Extended Data Table 5 Negative outcome controls

From: Quasi-experimental evaluation of a nationwide diabetes prevention programme

  1. *Effect estimate can be interpreted as the difference in number of events during the follow-up period at the threshold.
  2. †Sample restricted to those without any previous diagnosis of cancer or dementia, respectively. While healthy lifestyle behaviours are assumed to be protective factors for cancer and dementia onset, the median follow-up time (29 months) is sufficiently short to assume that any plausible effects are negligible.
  3. IUD = Intra-uterine device. Regression coefficients from local linear regression with MSE-optimal bandwidth, 95% confidence intervals (95% CI), P values and sample size. This table shows the main regression discontinuity analysis (primary cohort) for health outcomes which we did not expect to be plausibly affected by referral to the NHS DPP (‘negative outcome controls’). Effect estimates can be interpreted as the difference in percentage points at the threshold. The results for routine breast and cervical cancer screening as well as contraceptive use are restricted to women only.
Back to article page