Do Current Quality Measures Make a Difference?

The world is ablaze with quality measures, mandated by various provider and other organizations.  Do they really improve people’s health care or health status or medical outcomes?  Studies in the Journal of the American Medical Association provide a decidedly mixed verdict.  (JAMA Articles)  Most quality measures are “process” measures, they purport to describe the “right” way to deliver care, in the belief that certain processes lead to better clinical outcomes.  In the first study, quality measures around pregnancy were examined.  The measures were promulgated by the Joint Commission on Accreditation, one is elective deliveries before 39 weeks of gestation and one is caesarean deliveries in low-risk women.  The researchers compared performance on these measures with maternal or neonatal morbidity.  Hospital level rates on the measures were risk-adjusted.  Data was gathered from 43 hospitals in New York.  There was wide range in performance on the quality measures and on the outcome measures.  The quality measure results were correlated with each other and maternal and neonatal morbidity results were also correlated, but there was no significant interaction between performance on the quality measures and the clinical outcomes.  Actually, of the ten hospitals with the best performance on elective deliveries, only 3 were in the lowest quartile for maternal morbidity.  The variance in performance on the quality measures and in the outcomes suggests that there is room for quality improvement.  But these measures aren’t going to do the job, according to this research.

The second study involved certain skilled nursing facility quality indicators and hospital readmission rates.  Rates of readmission from SNFs tend to be high.  The data sources were various Medicare files from 2008 to 2010, including the Nursing Home Compare site.  The primary outcome was unplanned readmission or death from any cause within 30 days of hospital discharge.  The quality measures were percent of residents admitted after hospital acute care who had delirium in the nursing home, the percent with new or worsening pressure ulcers, the percent with moderate to severe pain, the staffing STAR rating and the site inspection rating.  The results were risk-adjusted.  In the fully-adjusted analysis, there was a statistically significant relationship between better site inspection ratings and lower risk of readmission or death with 30 days, but it was low.  A similar small relationship existed between fewer new or worse pressure ulcers and the outcomes.  Interestingly, smaller and not-for-profit facilities had better performance.  But again, the quality measures had a limited relationship to real clinical outcomes, suggesting that they are directing attention to care processes that may not matter.  Complying with quality measures is not free and they typically take effort away from other quality improvement efforts.  So it is important that these quality indicators are tested for value before they are used.