Quality achievement and disease prevalence in primary care predicts regional variation in renal replacement therapy (RRT) incidence: an ecological study

Diabetes Meillitus (DM) and hypertension (HT) are important causes of end-stage renal disease (ESRD) and renal replacement therapy (RRT) is the standard active treatment. Financially, incentivized quality initiatives for primary care include pay-for-performance (P4P) in DM and HT. Our aim was to examine any effect of disease prevalence and P4P on RRT incidence and regional variation. The incidence of RRT, sex and ethnicity data and P4P disease register and achievement data were obtained for each NHS locality. We calculated correlation coefficients for P4P indicators since 2004/05 and socio-demographic data for these 152 localities. We then developed a regression model and regression coefficient (R-2) to assess to what extent these variables might predict RRT incidence. Many of the P4P indicators were weakly but highly significantly correlated with RRT incidence. The strongest correlation was 2004/05 for DM prevalence and 2006/07 for HT quality. DM prevalence and the percentage with blood pressure control in HT target (HT quality) were the most predictive in our regression model R-2 = 0.096 and R-2 = 0.085, respectively (P < 0.001). Combined they predicted a fifth of RRT incidence (R-2 = 0.2, P < 0.001) while ethnicity and deprivation a quarter (R-2 = 0.25, P < 0.001). Our final model contained proportion of population > 75 years, DM prevalence, HT quality, ethnicity and deprivation index and predicted 40% of variation (R-2 = 0.4, P < 0.001). Our findings add prevalence of DM and quality of HT management to the known predictors of variation in RRT, ethnicity and deprivation. They raise the possibility that interventions in primary care might influence later events in specialist care.