Rating Hospitals:  Is Past Performance Indicative of Future Results?

By Ronald L. Scott

The Texas Health Care Information Council (THCIC) collects quality of care data from managed care organizations and other health care providers and issues reports available to the public.  Among other data, the state agency collects discharge data from hospitals in Texas.  See Tex. Health & Safety Code § 108.011 et seq. THCIC recently released Indicators of Inpatient Care in Texas Hospitals, 2000, a report evaluating hospital quality of care.  The report provides comparative data on 411 Texas hospitals. According to the report, the surveyed hospitals account for 95% of patient hospital admissions in Texas.

The report evaluates a number of “quality indicators” for common medical procedures and conditions.  Quality indicators include volume of procedures performed, mortality for inpatient procedures, mortality for inpatient conditions, and utilization.  The volume indicator tells how many times a hospital performed particular procedures on patients during the reporting period.  Other things being equal, many patients might prefer a facility that has performed a larger volume of procedures.  Utilization measures the rate of use by hospitals of procedures such as Cesarean section delivery and vaginal birth after Cesarean section.  Extremely high or low utilization rates may raise concerns that should be evaluated further.  The mortality indicator reports the number and percentage of patients who died after undergoing a particular procedure or while being treated for a particular condition.  The risk-adjusted mortality rate (RAMR) accounts for differences in outcome due to the age and health status of the patients. Hospitals have long objected to the public release of mortality data by arguing that their patients were sicker or older than average.

To illustrate how one may use the report, assume that your grandmother needs hip replacement surgery and has asked you to find a good hospital in Houston, Texas where she will have the surgery performed. Simply click on Hip Replacement under the mortality for inpatient procedures category. The state RAMR is 0.3%.  The report notes which hospitals have significantly lower or higher RAMRs. For example, Memorial Herman Northwest Hospital in Houston has a RAMR of 2.2, more than seven times the state average and the highest of any hospital in Houston for this procedure.  Hospitals are allowed to enter “comments” on their data, and Memorial Herman’s comment says in part:

Memorial Hermann Northwest Hospital supports the process of providing consumers with access to quality information.  We agree with the limitations on use of the data as published in the Agency for Healthcare Research on Quality's (AHRQ's) Guide to Inpatient Quality Indicators:  Quality of Care in Hospitals--Volume, Mortality, and Utilization and how the data should be used and interpreted (http://www.ahrq.gov/data/hcup/inpatqi.htm).  We utilize the data in our internal quality review process for performance improvement.

A look at the underlying data set forth in the “detailed table” shows that Memorial Hermann Northwest had one fatality while performing 34 hip replacements.  By comparison, Methodist Hospital performed 320 procedures with one fatality, resulting in an average RAMR.  St. Luke’s Episcopal Hospital (184 procedures) and Texas Orthopedic Hospital (112 procedures) both reported no fatalities.

How useful is this report to the typical patient?  Would you avoid sending your grandmother to Memorial Hermann Northwest on the basis of one fatality?  Or would you choose Methodist Hospital over Texas Orthopedic because Methodist performed three times as many procedures—even though Texas Orthopedic has a lower RAMR?  Statistically, these two hospitals are virtually equals on this indicator since neither has a statistically- lower RAMR than the state average. The report does not indicate what physician performed the procedures, and results within a facility could vary by physician.

The report can in some cases be useful to patients, but only if they understand the limitations of the data and are able to deal with the complexity of the report.  Insurance reimbursement issues also seriously limit the utility of the report. If grandmother’s health insurance only allows her to be treated at a particular Houston hospital, the report may simply increase her anxiety if such hospital has a low RAMR.  Further, her physician may only have hospital privileges at a few facilities, or the best hospital may be geographically inconvenient. Still, the THCIC is to be congratulated for publishing the report.  The report quantitatively evaluates quality of care and allows the public to access the findings.  You can bet that I will check out a hospital’s track record before admitting myself or a family member to the hospital.