Chapter 1.B.5--Evidence-Based Medicine

For additional discussion of John Wennberg's work, see Symposium, Variations Revisited, Health Affairs Web Exclusive (2004).  For more on evidence-based medicine, see Symposium, 24(1) Health Affairs (Jan. 2005).  Atul Gawande's article "The Bell Curve" in the Jan. 31, 2005 New Yorker, is a compelling account of the power of reports cards to motivate improvement in the quality of cystic fibrosis treatment.

Medicare now requires hospitals to participate in public reporting of 10 comparative quality measures (focused on heart failure and pneumonia) in order to obtain full reimbursement.  For more information see Medicare's Hospital Quality Initiative web page.

Questioning the utility of report cards, one study by prominent researchers found that cardiac surgery report cards in New York and Pennsylvania decreased patient and social welfare by prompting physicians and hospitals to avoid sicker or more severe cases.  David Dranove, Daniel Kessler, Mark McClellan, & Mark Satterthwaite, Is More Information Better? The Effects of 'Report Cards' on Health Care Providers, 111 J. Pol. Econ. 555 (2003).  Another study found that a hospital's or surgeon's ranking did not lead to a change in their market share.  Ashish K. Jha & Arnold M. Epstein, The Predictive Accuracy of the New York State Coronary Artery Bypass Surgery Report-Card System, 25(3) Health Aff. 844 (June 2006).

This is an example of a report used by the Pennsylvania Health Care Cost Containment Council, giving statistical ratings for cardiac surgeons who performed bypass operations in 1991.

Patients Who Died
Hospital Physician Practice Group and Surgeons Total Patients Actual Number Expected Range Statistical Rating P- Value
Hospital A   332   4  5.35 - 18.08   +   0.018
  Doctor A   206   2  2.64 - 12.81   +   0.029
  Doctor B   126   2  0.21 - 7.86   =   0.307
Hospital B   323   15  3.91 - 15.30   =   0.061
  Doctor A   58   1  0.00 - 4.50   =   0.632
  Doctor B   7   *NA   *NA   *NA   *NA
  Doctor C   50   3  0.00 - 3.26   =   0.079
  Doctor D   133   8  0.00 - 6.82   -   0.008
  Doctor E   75   3  0.24 - 6.34   =   0.873
Hospital C   467   21 6.24 - 20.09   -   0.026
  Doctor A   23   *NA   *NA   *NA   *NA
  Doctor B   62   1  0.00 - 4.52   =   0.518
  Doctor C   26   *NA   *NA   *NA   *NA
  Doctor D   40   1  0.00 - 2.98   =   0.986
  Doctor E   56   4  0.00 - 4.71   =   0.149
  Doctor F   43   3  0.00 - 29.4   -   0.039
 Statistical Rating Key

+   fewer deaths than expected

-   more deaths than expected

=   the number of deaths was not different than what was expected

*NA   not applicable because fewer than 30 patients were treated

Hospitals and Physicians may have commented on these data. Copies are available upon request.

These are examples of health plan performance measures included in the [HEDIS] Health Plan Employer Data and Information Set
Measurement Goal Category Measure
Quality Preventive medicine

Prenatal care

Acute and chronic disease

Mental health

Childhood immunization
Cholesterol screening
Mammography screening
Pap smears for cervical cancer

Low birthweight
Prenatal care in first trimester

Asthma admissions/readmissions
Diabetic eye exam
Treatment following heart attack

Readmission rate for treatment of
  major affective disorders

Access and satisfaction Access

Member Satisfaction


Percent of members age 42 to 64 with
  plan visit in previous two years
Waiting times
Members who had no problem receiving
  care or referrals

Percent of members who are "very satisfied"
  with plan
Provision of satisfaction surveys


Utilization High occurrence/high cost Frequency and average cost of five selected


Return to home page.