Medicare:

Improved Patient Outcome Analyses Could Enhance Quality Assessment

PEMD-88-23: Published: Jun 27, 1988. Publicly Released: Jul 11, 1988.

Additional Materials:

Contact:

Robert L. York
(202) 275-5885
contact@gao.gov

 

Office of Public Affairs
(202) 512-4800
youngc1@gao.gov

Pursuant to a congressional request, GAO examined the Health Care Financing Administration's (HCFA) analysis and use of existing administrative data to monitor the outcome of care received by Medicare patients.

GAO found that HCFA: (1) basically used three analytical approaches for its 1987 Medicare patient outcome analyses, involving 1986 and 1987 data and patient subgroup monitoring over time; (2) strengthened its 1987 analyses regarding patient-level data, clinically coherent diagnostic groups, comorbidity information, and more appropriate illness severity adjustments; and (3) primarily used its outcome analyses to require peer review organizations (PRO) to examine the hospitals the analyses identified as having potential problems. GAO also found that: (1) HCFA contractors used analytical approaches, involving nonintrusive outcomes, a risk-adjusted mortality index, disease staging, and hospital rate-setting, which could enhance data analysis; and (2) other analytical approaches, involving risk-adjusted monitoring and computerized identification of complications, could also enhance data analysis. In addition, GAO found that some of the various approaches: (1) used current Medicare administrative data to better advantage than did others; (2) demonstrated the great potential for identifying types of cases with unusually favorable or adverse outcomes; (3) lacked adequate validation regarding their effectiveness in targeting cases for quality review; (4) lacked capability to analyze any data other than mortality; and (5) did not adequately address the potential deviation or problems caused by random variation, particularly in instances involving only small numbers of cases.

Recommendations for Executive Action

  1. Status: Closed - Implemented

    Comments: Limited validation studies have been conducted, but not all have been released. Further validation studies have not been initiated.

    Recommendation: The Secretary of Health and Human Services should direct the Administrator, HCFA, to assess periodically the relative strengths and limitations of available approaches for analyzing Medicare patient outcomes data in terms of their substantive focus, technical adequacy, and degree of validation (i.e., their overall effectiveness in identifying patterns of patient care with quality problems). These assessments should guide the selection of analytical approaches used in future HCFA reviews of Medicare patient outcomes. HCFA should ensure that analyses of Medicare patient outcomes from administrative files employ approaches that have been validated to some degree through independent data sources, and any results publicly released should describe the extent of that validation.

    Agency Affected: Department of Health and Human Services

  2. Status: Closed - Not Implemented

    Comments: It has been more than 3 years since the report was issued. There is no prospect of further agency implementation.

    Recommendation: The Secretary of Health and Human Services should direct the Administrator, HCFA, to expand HCFA analysis of comparative outcomes among patient subgroups, such as those defined by diagnostic and demographic characteristics. If substantial differences in outcomes among such groups are found after adjusting for differences in patient severity, HCFA should experiment with strategies for targeting quality-of-care reviews based on these analyses.

    Agency Affected: Department of Health and Human Services

  3. Status: Closed - Implemented

    Comments: Hospital mortality analyses for 1988 included data from 1986 and 1987.

    Recommendation: The Secretary of Health and Human Services should direct the Administrator, HCFA, to employ data for several years when analyzing outcomes such as mortality rates for small groups of cases across individual hospitals. Hospitals that demonstrate a consistent pattern of observed outcomes that deviate significantly from the expected should be considered the prime candidates for intensified review, as should hospitals whose deviation beyond the range of expected mortality in a single year is based on a number of cases large enough to reduce the effect of random variation.

    Agency Affected: Department of Health and Human Services

  4. Status: Closed - Implemented

    Comments: Sophisticated adjustments have been developed by HCFA staff, but apparently will not be used in new hospital mortality analyses scheduled to be released soon.

    Recommendation: The Secretary of Health and Human Services should direct the Administrator, HCFA, to strengthen HCFA analyses of mortality data by testing and incorporating more sophisticated adjustments that exploit more fully the available diagnostic information.

    Agency Affected: Department of Health and Human Services

  5. Status: Closed - Not Implemented

    Comments: It has been more than 3 years since the report was issued. There is no prospect of further agency implementation.

    Recommendation: The Secretary of Health and Human Services should direct the Administrator, HCFA, to evaluate periodically, through medical record reviews of a nationally representative sample of Medicare patients, the percentage of cases with missing and inaccurate data in the Medicare Statistical System for each of the individual data elements used by HCFA to analyze Medicare outcomes. The results of such assessments should be publicly reported, and corrective action taken for those data elements crucial for reliable outcome analyses. Meanwhile, all analyses of Medicare mortality rates and other outcomes should include an explanation that their findings could be in error by an unknown amount due to potential data inaccuracies.

    Agency Affected: Department of Health and Human Services

 

Explore the full database of GAO's Open Recommendations »

Sep 29, 2016

Sep 28, 2016

Sep 15, 2016

Sep 14, 2016

Sep 12, 2016

Sep 9, 2016

Sep 6, 2016

Looking for more? Browse all our products here