External validation and update of a prediction rule for the duration of sickness absence due to common mental disorders

G. Norder, C.A.M. Roelen, J.J.L. Van Der Klink, U. Bültmann, J.K. Sluiter, K. Nieuwenhuijsen

Research output: Contribution to journalArticleScientificpeer-review

13 Downloads (Pure)

Abstract

Purpose:
The objective of the present study was to validate an existing prediction rule (including age, education, depressive/anxiety symptoms, and recovery expectations) for predictions of the duration of sickness absence due to common mental disorders (CMDs) and investigate the added value of work-related factors.

Methods:
A prospective cohort study including 596 employees who reported sick with CMDs in the period from September 2013 to April 2014. Work-related factors were measured at baseline with the Questionnaire on the Experience and Evaluation of Work. During 1-year follow-up, sickness absence data were retrieved from an occupational health register. The outcome variables of the study were sickness absence (no = 0, yes = 1) at 3 and 6 months after reporting sick with CMDs. Discrimination between workers with and without sickness absence was investigated at 3 and 6 months with the area under the receiver operating characteristic curve (AUC).

Results:
A total of 220 (37 %) employees agreed to participate and 211 (35 %) had complete data for analysis. Discrimination was poor with AUC = 0.69 and AUC = 0.55 at 3 and 6 months, respectively. When ‘variety in work’ was added as predictor variable, discrimination between employees with and without CMD sickness absence improved to AUC = 0.74 (at 3 months) and AUC = 0.62 (at 6 months).

Conclusions:
The original prediction rule poorly predicted CMD sickness absence duration. After adding ‘variety in work’, the prediction rule discriminated between employees with and without CMD sickness absence 3 months after reporting sick. This new prediction rule remains to be validated in other populations.
Original languageEnglish
Pages (from-to)202-209
JournalJournal of Occupational Rehabilitation
Volume27
Issue number2
DOIs
Publication statusPublished - 2017

Fingerprint

Area Under Curve
Occupational Health
Depression
Education

Cite this

Norder, G. ; Roelen, C.A.M. ; Van Der Klink, J.J.L. ; Bültmann, U. ; Sluiter, J.K. ; Nieuwenhuijsen, K. / External validation and update of a prediction rule for the duration of sickness absence due to common mental disorders. In: Journal of Occupational Rehabilitation. 2017 ; Vol. 27, No. 2. pp. 202-209.
@article{178ca20fda5b407e82a24dbeba02d7c8,
title = "External validation and update of a prediction rule for the duration of sickness absence due to common mental disorders",
abstract = "Purpose: The objective of the present study was to validate an existing prediction rule (including age, education, depressive/anxiety symptoms, and recovery expectations) for predictions of the duration of sickness absence due to common mental disorders (CMDs) and investigate the added value of work-related factors. Methods: A prospective cohort study including 596 employees who reported sick with CMDs in the period from September 2013 to April 2014. Work-related factors were measured at baseline with the Questionnaire on the Experience and Evaluation of Work. During 1-year follow-up, sickness absence data were retrieved from an occupational health register. The outcome variables of the study were sickness absence (no = 0, yes = 1) at 3 and 6 months after reporting sick with CMDs. Discrimination between workers with and without sickness absence was investigated at 3 and 6 months with the area under the receiver operating characteristic curve (AUC). Results: A total of 220 (37 {\%}) employees agreed to participate and 211 (35 {\%}) had complete data for analysis. Discrimination was poor with AUC = 0.69 and AUC = 0.55 at 3 and 6 months, respectively. When ‘variety in work’ was added as predictor variable, discrimination between employees with and without CMD sickness absence improved to AUC = 0.74 (at 3 months) and AUC = 0.62 (at 6 months). Conclusions: The original prediction rule poorly predicted CMD sickness absence duration. After adding ‘variety in work’, the prediction rule discriminated between employees with and without CMD sickness absence 3 months after reporting sick. This new prediction rule remains to be validated in other populations.",
author = "G. Norder and C.A.M. Roelen and {Van Der Klink}, J.J.L. and U. B{\"u}ltmann and J.K. Sluiter and K. Nieuwenhuijsen",
year = "2017",
doi = "10.1007/s10926-016-9646-1",
language = "English",
volume = "27",
pages = "202--209",
journal = "Journal of Occupational Rehabilitation",
issn = "1053-0487",
publisher = "Springer New York",
number = "2",

}

External validation and update of a prediction rule for the duration of sickness absence due to common mental disorders. / Norder, G.; Roelen, C.A.M.; Van Der Klink, J.J.L.; Bültmann, U.; Sluiter, J.K.; Nieuwenhuijsen, K.

In: Journal of Occupational Rehabilitation, Vol. 27, No. 2, 2017, p. 202-209.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - External validation and update of a prediction rule for the duration of sickness absence due to common mental disorders

AU - Norder, G.

AU - Roelen, C.A.M.

AU - Van Der Klink, J.J.L.

AU - Bültmann, U.

AU - Sluiter, J.K.

AU - Nieuwenhuijsen, K.

PY - 2017

Y1 - 2017

N2 - Purpose: The objective of the present study was to validate an existing prediction rule (including age, education, depressive/anxiety symptoms, and recovery expectations) for predictions of the duration of sickness absence due to common mental disorders (CMDs) and investigate the added value of work-related factors. Methods: A prospective cohort study including 596 employees who reported sick with CMDs in the period from September 2013 to April 2014. Work-related factors were measured at baseline with the Questionnaire on the Experience and Evaluation of Work. During 1-year follow-up, sickness absence data were retrieved from an occupational health register. The outcome variables of the study were sickness absence (no = 0, yes = 1) at 3 and 6 months after reporting sick with CMDs. Discrimination between workers with and without sickness absence was investigated at 3 and 6 months with the area under the receiver operating characteristic curve (AUC). Results: A total of 220 (37 %) employees agreed to participate and 211 (35 %) had complete data for analysis. Discrimination was poor with AUC = 0.69 and AUC = 0.55 at 3 and 6 months, respectively. When ‘variety in work’ was added as predictor variable, discrimination between employees with and without CMD sickness absence improved to AUC = 0.74 (at 3 months) and AUC = 0.62 (at 6 months). Conclusions: The original prediction rule poorly predicted CMD sickness absence duration. After adding ‘variety in work’, the prediction rule discriminated between employees with and without CMD sickness absence 3 months after reporting sick. This new prediction rule remains to be validated in other populations.

AB - Purpose: The objective of the present study was to validate an existing prediction rule (including age, education, depressive/anxiety symptoms, and recovery expectations) for predictions of the duration of sickness absence due to common mental disorders (CMDs) and investigate the added value of work-related factors. Methods: A prospective cohort study including 596 employees who reported sick with CMDs in the period from September 2013 to April 2014. Work-related factors were measured at baseline with the Questionnaire on the Experience and Evaluation of Work. During 1-year follow-up, sickness absence data were retrieved from an occupational health register. The outcome variables of the study were sickness absence (no = 0, yes = 1) at 3 and 6 months after reporting sick with CMDs. Discrimination between workers with and without sickness absence was investigated at 3 and 6 months with the area under the receiver operating characteristic curve (AUC). Results: A total of 220 (37 %) employees agreed to participate and 211 (35 %) had complete data for analysis. Discrimination was poor with AUC = 0.69 and AUC = 0.55 at 3 and 6 months, respectively. When ‘variety in work’ was added as predictor variable, discrimination between employees with and without CMD sickness absence improved to AUC = 0.74 (at 3 months) and AUC = 0.62 (at 6 months). Conclusions: The original prediction rule poorly predicted CMD sickness absence duration. After adding ‘variety in work’, the prediction rule discriminated between employees with and without CMD sickness absence 3 months after reporting sick. This new prediction rule remains to be validated in other populations.

U2 - 10.1007/s10926-016-9646-1

DO - 10.1007/s10926-016-9646-1

M3 - Article

VL - 27

SP - 202

EP - 209

JO - Journal of Occupational Rehabilitation

JF - Journal of Occupational Rehabilitation

SN - 1053-0487

IS - 2

ER -