170 Chapter 5 confirm that the relationship of interest differs depending on the average psychological therapy supply available to GPs. Table 3. Heterogeneity in the association of psychological therapy supply and antidepressants prescription rate by average number of psychologists in the local grou Average number of psychologists in the local group Main specification 1st Quartile (lowest) (1) 2nd Quartile (2) 3rd Quartile (3) 4th Quartile (4) (5) Psychologists/100,000 patients Within-term -0.826*** [-1.090,-0.562] -0.120 [-0.480,0.241] 0.277 [-0.393,0.946] -1.115* [-2.170,-0.0601] -0.305 [-0.706, 0.0954] Between-term -0.869 [-3.995,2.256] -9.906** [-17.40,-2.410] -1.804 [-11.00,7.395] -2.274** [-3.684,-0.864] -0.955** [-1.641, -0.269] Year dummies Yes Yes Yes Yes Yes GP, GP practice and Local context * covariates Yes Yes Yes Yes Yes Constant -19.63 [-141.0,101.7] 19.78 [-97.61,137.2] 161.8* [32.14,291.4] 88.22 [-47.88,224.3] 22.45 [-48.56,93.46] N 4,474 4,257 4,191 4,288 17,210 N Clusters 21 27 26 24 R2 within model 0.0966 0.0778 0.117 0.125 0.0756 Mean dep. variable 28.15 28.17 25.80 25.32 26.88 Min-max average psychologists 0-1.15 1.16-1.73 1-73-2.42 2.41-7.37 0-7.37 Notes: The dependent variable is the share of adults with depression in the GP list that has been prescribed antidepressants. Covariates are included in the model demeaned and centered (e.g. for each covariate two terms are included, one with the mean over the study period and other with the difference from the mean in each year). *Some of the covariates are reported at the municipality level, which corresponds roughly to the local group geographical boundaries. 95% confidence intervals in brackets. Standard errors clustered at the local group level (55 local groups in all models). * p < 0.05, ** p < 0.01, *** p < 0.001 For those in the local groups with lowest (1st quartile) and highest (4th quartile) supply, the within term is statistically significant and larger than in the main model ((5) of Table 2). This means that the addition of one psychologist/100,000 patients leads GPs to reduce their prescription in 0.826 pp (95% CI -1.090, -0.562) in local groups deprived from psychologists, and in 1.115 pp (95% CI -2.170, -0.0601) in local groups with the highest supply. These are, respectively, reductions of 2.9% and 4.4% when compared to the average prescription in each subgroup (28.15% and 25.32%).7 We also examined whether our results varied based on GP and practice characteristics presented in the literature as determinants of prescription. Within coefficients in Figure 3. are mostly not statistically significant except for the analysis by GP age category, which 7 We refrain from interpreting the between term in this stratified analysis because there is very little variation in the average number of psychologists in each subgroup, with exception of the 4th quintile.
RkJQdWJsaXNoZXIy MjY0ODMw