Multicentre validation of frequent sickness absence predictions

Area of Science:

• Occupational Health
• Epidemiology
• Predictive Modeling

Background:

• Previous research identified age, self-rated health (SRH), and prior sickness absence (SA) as predictors of frequent SA.
• A need exists to validate and adapt this prediction model for practical clinical application.

Purpose of the Study:

• To validate a prediction model for frequent sickness absence (SA) in a multicentre study.
• To develop and assess the clinical utility of a prediction rule derived from the model.

Main Methods:

• A multicentre study involving 2562 care of the elderly workers in Aarhus, Denmark.
• Data included baseline age, self-rated health (SRH), and prior SA to predict frequent SA over a 1-year follow-up.
• A prediction rule, 'SRH-prior SA', was derived and tested across 13 centres.

Main Results:

• The prediction model showed accurate predictions in 4 of 13 centres and good/fair discrimination in 10 centres.
• The 'SRH-prior SA' rule identified high-risk workers with sensitivities ranging from 0.17-0.54 and specificities of 0.86-0.96.
• Positive predictive values for the rule varied between 0.54-0.87 across the centres.

Conclusions:

• The prediction model effectively discriminated between workers at high and low risk of frequent SA in most participating centres.
• The 'SRH-prior SA' prediction rule is suitable for clinical use to identify individuals at high risk for frequent sickness absence.