Persistence, not avoidance, is associated with low back pain: an observational cohort study

Key points

1. 01Persistence behaviour at baseline predicted a 3.3-point increase in pain intensity over 12 months when combined with state anxiety
2. 02Avoidance behaviour was not associated with pain persistence and naturally decreased as pain improved
3. 0339% of participants showed persistence patterns at baseline, compared to only 6% showing avoidance
4. 04Movement control impairment (MCI) was common but did not independently predict ongoing pain intensity
5. 05Screening for activity patterns early in an acute LBP episode may help tailor treatment

How it was conducted

Design

Observational prospective cohort study

Participants

165 adults aged 18-65 with acute low back pain (onset within 4 weeks after a pain-free period of at least 3 months)

Follow-up

5 time points: baseline, 2, 3, 6, and 12 months

Primary outcome

Pain intensity (numeric rating scale 0-10) over 12 months

Key exposures

Avoidance and persistence behaviour (Avoidance-Endurance Questionnaire), movement control impairment (6-test clinical battery), state anxiety (STAI-S)

Analysis

Linear mixed-effects regression models with random intercepts and slopes, adjusted for baseline pain, state anxiety, sex, and age

What they found

• Pain intensity decreased significantly over 6 months: 5.2 at baseline, 3.2 at 2 months, 2.7 at 3 months, 2.2 at 6 months (F(3.00, 414) = 108.1, p < 0.001, partial eta2 = 0.44)
• In the adjusted model, a 1-point higher baseline AEQ-Persistence score was associated with a 3.072-point increase in pain intensity over time (95% CI 0.07-6.07, p = 0.05), interacting with state anxiety
• Avoidance behaviour was not associated with pain intensity in the adjusted model (coefficient -1.48, 95% CI -3.60 to 1.24, p = 0.21)
• AEQ-Persistence scores remained stable over time (F(3, 270) = 1.57, p = 0.19, partial eta2 = 0.02), while avoidance scores decreased significantly (F(3.00, 285) = 28.8, p < 0.001, partial eta2 = 0.28)
• MCI scores did not change significantly over 6 months (from 2.1 to 1.7; chi2(3) = 7.65, p = 0.54) and were not associated with pain intensity in any adjusted model
• Female sex was associated with higher pain intensity in all adjusted models (avoidance model: coefficient 0.82, 95% CI 0.11-1.53, p = 0.03)
• At 12 months, approximately two-thirds of participants still experienced some LBP, though many at low intensity
• Activity pattern classification at baseline: 49% pacing, 39% persistence, 6% avoidance, 6% not reported

Limitations

• Interrater reliability of the movement control impairment clinical tests was not formally assessed, introducing potential variability
• Treatment received during the study was not controlled for in the analysis, which may have confounded results
• Recruitment from both clinical settings and university campuses may have introduced population variability, limiting generalisability
• The operationalisation of pacing behaviour was vague and may not reliably distinguish it from avoidance or persistence

Why it matters

For patients

Patients who keep pushing through their normal activities without adjusting for pain, especially when feeling distressed, may be at greater risk of pain continuing beyond the acute phase.

For clinicians

Screening for persistence patterns early in an acute LBP episode is important because persistence, not avoidance, was the activity behaviour most associated with ongoing pain, and it remained stable over time without apparent self-correction.

For readers

This cohort study challenges the dominance of fear-avoidance models in LBP research by showing persistence behaviour is a clinically relevant and underappreciated risk factor for pain persistence.