To determine from recently published trials the efficacy of exercises in patients with acute, subacute or chronic back pain.

MEDLINE was searched from 1991 to the first quarter of 1995 using the following keywords: 'backache', 'musculoskeletal diseases', 'joint diseases', 'spinal diseases', 'physical therapy', 'evaluation study', 'outcome and process assessment'.

Study designs of evaluations included in the review

Randomised controlled trials (RCTs) of exercise therapy in patients with back pain were included. Excluded were abstracts, unpublished studies, studies in which exercise therapy was given in groups, one study in which not all patients had back pain before randomisation, and one study which did not study the efficacy of exercise therapy.

Specific interventions included in the review

The interventions studied were of the following types: flexion, extension, McKenzie therapy, extension exercises plus mobilisation, back school, workplace visit, exercise programme, waiting list and placebo. Intensive back exercising and normal exercising were studied.

Participants included in the review

The participants were patients with acute back pain defined as of less than 6 weeks' duration, subacute pain defined as being of 6 weeks' to 3 months' duration and chronic pain defined as being of more than 3 months' duration. Patients with specific causes of back pain (radicular syndrome, surgery for lumbar disc protrusion) were excluded.

Outcomes assessed in the review

The outcomes assessed included pain, disability, sick leave, recurrences, back mobility, fitness, strength and overall treatment effect.

How were decisions on the relevance of primary studies made?

The author does not state how the papers were selected for the review, or how many of the reviewers performed the selection.

The validity of the primary studies was assessed using the following criteria: homogeneity of study population, comparability of relevant baseline characteristics, adequacy of randomisation, drop-outs description, number of drop-outs, group size, interventions described, pragmatic study, cointerventions avoided, placebo-controlled, blinding of patients, relevant outcome measure, blinded outcome assessment, adequate follow-up, intention to treat analysis, and frequencies of most important outcomes presented for each treatment group. All the included studies were given a method score, out of a maximum of 100 points, based on the validity criteria. Scoring was performed by the author with doubtful studies being rescored by another general practitioner and a final decision being reached after discussion.

The author does not state how the data were extracted for the review, or how many of the reviewers performed the data extraction.

How were the studies combined?

The studies were combined in a narrative review.

How were differences between studies investigated?

The studies were grouped according to scores on validity, with results reported separately for high and low scoring studies.

Eleven RCTs (N=1,327) were used to assess the efficacy of exercise in back pain: 4 (N=722) studied acute back pain, 1 (N=98) studied subacute back pain, and 6 (N=507) studied chronic back pain.

Twenty-five per cent of the studies had a method score of less than 40 points, whilst 50% of the studies scored 50 points or more. The following shortcoming were present in the primary studies: insufficient description of drop-outs, number of drop-outs, small group sizes, cointerventions, no placebo control group and no intention to treat analysis.

Acute back pain: 2 trials with high method scores comparing exercise therapy with no therapy reported negative results of both flexion and extension exercises.

Subacute back pain: the only trial in this category reported positive results regarding return to work but had a method score of 59.

Chronic back pain: only 1 study comparing exercise with placebo or waiting list control had an appropriate follow-up period of 12 months and a high method score. This showed a positive result for exercise. Two trials comparing intensive back exercises with normal exercises, and 1 trial comparing a fitness programme with normal exercising, reported better results with intensive exercising after 3 to 6 months, but no benefit from intensive exercising after 12 months. The 2 trials comparing extension with flexion had low methods scores and short follow-up periods.

In patients with acute back pain exercise therapy is ineffective. The graded activity programmes with exercises in patients with subacute back pain and intensive extension exercises or fitness exercises in patients with chronic back pain deserve attention. More research on McKenzie therapy, on exercises with a graded activity programme and on different types of exercising in patients with chronic back pain is necessary.

The author diligently scored all included studies on methodological criteria and the review includes tabulation of the scores obtained by the individual studies. The narrative review includes an attempt to explain the conflicting results of the primary studies by considering the method scores obtained. By limiting the literature search to one database some relevant studies may have been omitted. The number of good quality studies considered in this review is small and the addition of further good quality studies may have added to the evidence offered. Though inclusion and exclusion criteria are stated no details are given of the methods used to select studies for inclusion, or of the methods used to extract data. The quality assessment was not independently checked for all studies but only "in cases of doubt" and this term was not defined. Some of the included studies have a very short follow-up period (5 days, 2 weeks, 4 weeks) and it may have been more appropriate to consider including only studies with an adequate follow-up period rather than scoring adequacy of follow-up as one of the validity criteria. Fuller details of the primary studies would have been welcome, including definitions used for back pain, patient characteristics, duration of treatment, experience of therapist, description of interventions, drop-outs and ascertainment of outcome measures used. By contacting the original authors sufficient information may have been obtained to allow an intention-to-treat analysis on raw data. This may have produced different results. Although some attempt was made to explain the heterogeneity among studies, more investigation of this heterogeneity could have been performed by considering specific aspects of validity criteria.

The quality of evidence offered is, as the author states, limited by the small number of studies found. Given the methodological flaws mentioned above, and the lack of relevant detail of the primary studies, it is not possible to either support or oppose the use of exercise in patients with back pain.

Good-quality controlled trials of defined exercise therapy on specific patient groups assessed by defined and relevant outcome measures are required to determine the most cost-effective management of patients with back pain.

The Group Health Foundation; The Prudential Centre for Health Care Research; Wyeth-Ayerst Laboratories.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.