Effects of lifestyle interventions on the improvement of chronic non-specific low back pain

Key points

1. 01Cognitive therapy combined with functional exercise programs ranked highest for reducing pain intensity; functional exercise programs ranked highest for reducing functional disability.
2. 02When both outcomes were assessed together, functional exercise programs followed by aquatic and aerobic exercise produced the best combined results.
3. 03Simply providing health information to patients was among the least effective approaches and appears insufficient without professional follow-up and motivational support.
4. 04Motivation, positive feedback, and ongoing professional follow-up were identified as key factors in successful lifestyle interventions.
5. 05High heterogeneity (I2 up to 99.54% for pain intensity) limits confidence in precise effect size estimates across comparisons.

How it was conducted

Design

Systematic review and network meta-analysis (univariate and multivariate NMA); registered with PROSPERO (CRD42022315090); conducted per PRISMA guidelines

Databases searched

PubMed, Web of Science, Scopus, SportDiscus; search ended 19 March 2022

Included studies

20 studies for qualitative analysis; 16 RCTs included in quantitative NMA (moderate-to-high methodological quality by PEDro scale, mean score 6.5/10)

Participants

2,525 adults with chronic non-specific low back pain (at least 3 months); mean age 46.4 +/- 7.7 years; 36.0% male, 61.9% female

Interventions compared

13 intervention categories including cognitive therapy, functional exercise programs, lumbar stabilization, resistance training, aerobic exercise, aquatic exercise, yoga, tai chi, qigong, Pilates, sedentary behavior reduction, health information, and standard care

Primary outcomes

Pain intensity (VAS or NRS 0-10) and functional disability (ODI or RMDQ); standardized mean difference (SMD) and mean difference (MD) used

What they found

• For pain intensity (univariate NMA), the highest P-score ranking was: cognitive therapy first, lumbar stabilization exercise second, resistance training third; cognitive therapy vs. health information showed a significant difference of -19.42 (95% CI -35.96, -2.87) in favor of cognitive therapy.
• For functional disability (univariate NMA), the highest P-score ranking was: functional exercise programs first, aerobic training second, standard care third; functional exercise programs vs. control showed a significant difference of -38.02 (95% CI -53.84, -22.20) in favor of functional exercise programs.
• In the multivariate NMA combining both outcomes, functional exercise programs followed by aquatic exercise plus functional exercise programs and aerobic exercise had the highest P-scores.
• Global heterogeneity was very high: I2 = 99.54% (95% CI 99.44%, 99.618%), tau2 = 199.50 for pain intensity; I2 = 94.44% (95% CI 88.86%, 97.22%), tau2 = 34.20 for functional disability.
• In the multivariate NMA, heterogeneity was significant (Cochrane Q X2(11) = 438.21, p < 0.001); I2 ranged from 93.36% to 93.52% across pain scales and 93.40% to 93.45% across disability scales.
• Publication bias tests were not significant for pain intensity (Egger t(25) = -0.19, p = 0.854; Thompson-Sharp t(25) = -1.77, p = 0.088) or functional disability (Egger t(12) = 0.88, p = 0.397).
• For functional disability, aerobic training vs. control was significantly different: -16.10 (95% CI -27.74, -4.46) in favor of aerobic training; complementary therapy vs. control: -36.10 (95% CI -48.56, -23.64) in favor of complementary therapy.

Limitations

• Very high heterogeneity (I2 up to 99.54%) due to variability in participant characteristics, intervention types and durations, and outcome measurement tools limits the reliability of precise effect estimates.
• A significant proportion of included RCTs had high risk of bias (6 of 16 RCTs rated high risk by RoB2), and three non-randomized studies carried serious overall risk of bias.
• The male/female ratio was unbalanced across comparisons involving cognitive therapy or health information, meaning sex may act as a confounding variable in those comparisons.
• All outcomes were evaluated short-term only (pre/post-treatment); whether improvements are maintained long-term is unknown.

Why it matters

For patients

People with long-term back pain are most likely to benefit from programs that combine movement, pain education, and strategies for self-management rather than relying on advice alone.

For clinicians

Multimodal programs integrating cognitive-functional therapy with structured exercise should be prioritized over passive or single-modality approaches; ongoing feedback and follow-up appear essential for adherence and outcomes.

For readers

This is the first network meta-analysis directly comparing lifestyle-focused interventions for chronic non-specific low back pain, providing a ranked comparison across 13 intervention types, though high heterogeneity means rankings should be interpreted with caution.