TY  - JOUR
PY  - 2019//
TI  - Predicting prolonged sick leave among trauma survivors
JO  - Scientific reports
A1  - von Oelreich, Erik
A1  - Eriksson, Mikael
A1  - Brattström, Olof
A1  - Discacciati, Andrea
A1  - Strömmer, Lovisa
A1  - Oldner, Anders
A1  - Larsson, Emma
SP  - e58
EP  - e58
VL  - 9
IS  - 1
N2  - Many survivors after trauma suffer from long-term morbidity. The aim of this observational cohort study was to develop a prognostic prediction tool for early assessment of full-time sick leave one year after trauma. Potential predictors were assessed combining individuals from a trauma register with national health registers. Two models were developed using logistic regression and stepwise backward elimination. 4458 individuals were included out of which 488 were on sick leave full-time 12 months after the trauma. One comprehensive and one simplified model were developed including nine and seven predictors respectively. Both models showed excellent discrimination (AUC 0.81). The comprehensive model had very good calibration, and the simplified model good calibration. Prediction models can be used to assess post-trauma sick leave using injury-related variables as well as factors not related to the trauma per se. Among included variables, pre-injury sick leave was the single most important predictor for full-time sick leave one year after trauma. These models could facilitate a more efficient use of resources, targeting groups for follow-up interventions to improve outcome. External validation is necessary in order to evaluate generalizability.<p />  <p>Language: en</p>
LA  - en
SN  - 2045-2322
UR  - http://dx.doi.org/10.1038/s41598-018-37289-w
ID  - ref1
ER  -