166 Chapter 5 Table 1. Descriptive statistics of the study population. Mean Standard deviation Min Max Overall between within Ratio Nurses to GP 1.117 0.374 0.348 0.158 0.312 6.250 Incidence of depression [%] 9.235 3.905 3.372 2.137 0.328 56.281 Local group*-level variables Psychologists/100,000 patients 1.976 1.314 1.246 0.476 0.000 7.586 Integrated Care (ULS) [0,1] 0.126 0.332 0.337 0.000 0.000 1.000 Population density* 1543.426 2149.705 2174.459 177.349 4.154 14780.560 City [0,1] 0.449 0.497 0.496 0.026 0.000 1.000 Town or suburb [0,1] 0.382 0.486 0.483 0.054 0.000 1.000 Rural area [0,1] 0.169 0.374 0.377 0.051 0.000 1.000 Average monthly earnings (employed)* 1051.025 220.702 221.050 30.594 702.200 2331.200 % Unemployment* 7.091 2.536 2.286 1.146 2.400 16.100 N** 17,210 Notes: # Since a reform in 2006, the previously existing GP practices are called personalised health care units (PHCU) and coexist in the Portuguese ecosystem with two types of performance-incentivised team-based practices: family health units (FHU) of type A and B (FHU-B). In the initial stage, all newly created FHUs were of type A and had a team-based organization and group incentives. FHU-A could then apply to migrate to FHU-B, with GP individual pay-for-performance incentives and more stringent performance targets.† All the GP practices were subjected to a reform in 2006, which is therefore considered as starting year. *Some of the covariates are reported at the municipality level, which corresponds roughly to the local group’s geographical boundaries. **GP-year observations: 2015=4,045; 2016=4,358; 2017=4,612; 2018=4,195; corresponding to 5,359 unique GPs working in 888 GP practices of 55 local groups. Figure 1 provides some insight into the time and regional variation of our key variables. The antidepressant prescription share fell from 27.6% to 25.7% from 2015 to 2018; the top panel shows considerable asymmetries across regions. These patterns are only partially explained differences in the prevalence of depressive disorders (Figure A2 of the appendix). Certain regions with low prevalence have higher rates of prescription, notably some regions in the great Lisbon area. The maps also attest to the persistence of the antidepressant prescription over time. The average number of psychologists in the local groups increased during the study period, from 1.8 (2015) to 2.3 (2018) FTE/100,000 patients registered in the local groups. However, the bottom panel of Figure 1 shows that this increase was not consistent across the country, with some regions still having 0 psychologists in their local group in 2018, or other regions in which the number decreased. Figure 2 sheds some initial insight into the relationship between GP antidepressant prescription and the number of psychologists in the local group. It shows that there is considerable variation in antidepressant prescription across local groups and that local groups where GP prescribe antidepressants to less than 20% of the patients diagnosed with depression are also more often in the highest quartile of average psychologist supply.
RkJQdWJsaXNoZXIy MjY0ODMw