About method

Data
The retrospective cohort analysis was conducted using the nation-wide registries covering the period 1994-2015: the Danish National Patient Registry (NPR) comprising of 6.9 M patients and the Cancer Registry of 0.7 M cancer patients. We did not include data before 1994, since these were coded in the older version ICD-8. Patient-specific information could be linked between the Cancer Registry and NPR by using the person-specific Danish registration numbers.

Method
We applied a previously published method (A. B. Jensen et al., 2014) to calculate significant directional diagnosis pairs. The strength of correlation between a diagnosis pair was estimated using relative risk (RR) scores within 15 years. The RR scores and associated p-values were calculated using a sampling approach between case and control populations, which were matched by age, gender, type of hospital encounter (inpatient, outpatient and emergency room) and dispatch week. We selected all significant directional diagnosis pairs with a RR > 1 that comprise of a non-cancer diagnosis leading to a cancer diagnosis (D1 → C1). For each diagnosis pair, the differential time (in years) was calculated between D1 and C1 for each patient. The average differential time across patients was then used as an estimate of the time occurrence of D1 before C1.