Introduction
Musculoskeletal and mental disorders are now the leading causes of work disability, with chronic back pain (CBP) as a primary contributor among musculoskeletal conditions. In primary care, the STarT Back Tool (SBT) is used to stratify patients with acute low back pain aiming to reduce transition to CBP by directing management to advice/guidance, normal physiotherapy or psychologically informed physiotherapy. We were interested in whether consideration of genetic predisposition to CBP would improve prediction by SBT. This study investigated the impact of a polygenic risk score (PRS) for CBP on SBT prediction of the incidence rate ratio (IRR) of work disability days and unemployment in a large population-based sample.
Methods
The study used 1,996 participants from the Northern Finland Birth Cohort 1966, who had genotype and SBT questionnaire and registry data available. Outcome data included work disability days (including disability pension and sick leave days) and unemployment days obtained from two national Finnish registers with a two-year follow-up period. CBP PRS was developed from a genome-wide association study of CBP in a large sample from UK Biobank. A zero-inflated negative binomial (ZINB) regression model was employed to evaluate the impact of SBT score (higher values indicate worse prognosis) and CBP PRS on the IRR of outcomes, adjusting for sex, BMI, smoking, education level, and occupational status.
Results
The ZINB model showed that an increase in both SBT score and CBP PRS was significantly associated with a higher IRR of work disability days. Specifically, current and former smoking, lower educational attainment, and higher BMI were significantly correlated with increased work disability IRR. Each one-unit increase in CBP PRS was associated with a 33% increase in work disability days (IRR = 1.332, CI: 1.175-1.510), holding all other model variables constant. By comparison, with including SBT, for each one-unit increase in SBT score, the expected number of work disability days increased by 16% (IRR = 1.164, CI: 1.076-1.256). The ZINB model’s logit component for predicting excess zeros indicated that individuals with higher SBT scores were less likely to report no work disability days (OR = 0.805, CI: 0.752-0.851). No significant association was observed between unemployment days and either SBT or PRS scores.
Discussion
This study underscores the differential impact of CBP screening and genetic predisposition on work disability outcomes. While both the SBT and chronic back pain PRS were associated with incidence of work disability days, the PRS showed a greater effect. These results suggest that genetic predisposition may better capture the various bio-psycho-social components contributing to work disability than questionnaire screening alone. In conclusion, this study highlights the potential benefit of a combined genetic and clinical screening approach for predicting work disability risk.