Dose-response relationship and effect modifier of stabilisation exercises in non-specific low back pain: a project-wide individual patient data re-analysis on 1,483 intervention participants

Finding Predictive Factors of Stabilization Exercise Adherence in Randomized Controlled Trials on Low Back Pain: An Individual Data Reanalysis Using Machine Learning Techniques

OBJECTIVE

To identify predictors of adherence in supervised and self-administered exercise interventions for individuals with low back pain.

DESIGN

Cohort study.

SETTING

Rehabilitation.

PARTICIPANTS

This preplanned reanalysis within the Medicine in Spine Exercise Network included 1511 participants with low back pain (57% female, mean age 40.9 years, SD ±14 years).

INTERVENTIONS

Participants underwent an initial 3-week supervised phase of sensorimotor exercises, followed by a 9-week self-administered phase.

MAIN OUTCOME MEASURES

Biological, psychological, and social factors potentially impacting training adherence were evaluated. During the supervised phase, adherence was tracked through a standardized training log. During the self-administered phase, adherence was monitored via a diary, with adherence calculated as the percentage of scheduled versus completed sessions. Adherence was analyzed both as an absolute percentage and as a dichotomized variable (adherent vs nonadherent, with a 70% adherence cut-off). Predictors for adherence were identified using Gradient Boosting Machines and Random Forests (R package caret). Seventy percent of the observations were used for training, whereas 30% were retained as a hold-out test-set.

RESULTS

The average overall adherence was 64% (±31%), with 81% (±28%) adherence during the supervised phase and 58% (±39%) in the self-administered phase. The root mean square error for the test-set ranged from 36.2 (R2=0.18, self-administered phase) to 19.3 (R2=0.47, supervised phase); prediction accuracy for dichotomized models was between 64% and 83%. Predictors of low to intermediate adherence included poorer baseline postural control, decline in exercise levels, and fluctuations in pain intensity (both increases and decreases).

CONCLUSION

Identified predictors could aid in recognizing individuals at higher risk for nonadherence in low back pain exercise therapy settings.

Machine learning-based identification of determinants for rehabilitation success and future healthcare use prevention in patients with high-grade, chronic, nonspecific low back pain: an individual data 7-year follow-up analysis on 154,167 individuals

ABSTRACT

To individually prescribe rehabilitation contents, it is of importance to know and quantify factors for rehabilitation success and the risk for a future healthcare use. The objective of our multivariable prediction model was to determine factors of rehabilitation success and the risk for a future healthcare use in patients with high-grade, chronic low back pain. We included members of the German pension fund who participated from 2012 to 2019 in multimodal medical rehabilitation with physical and psychological treatment strategies because of low back pain (ICD10:M54.5). Candidate prognostic factors for rehabilitation success and for a future healthcare use were identified using Gradient Boosting Machines and Random Forest algorithms in the R-package caret on a 70% training and a 30% test set. We analysed data from 154,167 patients; 8015 with a second medical rehabilitation measure and 5161 who retired because of low back pain within the study period. The root-mean-square errors ranged between 494 (recurrent rehabilitation) and 523 (retirement) days ( R2 = 0.183-0.229), whereas the prediction accuracy ranged between 81.9% for the prediction of the rehabilitation outcome, and 94.8% for the future healthcare use prediction model. Many modifiable prognostic factors (such as duration of the rehabilitation [inverted u-shaped], type of the rehabilitation, and aftercare measure), nonmodifiable prognostic factors (such as sex and age), and disease-specific factors (such as sick leave days before the rehabilitation [linear positive] together with the pain grades) for rehabilitation success were identified. Inpatient medical rehabilitation programmes (3 weeks) may be more effective in preventing a second rehabilitation measure and/or early retirement because of low back pain compared with outpatient rehabilitation programs. Subsequent implementation of additional exercise programmes, cognitive behavioural aftercare treatment, and following scheduled aftercare are likely to be beneficial.

Personalized assessment and management of non-specific low back pain

BACKGROUND AND OBJECTIVE

Low back pain (LBP), and in particular non-specific low back pain (NSLBP), which accounts for approximately 90% of LBP, is the leading cause of years lived with disability worldwide. In clinical trials, LBP is often poorly categorized into 'specific' versus 'non-specific' and 'acute' versus 'chronic' pain. However, a better understanding of the underlying pain mechanisms might improve study results and reduce the number of NSLBP patients.

DATABASES AND DATA TREATMENT

Narrative review.

RESULTS

NSLBP is a multi-dimensional, biopsychosocial condition that requires all contributing dimensions to be assessed and prioritized. Thereby, the assessment of the contribution of nociceptive, neuropathic and nociplastic pain mechanisms forms the basis for personalized management. In addition, psychosocial (e.g. anxiety, catastrophizing) and contextual factors (e.g. work situation) as well as comorbidities need to be assessed and individually weighted. Personalized treatment of NSLBP further requires individually choosing treatment modalities, for example, exercising, patient education, cognitive-behavioural advice, pharmacotherapy, as well as tailoring treatment within these modalities, for example, the delivery of tailored psychological interventions or exercise programs. As the main pain mechanism and psychosocial factors may vary over time, re-assessment is necessary and treatment success should ideally be assessed quantitatively and qualitatively.

CONCLUSIONS

The identification of the main contributing pain mechanism and the integration of the patients' view on their condition, including beliefs, preferences, concerns and expectations, are key in the personalized clinical management of NSLBP. In research, particular importance should be placed on accurate characterization of patients and on including outcomes relevant to the individual patient.

SIGNIFICANCE STATEMENT

Here, a comprehensive review of the challenges associated with the diagnostic label 'non-specific low back pain' is given. It outlines what is lacking in current treatment guidelines and it is summarized what is currently known with respect to individual phenotyping. It becomes clear that more research on clinically meaningful subgroups is needed to best tailor treatment approaches.

How to tackle non-specific low back pain among adult patients? A systematic review with a meta-analysis to compare four interventions

OBJECTIVE

To tackle non-specific low back pain (NSLBP) among patients and find the most effective solution and to quantitatively synthesize the overall effect of motor control training (MCT) compared with Pilates, McKenzie method, and physical therapy (PT) in pain and physical function.

METHODS

Randomized controlled trials (RCTs) of four types of intervention (MCT, Pilates, McKenzie method, and PT) for LBP were collected by searching PubMed, Web of Science, EBSCOhost (Cochrane Central Register of Controlled Trials), and Scopus databases from the establishment of the database to September 30, 2023. The risk of bias was evaluated for included studies using the Revised Cochrane Risk of Bias tool for randomized trials (RoB 2.0). Taking pain and physical function in the experimental and control groups as outcome indicators, subgroup analysis was performed according to the intervention method to calculate the standardized mean difference (SMD) and 95% confidence interval (CI).

RESULTS

A total of 25 RCTs, including 1253 patients, were included. Meta-analysis showed that MCT effectively relieved pain [SMD = -0.65, 95% CI (- 1.00, - 0.29), p < 0.01] and improved physical function [SMD = -0.76, 95% CI (- 1.22, - 0.31), p < 0.01] comparing with other 3 types of intervention. Subgroup analysis suggested that MCT could alleviate pain [SMD = -0.92, 95% CI (- 1.34, - 0.50), p < 0.01] and improve physical function [SMD = -1.15, 95% CI (- 1.72, - 0.57), p < 0.01] compared with PT, but it had no statistical significance compared with Pilates [pain: SMD = 0.13, 95% CI (- 0.56, 0.83), p = 0.71; physical function: SMD = 0.10, 95% CI (- 0.72, 0.91), p = 0.81] and the McKenzie method [pain: SMD = -0.03, 95% CI (- 0.75, 0.68), p = 0.93; physical function: SMD = -0.03, 95% CI (- 1.00, 0.94), p = 0.95].

CONCLUSIONS

MCT can effectively relieve pain and improve physical function in patients with NSLBP. It is more effective compared with PT for LBP, while no differences were detected between MCT and Pilates, as well as McKenzie method. Therefore, MCT, Pilates, and the McKenzie method should be encouraged as exercise interventions for NSLBP rehabilitation.