New Report Helps Policymakers Understand Health Services Research Findings

10/7/2002

From: Hollis Hope or Deborah Rogal, 202-292-6700 both of AcademyHealth

WASHINGTON, Oct. 7 -- Health services researchers generate a torrent of information about pressing health policy issues, but making their findings accessible to decision-makers is not always easy. Even when the research finds its way into the hands of decision-makers, the statistical analyses may be misinterpreted.

In a new report from The Robert Wood Johnson Foundation's Changes in Health Care Financing and Organization (HCFO) initiative, Bryan Dowd, Ph.D., and Robert Town, Ph.D., of the University of Minnesota discuss two specific analytic problems that can lead policymakers (and analysts) to draw incorrect conclusions from simple data analyses: omitted variables bias and reverse causality. They also discuss the common econometric solutions to those problems.

"Policymakers face a formidable problem dealing with the mass of data that crosses their desk, including study findings," says Dowd.

"Sometimes those studies are poorly done, or the results are subject to selective interpretation. This paper provides some background material that we hope will help policymakers and their staffs identify flaws in studies, or understand the methods that are being used."

The report grew from a HCFO meeting in which Dowd explained causality in the context of how managed care spillover might be used to inform policy discussions. He pointed out that while research and policy development often require using imperfect information, it is important for those using research findings to be able to identify and account for such imperfections.

"The greater the potential harm from bad decisions, or the potential gain from deception, the more carefully policymakers must guard against it," Dowd concludes.

To download "Does X Really Cause Y?" visit the HCFO Web site at www.hcfo.net. For hard copies, e-mail colleen.king@academyhealth.org.

AcademyHealth (www.academyhealth.org) serves as the national program office for the HCFO program. The program's mission is to serve as a bridge between the policy and research communities, funding the production of useable and timely information on health care policy and market developments for dissemination to stakeholders in the public and private policy arenas.