Central and statistical data monitoring in the Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage (CRASH-2) trial

Abstract

Background The purpose of monitoring in clinical trials is to ensure the rights, safety, and well-being of trial patients and the accuracy of the trial data. In the Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage (CRASH-2) trial, which recruited over 20,000 adult trauma patients worldwide, the nature and extent of monitoring was based on a risk assessment undertaken before recruitment started. Purpose We report the methods used for central and statistical monitoring in the CRASH-2 trial and explain how central monitoring was used to target on-site investigations. Methods To ensure that trial participants met the inclusion criteria, we monitored event rates for the primary (death) and secondary outcomes (blood transfusion given). We monitored four quantitative variables (systolic blood pressure (SBP), heart rate (HR), respiratory rate, and capillary refill time) as indicators of the severity of bleeding. We used the coefficient of variation (CV) to identify sites with too much or too little variability. To ensure the accuracy of the data on side effects, we monitored thromboembolic events at each site. Sites with higher or lower than expected event rates were identified for further evaluation. Results A total of 274 sites recruited patients: 145 sites recruited ≥20; patients, and 52 sites recruited ≥100 patients. Sites with low case fatality and low blood transfusion rates were found to be including patients with relatively mild haemorrhage. One site with a high rate of thromboembolic events was found to be using clinical judgement alone. Measurements of SBP and HR varied by about one-fifth of their average value, and capillary refill time measurements varied by around one-third of their average; between-site variation was lowest for blood pressure. Limitations A comparison of mean and median CV indicated that the distributions are slightly skewed to the right. Our simple approach to calculating 95% confidence intervals for the CV may be improved by using a logarithmic transformation of CV for each variable. Conclusions Central and statistical monitoring of data can be used to monitor clinical trials, particularly large, pragmatic, international trials where 100% on-site monitoring is neither necessary nor cost-effective. In the CRASH-2 trial, re-education about trial protocol and the development of guidance helped resolve the issues identified during monitoring.

Truths, lies, and statistics

“…Statistical Monitoring is made possible by the structured nature of clinical data. In regard to multicenter trails often data is collected utilizing similar collection forms, making it easy to run analyses evaluating any abnormalities in all variables or specific variables of one center compared to others (49)(50)(51)(52). Randomized trails may also take advantage of the structured nature of variables for statistical monitoring.…”

Data fraud in clinical trials

“…First, statistical monitoring relies on the highly structured nature of clinical data, since the same protocol is implemented identically in all participating centers, where data are collected using the same case report form, whether on paper or electronically. Abnormal trends and patterns in the data can be detected by comparing the distribution of some or all variables in each center against all other centers [73–76]. Similar comparisons can also be made between other units of analysis, if the structure of the trial warrants it.…”

“…Different implementations of statistical monitoring have been proposed in the literature [73–76,79–80]. The most popular approach is based on ‘key risk indicators,’ which are clinical data variables identified as important, and monitored throughout the trial against pre-specified thresholds [73,79].…”

“…The most popular approach is based on ‘key risk indicators,’ which are clinical data variables identified as important, and monitored throughout the trial against pre-specified thresholds [73,79]. A site that exceeds the threshold for a key risk indicator is flagged for further scrutiny.…”

Statistical monitoring of data quality and consistency in the Stomach Cancer Adjuvant Multi-institutional Trial Group Trial

“…CSM methods have recently received increased attention in the clinical trial literature [8][9][10][11][12] as well as from regulatory agencies [5,6]. The statistical concepts underlying CSM were proposed two decades ago to detect fraud [13][14][15][16], but there has since been a growing recognition that the same concepts can be extended to assess data quality in general [17].…”