use in industry. A need was perceived for a way to circumvent the methodological limitations of previous healthcare studies and use statistical advancements developed in bankruptcy predictions to produce a model predictive of hospital closure that was both practical and accurate.

Study sample

The financial data to be studied was obtained from HCFA Annual Medicare Hospital Cost Reports, 1984 to 1987. Twenty-one financial ratios were selected as possible predictors of hospital closure. These ratios were chosen on the basis of their predictive accuracy in other bankruptcy studies, their relevance to hospital finance, and their availability in the HCFA cost reports. The ratios were grouped into categories of leverage, liquidity, capital efficiency, and asset availability, and then were calculated at both one and two years prior to closure.

Hospital closures were identified from a list of closed community hospitals compiled from the American Hospital Association Guide.(e) Hospitals which joined multihospital systems, changed their names, or were absorbed into other organizations through mergers or consolidations were not included in the sample. Closed hospitals for which two years of financial data prior to the year of closure were not reported in the HCFA Cost Reports also were not included in the sample.

Seventy-one hospitals that closed in 1986 or 1987 met the sample criteria and were selected for use in the study.(f) Bed size of the closed hospitals in the sample ranged from nine to 604, with a mean bed size of 64 and median bed size of 39. Two-thirds (64.8 percent) of the hospitals were located in a rural area (that is, located outside a metropolitan statistical area).

To compare the financial ratios between closed and open hospitals, a matched sample of open hospitals then was selected. Hospitals operating in 1986 and 1987 were identified as matching sample candidates if they met three criteria: they had to match the closed hospitals in urban or rural status, number of beds, and be located in the same state; they had to have remained open throughout the study period; and two years of their Medicare cost reports had to be available.

Seventy-one pools of open hospitals were generated, each pool made up of hospitals matching an individual closed hospital. From each pool, an open hospital was selected at random. The final study sample thus matched 71 closed and 71 open hospitals.

Study findings

Twenty-one financial ratios were calculated for each of the closed and open hospitals (from financial data at both one year and two years prior to date of closure of the closed hospital). Four statistical manipulations--means and standard deviations of the ratios; chi-square tests; logit analysis(g); and TABULAR DATA OMITTED analysis of variance--provided insight into the differences between the open and closed hospitals.

First, with respect to mean ratio performance, one would expect that closed hospitals would have higher leverage (debt) ratios than the open hospitals, and lower liquidity, capital efficiency, and asset availability ratios as well. This expectation was found to be true, as both the mean and standard deviations in Exhibit 1 indicate at both one and two years prior to closure.

Chi-square tests were used to measure differences between the ratio means of open and closed hospitals. The chi-square statistic indicates the significance of any difference existing between these means. The results, shown in Exhibit 2, indicate that seventeen of the ratios are significantly different (p |is less than~ .05) between closed and open hospitals one year before closure. Two years before closure, only eight ratios have significantly different means. It is not surprising that predictability decreases when measured two years prior to closure since as a hospital approaches failure the financial factors which signal its demise are accentuated.