Abstracts of the Third Global Symposium on Health Systems Research: 30 September – 3 October 2014 The Science and Practice of People-Centred Health Systems Cape Town South Africa Thursday 2 October 2014 16.30 - 18.00 Auditorium 1 Long-term effects of a randomized policy experiment on quality improvement in the Philippines Quimbo, Stella 1 ; Wagner, Natascha 2 ; Florentino, Jhiedon 3 ; Solon, Orville 1 ; Peabody, John 4 1 University of the Philippines School of Economics; Philippines 2 International Institute of Social Studies, Erasmus University; The Netherlands 3 UPecon Foundation; Philippines 4 Qure Healthcare; USA ABSTRACT We investigate the possible long-term effects of a randomized policy experiment conducted in the Philippines between 2003 and 2008. The policy experiment, known as the Quality Improvement Demonstration Study (QIDS), consisted of a large-scale community level policy intervention in 30 hospitals. Five years after interventions were discontinued, we re-surveyed doctors who had previously participated in the experiment to determine if the intervention effects continue to persist. The hospitals had been randomly assigned into two intervention sites plus control sites. The first policy intervention was an expanded health insurance program targeted to young children. The second was a pay-for- performance scheme to hospital staff conditioned on improving the quality of care. The objective was to evaluate the effects of these two interventions on the quality of hospital care and the health status of children. After three years, both QIDS interventions were found to have short-term impacts. For our follow-up study, we collected data from February to August 2013 on previously participating doctors to assess whether hospitals in the treatment sites continue to have better quality compared to those in the control sites. We specifically test whether quality measurement and performance feedback, last documented in August 2008, might provide a sufficient inoculum that persisted over time in the intervention sites relative to controls. A hypothesis is that clinical skills, once acquired, would become routine and not deteriorate establish a culture of performance and quality expectation. Alternatively, they may also diminish with disruptions in measurement, feedback, and incentive. We found that the intervention sites continued to have significantly higher quality compared to the control sites. The previously documented quality improvement in intervention sites appears to be sustained; moreover it was subject to a very low (less than 1 percent per year) rate of decay in quality scores.