Background and purpose

There is on-going debate about a possible link between manipulation and stroke in patients, and a growing interest in other treatment reactions such as increased pain. Evidence about manipulation is contradictory. There is little published information about outcomes in osteopathy. We aimed to address this gap.

Study purpose and background

There is growing concern that current outcome measures of back pain trials do not comprehensively capture what is important to patients. Some researchers believe we need to incorporate patients in the development of ‘next-generation’ outcomes. As a preliminary step to patient-interviews, we aimed to articulate ‘important change’ from the patients' perspective, as represented in reports of qualitative investigations inadvertently or directly exploring this.

Background and Purpose: How outcomes of clinical trials are reported alters the way treatment effectiveness is perceived. Clinicians interpret outcomes of trials more favourably when results are presented in relative rather than in absolute terms. However, the face validity of different methods is unclear. We aimed to explore which methods clinicians find clearest, most interpretable, and useful.

Methods and Results: We purposively sampled clinicians who see patients with low back pain (LBP) and presented them with summary reports of a hypothetical trial, reporting the results using a variety of different methods. We explored participants’ preferences for these different methods and how they would like to see future trials reported. We interviewed 14 clinicians (GPs, manual therapists, psychologists, a rheumatologist, and surgeons). Participants felt that clinical trial reports were not written with them in mind. They were familiar with mean differences, proportion improved, and number needed to treat (NNT); and unfamiliar with standardised mean difference (SMD), odds ratios and relative risk. They found the proportion improved, relative risk and NNT more intuitively understandable, and were concerned that between-group mean difference, relative risk and odds ratios may mislead. Participants thought each method uniquely contributed to their overall understanding, and that using a variety of methods to report future trials may prevent erroneous portrayal of treatment effect.

Conclusion: Clinicians who see patients with low back pain currently find it difficult to interpret LBP trials. Using a suite of methods to report outcomes may aid clinicians’ interpretation and the transition of research into practice.

Conflict of Interest: None

Sources of Funding: Barts and the London Charity

Background and Purpose: Clinicians have expressed frustration at the difficulty of interpreting low back pain (LBP) trial outcomes. Using a suite of methods to report outcomes may aid interpretation and transition of research into practice. We aimed to facilitate consensus between LBP experts on how future trials are reported.

Methods and Results: We invited SBPR and LBP Forum members, and authors of LBP trials to participate. In the first round, participants were presented with results of a qualitative study on clinicians’ preferences for different reporting methods. They were asked to rate and comment on the appropriateness of including different reporting methods in a standardised set. In the second round, they reviewed other participants’ ratings and comments, re-rated methods, and edited a statement of recommendation for future reporting. In the final round, participants were asked if they approved of a revised statement. Consensus was measured using the RAND/UCLA appropriateness method and ratified in a meeting at LBP Forum X. Sixty-three experts participated in the study. Ninety-eight percent of participants approved a statement recommending that, where possible, results of LBP trials are reported using between-group mean differences (including advice on clinically important difference), proportion of patients improving in each group, NNT to achieve a minimally important change, and the proportion of deteriorating in each group (all with 95% CIs). Also, additional reporting methods were recommended according to needs of particular trials.

Conclusion: A high level of consensus was reached amongst LBP experts on a statement recommending a standardised set of reporting methods.

Conflict of Interest: None

Source of Funding: Barts and the London Charity

Background and Purpose: Using numbers needed to treat (NNT) to report trial outcomes has been publicised as improving the interpretability of trial results. However, authors of back pain trials rarely report continuous outcomes using NNT, and if they do, rarely consider deteriorating patients. We performed a reanalysis of the UK BEAM dataset to explore the effects of reporting NNT on interpretation, and articulate difficulties with this approach.

Method and Results: 1,334 patients consulting for low back pain were randomised to receive either best GP care alone, or best GP care and exercise, manipulation or combined exercise and manipulation. The primary outcome was Roland Morris Disability Questionnaire (RMDQ) score at three and 12 months. Patient-response to a health transition question (TQ) was used as a secondary outcome. Using consensus thresholds for individual change on the RMDQ, and using the TQ, we calculated NNTs for improvement, and for benefit (incorporating one additional improvement gained or an additional deterioration prevented). In contrast to the small-to-moderate mean differences originally reported, NNTs were low, ranging from two to eight. Detailed results have been submitted for publication and will be presented at the meeting. Estimating NNT can be challenging due to difficulties in defining thresholds of individual change on continuous scales. However, using a TQ avoids some of these challenges.

Conclusion: In contrast to the small-to-moderate mean effects originally reported, corresponding numbers needed to treat for the interventions are attractive. Where possible, NNT should be considered for future use alongside more conventional reporting methods.

Background and Purpose: A number of authors have calculated thresholds of individually minimally important individual change (MIC) for the Roland Morris Disability Questionnaire (RMDQ). However, at an individual level, the imprecision of the measurement instrument often exceeds the MIC. In recognition of this, there is some consensus on appropriate individual change thresholds (absolute and baseline-specific) for the RMDQ. These were intended to be re-evaluated as further data become available. With this in mind, we calculated the MIC and minimally detectable change (MDC) of the RMDQ in the UK BEAM population.

Method and Results: 1,334 patients enrolled in the UK BEAM trial completed the RMDQ. We calculated MIC using ROC curves from three and 12 month follow-up data, and using a seven-point health transition question as the external criterion. We performed sub-analyses of MIC for bands of baseline severity and considered adjusting these for regression to the mean (RTM). We calculated MDC based on within-person and residual error variances of stable patients’ repeated measurements. The overall MIC was four points and the MDC 7.6 points. These values fall around the centre of the ranges considered by the consensus study team. Higher MIC values were observed for more disabled patients; this may be partially an artefact of RTM. In our positively skewed population, more disabled patients required more than the consensus recommendation of 30% change from baseline.

Conclusion: This is further evidence that five RMDQ points is an appropriate threshold by which to judge individual change on the RMDQ. Proportional change from baseline may be more population-specific than previously thought.

Randomised controlled trials (RCTs) of non-specific low back pain (LBP) interventions commonly estimate sizes of treatment effect from subjective scales. It is not clear which scales are most commonly used. Moreover, the standardized effect sizes, d, of interventions are typically small (d ≈0.3) and many trials may be underpowered to detect d at this magnitude, regardless of scale used. We identified RCTs reported between 1980 and 2000 from the European guidelines for low back and extended the search to 2006 using MEDLINE, EMBASE, PsycINFO, Cochrane, and Lilacs. We extracted data from identified studies concerning outcome measure use and power to detect d. Following analysis of 222 papers, the most commonly used outcome measure was the visual analogue pain scale (VAS-P), used by 42% of trials, followed by the Roland Morris disability questionnaire (RMDQ), used by 34% of trials. Secondary measure use was diverse; 5% of trials included 10 or more measures and 40% involved at least one objective measure. 57% of the trials were powered to detect d = 0.8, only 37% could detect d = 0.5, only 6% were able to detect d = 0.3 and only one trial could detect Cohen’s definition of a small effect, d = 0.2. In sum, the VAS-P is the most commonly used outcome measure in trials of non-specific LBP and the majority of trials are too small to show effect sizes that might realistically occur.