Association of short poor work ability measure with increased mortality risk: a prospective multicohort study

Abstract

Objectives: To examine whether a single-item measure of self-rated work ability predicts all-cause mortality in three large population-based samples collected in 1978-1980, 2000 and 2017.

Setting: A representative sample of the population of Finland.

Participants: The study population comprised 17 178 participants aged 18 to 65 from the population-based Mini-Finland, Health 2000 and FinHealth 2017 cohort studies, pooled together. In all cohorts, self-rated work ability was assessed at baseline (1978-80, 2000-2001 and 2017) using three response alternatives: completely fit (good work ability), partially disabled (limited work ability) and completely disabled (poor work ability) for work.

Primary and secondary outcome measures: All-cause mortality from national registers. Cox proportional hazards models were adjusted for socioeconomic characteristics, lifestyle factors, self-rated health and mental health problems.

Results: Of the participants, 2219 (13%) were classified as having limited and 991 (5.8%) poor work ability and 246 individuals died during the 4 year follow-up. The age- and sex-adjusted HR for mortality risk was 7.20 (95% CI 5.15 to 10.08) for participants with poor vs good work ability and 3.22 (95% CI 2.30 to 4.43) for participants with limited vs good work ability. The excess risk associated with poor work ability was seen in both genders, all age groups, across different educational levels, self-rated health levels and in those with and without mental health problems. The associations were robust to further adjustment for education, health behaviours, self-rated health and mental health problems. In the multivariable analyses, the HR for mortality among those with poor vs good work ability was 5.75 (95% CI 3.59 to 9.20).

Conclusions: One-item poor self-rated work ability -measure is a strong predictor of increased risk of all-cause mortality and may be a useful survey-measure in predicting severe health outcomes in community-based surveys.

Keywords: EPIDEMIOLOGY; OCCUPATIONAL & INDUSTRIAL MEDICINE; PUBLIC HEALTH.

Figure 1

Receiver operating characteristic (ROC) curves of sensitivity and specificity of the model (work ability predicting mortality risk) within four follow-up years including age and sex (first row columns) and age, sex and work ability (second row columns).

Figure 2

Net reclassification indexes (NRI) of all included variables predicting mortality risk (the accuracy of the predicted classification of mortality during the 4 year follow-up).