Single-dose clinical trial methods for evaluating analgesics have been used successfully for over 50 years. The aims of this review were to examine which pain measurement scales have been used in high quality acute pain trials, to investigate other common measurements or characteristics, to confirm that different scales used by standard methods give the same estimate of analgesic effect, and to investigate remedication methodologies and the potential of 'time to remedication' as a standard outcome. Published reports of randomised, double blind, placebo-controlled trials, investigating at least 20 adult patients (10 patients per treatment arm) experiencing moderate or severe pain using at least one standard pain intensity or pain relief scale were sought. Key design features, outputs and outcomes were cataloged for each report. The main outcomes reported were misleading, detailing only the mean values of data with a demonstrably skewed distribution. After 50 years, the reporting of results from acute pain trials warrants a fresh look. Possible improvements include reporting the number of patients with certain levels of pain relief, or the actual number (percentage) of patients with a certain level of pain relief at a certain time, or more useful information on remedication from trials of at least 12 hours duration. Most useful would be all three. Further exploration would only be possible from analysis at the individual patient level.
ContributorsBarden, Jodie and Jayne Edwards, Lorna Mason, Henry McQuay, R. Andrew Moore
Disease Category: Anaesthesia & pain control
Disease Name: Pain (acute)
Age Range: 18 - 100
Sex: Either
Nature of Intervention: Drug
- None
- Systematic review of outcome measures/measurement instruments
- Systematic review of outcomes measured in trials
- Systematic review
Clinical trials were drawn from existing systematic reviews of single-dose oral aspirin, ibuprofen and paracetamol where large numbers of trials were available. Studies had to be randomised and double blind, include a placebo comparison, investigate adult patients with established pain of moderate to severe intensity, use at least one standard pain intensity or pain relief scale, and include more than 10 patients per treatment arm.
Pain measures validated and accepted for the calculation of TOTPAR and SPID were five-point categorical pain relief (PR) scales with comparable wording to ‘none, slight, moderate, good or complete’ (scales using percentages or money as anchors were not permitted); four-point categorical pain intensity (PI) scales with comparable wording to ‘none, mild, moderate, severe’; visual analogue scales (VAS) for pain relief and VAS for pain intensity. Dichotomous information from the top two categories of a global evaluation were accepted provided it was a standard five-point categorical scale (‘poor, fair, good, very good, excellent’) over 4–6 hours and completed by the patient (not the investigator or nurse).
Each report that could possibly be described as a randomised controlled trial was read independently by at least three authors and scored using a commonly used three item, 1–5 score, quality scale. Consensus was then achieved. The maximum score of an included study was 5 and the minimum score was 2. Authors were not blinded because they were already familiar with the literature. The unit of the analyses was published clinical reports. Key design features, outputs and outcomes detailed in the methods of each report were catalogued. Where trials had been included in more than one of the original reviews they were counted only once to avoid duplication.
The derived outcomes extracted were TOTPAR and SPID for both categorical and visual analogue scales, and the number of patients reporting the top two categories (‘very good’ and ‘excellent’) on the global scale. Data from standard measurement scales were dichotomised for all outcomes. The mean TOTPAR, SPID, VAS TOTPAR or VAS SPID values for active and placebo were converted to %max TOTPAR or %max SPID by division into the calculated maximum value (Cooper, 1991). The proportion of patients in each treatment group who achieved at least 50%max TOTPAR or %max SPID was calculated using verified methods. These proportions were then converted into the number of patients achieving at least 50%max TOTPAR by multiplying by the total number of patients in the treatment group. Relative benefit was calculated with 95% confidence intervals using a fixed effects model. The number needed to treat (NNT) with confidence intervals was calculated by the method of Cook and Sackett (1995) from the sum of all events and patients for treatment and placebo.
To compare differences between pain measures, trials reporting the same multiple measures were grouped and the NNT from each compared. For instance, results from those trials reporting both categorical TOTPAR and categorical SPID were pooled separately for each measure and the NNTs for each measure compared. This was repeated for each subsequent paired comparison. Potential differences between remedication methods were explored by sensitivity analysis where trials were grouped by method and the results compared. The statistical significance of the difference between NNT values derived from the different measures at the same doses of analgesic and the different remedication methods was established using the z test. Comparisons were only made when there was information from at least 300 patients to exclude spurious random effects. Calculations were performed using Microsoft Excel 2001 on a Power Macintosh G4.