MBAX 6332 Week 5 Surveillance Bias Policy Brief Example

Policy Brief: Ask Medicare to Report How Hard Hospitals Look for Infections Beside How Many They Find

Student Name

Walden University

MBAX 6332: Health Policy and Systems Thinking

Instructor Name

Month Day, Year

Policy Brief: Ask Medicare to Report How Hard Hospitals Look for Infections Beside How Many They Find

To

Policy Committee, composite State Hospital Association, preparing comments on the annual Medicare inpatient payment proposed rule.

The Problem

The Hospital-Acquired Condition Reduction Program penalizes the lowest-ranked fourth of hospitals on safety indicators and infections. The infection measures count infections a hospital finds, but not how hard it looks for them. A hospital that sends more blood and urine cultures, or tests more patients for C. difficile, will find more infections than a hospital that tests less, even if its patients are no more likely to be harmed. Haut and Pronovost (2011) described this problem as surveillance bias and warned that outcome measures can penalize vigilance. In the program's first year, major teaching hospitals, which tend to test more, were penalized at more than twice the rate of nonteaching hospitals (Rajaram et al., 2015). That pattern does not prove surveillance bias, but it is what surveillance bias would produce.

Why It Matters to Members

Members bear two risks. The first is financial: a year of Medicare inpatient payments cut by one percent, which for a mid-sized hospital can exceed $1 million. The second is clinical: if a hospital's leaders believe that testing less would improve their score, the program creates pressure, however slight, against cultures that clinicians consider appropriate. Infection preventionists in several member hospitals have told the association that they feel that pressure. Merton (1936) warned that purposive action often produces consequences its designers did not intend, and a measure that discourages looking is one.

The Recommendation

Ask the agency, in the association's comments on the proposed rule, to publish each hospital's testing intensity beside its infection measures, such as how many blood cultures and C. difficile tests it runs for every thousand inpatient days, and to study whether the infection scores should be adjusted for it. This does not require a change in the law, which sets the cutoff and the penalty amount but leaves the measures and their presentation to the agency. It would let hospitals, payers and the public see whether a low infection count reflects safe care or little testing, and it would give the agency data to evaluate the concern.

What the Recommendation Does Not Do

It does not ask the agency to stop counting infections or to excuse hospitals with high rates. It does not claim that surveillance bias explains all of the differences between hospitals; sicker patients and real differences in care explain much of it. And it does not address the fixed quartile, which only Congress can change and which the association may wish to raise separately.

Anticipated Objection

The agency may respond that testing data would add reporting burden. The answer is that most hospitals already record cultures and tests in their laboratory systems, and the CDC's surveillance network already collects some testing information for C. difficile. A pilot using existing data would cost little, and Campbell (1979) argued that building evaluation into a program from the start is the best protection against measures that drift from their purpose.

Suggested Comment Language

"The association asks CMS to publish, alongside each hospital's infection measures, a measure of testing intensity, and to evaluate whether infection scores should account for it, so that the program rewards prevention rather than reduced testing."

References

Campbell, D. T. (1979). Assessing the impact of planned social change. Evaluation and Program Planning, 2(1), 67-90. https://doi.org/10.1016/0149-7189(79)90048-X

Haut, E. R., & Pronovost, P. J. (2011). Surveillance bias in outcomes reporting. JAMA, 305(23), 2462-2463. https://doi.org/10.1001/jama.2011.822

Merton, R. K. (1936). The unanticipated consequences of purposive social action. American Sociological Review, 1(6), 894-904. https://doi.org/10.2307/2084615

Rajaram, R., Chung, J. W., Kinnier, C. V., Barnard, C., Mohanty, S., Pavey, E. S., McHugh, M. C., & Bilimoria, K. Y. (2015). Hospital characteristics associated with penalties in the Centers for Medicare & Medicaid Services Hospital-Acquired Condition Reduction Program. JAMA, 314(4), 375-383. https://doi.org/10.1001/jama.2015.8609