Model changes: time periods | Time period 3 months | Time period 5 months | Time period 6 months | Time period 12 months | ||||
First visit only | First and subsequent visits | First visit only | First and subsequent visits | First visit only | First and subsequent visits | First visit only | First and subsequent visits | |
Intervention effect (first visit) | −0.168** (0.073) | −0.224*** (0.081) | −0.114* (0.061) | −0.157** (0.066) | −0.051 (0.058) | −0.107* (0.063) | 0.019 (0.055) | −0.011 (0.061) |
Intervention effect (subsequent visits) | 0.035* (0.021) | 0.026* (0.016) | 0.033** (0.015) | 0.015 (0.013) | ||||
N | 1428 | 1350 | 1270 | 932 |
Model changes: data analysis method and sample | Fixed-effects model without control variables | Negative binomial regression with clustered SEs | Sample with all observations | Sample excluding outliers | ||||
First visit only | First and subsequent visits | First visit only | First and subsequent visits | First visit only | First and subsequent visits | First visit only | First and subsequent visits | |
Intervention effect (first visit) | −0.088 (0.069) | −0.138* (0.075) | −0.093 (0.074) | −0.117 (0.08) | −0.016 (0.037) | −0.017 (0.04) | −0.103 (0.067) | −0.166** (0.074) |
Intervention effect (subsequent visits) | 0.031* (0.019) | 0.014 (0.021) | 0.000 (0.008) | 0.038** (0.018) | ||||
N | 1388 | 1664 | 3094 | 1380 |
We report the main coefficients of interest from the sensitivity analyses (all other variables from main analyses included but not shown) and present robust SEs in parentheses. Statistical significance levels: *p<0.10, **p<0.05, ***p<0.01.