June 2017 - Quality in primary care
On Thursday June 1st, Rita Santos (University of York) held a presentation about her latest research: Quality in primary care: Is it possible to identify peer effects among English GP practices? The seminar was organised by Centre of Expertise Healthwise and the signature area Individual Health & The Economic Environment.
About Rita Santos
Rita Santos is a research fellow at the Centre for Health Economics at the University of York. She is also a member of the Health Policy Team.
Rita’s previous work focused on topics such as GP gender pay differences, GP practice choice, Hospital competition and GP practices competition. Currently she is working on measuring and explaining primary care quality variation under an NIHR doctoral fellowship.
Recent research includes:
- Equity in primary care
- Incentives for quality in primary care
- Geographical distribution of primary care quality
- Impact of competition on GP practice quality
- Impact of competition on hospital quality
- The application of micro and spatial econometrics to health economics
Rita holds a BA in Economics and an MSc in Applied Economics from the Faculty of Economics, University of Coimbra. She worked at the Department of Geography, University of Coimbra (2004-2009) where she conducted research into environmental impacts on health outcomes.
Quality in primary care: Is it possible to identify peer effects among English GP practices?
Quality is one of the most important features of health services. We measure quality by the number of emergency admissions in a practice for conditions where better care should reduce admissions (Ambulatory Care Sensitive Conditions-ACSCs). We test if it is possible to distinguish endogenous and contextual effects, i.e. whether the correlation amongst the quality of GP practices in an area is due to the effect of peer groups of practices or whether it is due to common unobserved factors affecting all practices in an area. Manski (1993, 2000) has stressed the difficulties of identifying peer effects. However, recent work by Lin (2010) and Lee et al (2010) suggests that spatial methods can be used to test for peer effects.
We have a panel of data on all English GP from 2006/7 to 2011/12 including ACSCs, GP practice characteristics, patient demographics, morbidity, and area deprivation.
We estimate a Spatial Durbin Model to distinguish between the endogenous and contextual effects while including group fixed effects to control for correlated effects and unobservables in a peer group.
A higher number of GPs per practice and more experienced GPs decrease the number of ACSCs. When we allow for endogenous and contextual effects the coefficient on the number of GPs is positive and significant, which is unexpected. We discuss whether this counter-intuitive result casts doubts on the use of spatial methods to distinguish practice, contextual and endogenous effects, or if when we introduce contextual effects we are actually capturing an area effect.
Authors: Rita Santos, Hugh Gravelle and Nigel Rice
|Last modified:||11 June 2018 09.59 a.m.|