Productivity and its Drivers in Finnish Primary Care 1988-2003

We measure productivity changes of primary care in Finland between 1988 and 2003 as a ratio of key services produced and real operating costs. In the second stage we estimate a truncated regression model that quantifies the contribution of certain internal and exogenous factors to productivity. We use newly developed techniques to correct asymptotic bias in non-parametric efficiency scores and bootstrap the confidence intervals for the explanatory model parameter estimates. The bias accounts on average 2.8 percent decrease in efficiency level. From 1997 to 2003 the average productivity declined 13.7 percent; the result is insensitive to estimated bias. Even if standard parametric confidence intervals do not generally apply when efficiency scores are regressed, our bootstrapped intervals are almost equal to parametric ones. Of the correlates used the increased income subject to municipal taxation accounted for three percentage points of the productivity decrease. The correlates, that are expected to decrease the need of primary care services, had a negative impact on productivity, implying that health centres have not been able to adjust their resource usage correspondingly. Organisational changes that have taken place within primary care have not resulted in desired productivity improvements.