Police reported road traffic casualties and Hospital patients of road traffic crashes were linked and filtered to patients with severity MAIS2+. -Some patients were not reported by the police, -Some linked patients conflict with casualties in their mode of transport -Some patients have missing values -Only for linked patients a region of the crash is present.

The challenge is to make an estimate of the total number of casualties for each of the covariates present in this dataset, preferably with a margin.

The dataset contains info for three years: 1994, 2009 and 2013 by Year, Mode_hospital, Mode_police, MaxAIS, ProvZH_hospital, ProvAcc_police, ExtCause, Gender, Agegroup, Number

The sum of Number is 46884, distributed over the years and Linked=Yes/No:

Year ------ Yes ------- No ------ Sum 1994 --- 9.975 ---- 6.988 -- 16.963 2009 --- 4.440 -- 10.446 -- 14.886 2013 --- 2.617 ---12.418 -- 15.035 sum -- 17.032 -- 29.852 -- 46.884

Especially in the last year the intersection between the two sources is small because of poor police reporting. The hospital file is assumed to be complete, although some patients after traffic injury do not receive a proper ExternalCause, so these can only be identified by linking.