Table 2.
Triple-differences marginal effects on the impact of the priority review voucher on clinical trial registrations
| Variable | Model 1 | Model 2 | ||
|---|---|---|---|---|
| Marginal effect (SE) | p-Value | Marginal effect (SE) | p-Value | |
| After × Eligible × ClinicalTrials.gov | 0.41 (0.40) | 0.31 | ||
| Eligible × ClinicalTrials.gov × Lag | ||||
| 0–1 | 0.44 (0.57) | 0.45 | ||
| 2–3 | 0.40 (0.56) | 0.49 | ||
| 4–5 | − 0.73 (1.03) | 1.35 | ||
| 6–7 | 0.19 (0.72) | 0.81 | ||
| 8–9 | − 0.08 (0.97) | 1.06 | ||
| 10–11 | − 0.34 (0.90) | 1.24 | ||
| Registry × Disease fixed effects | Yes | Yes | ||
| Year fixed effects | Yes | Yes | ||
| No. of observations (N) | 2941 | 2941 | ||
All regressions include the control variables, as well as registry_disease and year fixed effects (Poisson model). In Model 1, After × Eligible × ClinicalTrials.gov is the triple-difference estimator and captures the effect of the PRV on clinical trials registration. In Model 2, lags are bundled together by blocks of 2 years. The first block (Lag0–1) includes the policy year and the year after, and the second block (Lag2–3) includes years 2 and 3 post policy implementation, etc. Fixed effects capture time-invariant unobserved characteristics that are specific to the registry, disease and year
SE standard error, PRV priority review voucher