Inclusion and exclusion criteria used in non-specific low back pain trials: a review of randomised controlled trials published between 2006 and 2012.

BACKGROUND: Low back pain is a common health complaint resulting in substantial economic burden. Each year, upwards of 20 randomised controlled trials (RCTs) evaluating interventions for non-specific low back pain are published. Use of the term non-specific low back pain has been criticised on the grounds of encouraging heterogeneity and hampering interpretation of findings due to possible heterogeneous causes, challenging meta-analyses. We explored selection criteria used in trials of treatments for nsLBP. METHODS: A systematic review of English-language reports of RCTs in nsLBP population samples, published between 2006 and 2012, identified from MEDLINE, EMBASE, and the Cochrane Library databases, using a mixed-methods approach to analysis. Study inclusion and exclusion criteria were extracted, thematically categorised, and then descriptive statistics were used to summarise the prevalence by emerging category. RESULTS: We included 168 studies. Two inclusion themes (anatomical area, and symptoms and signs) were identified. Anatomical area was most reported as between costal margins and gluteal folds (n = 8, 5%), while low back pain (n = 150, 89%) with or without referred leg pain (n = 27, 16%) was the most reported symptom. Exclusion criteria comprised 21 themes. Previous or scheduled surgery (n = 84, 50%), pregnancy (n = 81, 48%), malignancy (n = 78, 46%), trauma (n = 63, 37%) and psychological conditions (n = 58, 34%) were the most common. Sub-themes of exclusion criteria mostly related to neurological signs and symptoms: nerve root compromise (n = 44, 26%), neurological signs (n = 34, 20%) or disc herniation (n = 30, 18%). Specific conditions that were most often exclusion criteria were spondylolisthesis (n = 35, 21%), spinal stenosis (n = 31, 18%) or osteoporosis (n = 27, 16%). CONCLUSION: RCTs of interventions for non-specific low back pain have incorporated diverse inclusion and exclusion criteria. Guidance on standardisation of inclusion and exclusion criteria for nsLBP trials will increase clinical homogeneity, facilitating greater interpretation of between-trial comparisons and meta-analyses. We propose a template for reporting inclusion and exclusion criteria.

The Power of Low Back Pain Trials: A Systematic Review of Power, Sample Size, and Reporting of Sample Size Calculations Over Time, in Trials Published Between 1980 and 2012.

STUDY DESIGN: A systematic review of nonspecific low back pain trials published between 1980 and 2012. OBJECTIVE: To explore what proportion of trials have been powered to detect different bands of effect size; whether there is evidence that sample size in low back pain trials has been increasing; what proportion of trial reports include a sample size calculation; and whether likelihood of reporting sample size calculations has increased. SUMMARY OF BACKGROUND DATA: Clinical trials should have a sample size sufficient to detect a minimally important difference for a given power and type I error rate. An underpowered trial is one within which probability of type II error is too high. Meta-analyses do not mitigate underpowered trials. METHODS: Reviewers independently abstracted data on sample size at point of analysis, whether a sample size calculation was reported, and year of publication. Descriptive analyses were used to explore ability to detect effect sizes, and regression analyses to explore the relationship between sample size, or reporting sample size calculations, and time. RESULTS: We included 383 trials. One-third were powered to detect a standardized mean difference of less than 0.5, and 5% were powered to detect less than 0.3. The average sample size was 153 people, which increased only slightly (∼4 people/yr) from 1980 to 2000, and declined slightly (∼4.5 people/yr) from 2005 to 2011 (P < 0.00005). Sample size calculations were reported in 41% of trials. The odds of reporting a sample size calculation (compared to not reporting one) increased until 2005 and then declined (Equation is included in full-text article.). CONCLUSION: Sample sizes in back pain trials and the reporting of sample size calculations may need to be increased. It may be justifiable to power a trial to detect only large effects in the case of novel interventions. LEVEL OF EVIDENCE: 3.

A Systematic Review of Outcome Measures Use, Analytical Approaches, Reporting Methods, and Publication Volume by Year in Low Back Pain Trials Published between 1980 and 2012: Respice, adspice, et prospice.

BACKGROUND: Increasing patient-reported outcome measures in the 1980s and 1990s led to the development of recommendations at the turn of the millennium for standardising outcome measures in non-specific low back pain (LBP) trials. Whether these recommendations impacted use is unclear. Previous work has examined citation counts, but actual use and change over time, has not been explored. Since 2011, there has been some consensus on the optimal methods for reporting back pain trial outcomes. We explored reporting practice, outcome measure use, and publications over time. METHODS: We performed a systematic review of LBP trials, searching the European Guidelines for the management of LBP, extending the search to 2012. We abstracted data on publications by year, outcome measure use, analytical approach, and approaches taken to reporting trials outcomes. Data were analysed using descriptive statistics and regression analyses. RESULTS: We included 401 trials. The number of published trials per year has increased by a factor of 4.5 from 5.4 (1980-1999) to 24.4 (2000-2012). The most commonly used outcome measures have been the Visual Analogue Scale for pain intensity, which has slowly increased in use since 1980/81 from 20% to 60% of trials by 2012, and the Roland-Morris Disability Questionnaire, which rose to 55% in 2002/2003, and then fell back to 28% by 2012. Most trialists (85%) report between-group mean differences. Few (8%) report individual improvements, and some (4%) report only within-group analyses. Student's t test, ANOVA, and ANCOVA regression, or mixed models, were the most common approaches to analysis. CONCLUSIONS: Recommendations for standardising outcomes may have had a limited or inconsistent effect on practice. Since the research community is again considering outcome measures and modifying recommendations, groups offering recommendations should be cognisant that better ways of generating trialist buy-in may be required in order for their recommendations to have impact.

The effect of journal impact factor, reporting conflicts, and reporting funding sources, on standardized effect sizes in back pain trials: a systematic review and meta-regression.

BACKGROUND: Low back pain is a common and costly health complaint for which there are several moderately effective treatments. In some fields there is evidence that funder and financial conflicts are associated with trial outcomes. It is not clear whether effect sizes in back pain trials relate to journal impact factor, reporting conflicts of interest, or reporting funding. METHODS: We performed a systematic review of English-language papers reporting randomised controlled trials of treatments for non-specific low back pain, published between 2006-2012. We modelled the relationship using 5-year journal impact factor, and categories of reported of conflicts of interest, and categories of reported funding (reported none and reported some, compared to not reporting these) using meta-regression, adjusting for sample size, and publication year. We also considered whether impact factor could be predicted by the direction of outcome, or trial sample size. RESULTS: We could abstract data to calculate effect size in 99 of 146 trials that met our inclusion criteria. Effect size is not associated with impact factor, reporting of funding source, or reporting of conflicts of interest. However, explicitly reporting 'no trial funding' is strongly associated with larger absolute values of effect size (adjusted ß=1.02 (95 % CI 0.44 to 1.59), P=0.001). Impact factor increases by 0.008 (0.004 to 0.012) per unit increase in trial sample size (P<0.001), but does not differ by reported direction of the LBP trial outcome (P=0.270). CONCLUSIONS: The absence of associations between effect size and impact factor, reporting sources of funding, and conflicts of interest reflects positively on research and publisher conduct in the field. Strong evidence of a large association between absolute magnitude of effect size and explicit reporting of 'no funding' suggests authors of unfunded trials are likely to report larger effect sizes, notwithstanding direction. This could relate in part to quality, resources, and/or how pragmatic a trial is.