Impaired sleep predicts onset of low back pain and burnout symptoms: evidence from a three-wave study

Key points

1. 01Sleep problems predicted new LBP two years later (beta = .15, p = .011) and new burnout two years later (beta = .24, p < .001).
2. 02Sleep problems also predicted burnout at three years (beta = .18, p = .002) but not LBP at three years (beta = .07, p = .127).
3. 03The sleep-to-burnout risk was much stronger in older workers (age 45+) than younger ones (beta = .42 vs. .13, p < .001) at two years.
4. 04No significant cross-lagged effect was found from LBP to burnout or from burnout to LBP in the full sample; burnout predicted later LBP only in younger women.
5. 05All participants were free of LBP and emotional exhaustion at baseline, ruling out reverse causation at study entry.

How it was conducted

Design

Three-wave prospective longitudinal questionnaire study with baseline (2003), first follow-up (2005, ~24 months), and second follow-up (2006, ~36 months)

Participants

405 working adults in Switzerland, initially free of LBP and emotional exhaustion; 38.8% women, 46.4% aged 45 or older

Sleep measure

Two-item brief Sleep Standard Evaluation Questionnaire: quality of sleep and daytime tiredness due to sleep problems (0-6 scale each)

Primary outcomes

New-onset self-reported LBP (three-item composite: treatment status, pain drawing, pain intensity) and burnout (three single items covering emotional exhaustion, depersonalization, reduced performance)

Analysis

Longitudinal structural equation modelling (SEM) with latent variables in AMOS 24.0; model fit assessed by RMSEA, CFI, and SRMR; subgroup comparisons by sex and age using Fisher Z tests

What they found

• Sleep problems predicted LBP at 2-year follow-up: beta = .15, p = .011
• Sleep problems predicted burnout at 2-year follow-up: beta = .24, p < .001
• Sleep problems predicted burnout at 3-year follow-up: beta = .18, p = .002
• Sleep problems did not significantly predict LBP at 3-year follow-up: beta = .07, p = .127
• Overall model fit was good: RMSEA = .05 (90% CI [.03, .06]), CFI = .97, SRMR = .06
• Sleep-to-burnout path at 2 years was stronger in older (age 45+) than younger participants: beta = .42 vs. .13, p < .001
• Sleep-to-burnout path at 3 years was stronger in men than women: beta = .25 vs. .03, p = .028
• Sleep-to-LBP path at 2 years was stronger in older women than younger women: beta = .39 vs. .07, p = .036; no significant age difference in men
• Cross-lagged path from burnout at 2 years to LBP at 3 years was significant only in younger women: beta = .23, p = .035 (n = 81)
• Pearson correlation between baseline sleep problems and LBP at follow-up 1: r = .11, p = .021; at follow-up 2: r = .11, p = .027
• Pearson correlation between baseline sleep problems and burnout at follow-up 1: r = .19, p < .001; at follow-up 2: r = .23, p < .001

Limitations

• All measures relied on self-report only, which may inflate correlations through common-method variance; objective sleep actigraphy was not used.
• Burnout dimensions were each assessed with a single item, which limits measurement reliability despite evidence that single items can be valid for these constructs.
• The sample excluded 77.9% of the original population-based cohort (those with existing LBP or exhaustion, retirees, unemployed), which may limit generalizability.
• The study cannot confirm whether sleep problems worsen after LBP or burnout onset at follow-up, as sleep was only assessed at baseline.

Why it matters

For patients

If you have trouble sleeping, it may raise your risk of developing back pain or burnout over the next two to three years, so addressing sleep problems early could be protective.

For clinicians

Screen working patients for sleep quality as an early modifiable risk factor for both LBP and burnout, particularly in workers aged 45 and older where the burnout risk is markedly higher.

For readers

This longitudinal study provides prospective evidence that sleep impairment precedes rather than follows the onset of LBP and burnout, supporting sleep as a genuine upstream risk factor rather than a byproduct.