The effect of a quality improvement initiative on the quality of other aspects of health care: the law of unintended consequences?

Abstract

Problem: Policymakers and clinicians are concerned that initiatives to improve the quality of care for some conditions may have unintended negative consequences for quality in other conditions.

Objective: We sought to determine whether a practice redesign intervention that improved care for falls, incontinence, and cognitive impairment by an absolute 15% change also affected quality of care for masked conditions (conditions not targeted by the intervention).

Design, setting, and participants: Controlled trial in 2 community medical groups, with 357 intervention and 287 control patients age 75 years or older who had difficulty with falls, incontinence, or cognitive impairment.

Intervention: Both intervention and control practices implemented case-finding for target conditions, but only intervention practices received a multicomponent practice-change intervention. Quality of care in the intervention practices improved for 2 of the target conditions (falls and incontinence).

Main outcome measures: Percent of quality indicators satisfied for a set of 9 masked conditions measured by abstraction of medical records.

Results: Before the intervention, the overall percent of masked indicators satisfied was 69% in the intervention group and 67% in the control group. During the intervention period, these percentages did not change, and there was no difference between intervention and control groups for the change in quality between the 2 periods (P=0.86). The intervention minus control difference-in-change for the percent of masked indicators satisfied was 0.2% (bootstrapped 95% confidence interval, -2.7% to 2.9%). Subgroup analyses by clinical condition and by type of care process performed by the clinician did not show consistent results favoring either the intervention or the control group.

Conclusion: A practice-based intervention that improved quality of care for targeted conditions by an absolute 15% change did not affect measurable aspects of care on a broad set of masked quality measures encompassing 9 other conditions.