Skip main navigation

Learning From Null

Negative Trials Are Important, Particularly in Low- to Middle-Income Countries
Originally publishedhttps://doi.org/10.1161/CIRCOUTCOMES.119.006280Circulation: Cardiovascular Quality and Outcomes. 2019;12:e006280

See Article by Owolabi et al

The importance of universal trial reporting, regardless of result, is clear. Not only are negative and neutral results essential to understanding the world but they also can inform subsequent trials. To execute a successful and informative randomized controlled trial, innumerable pieces have to fall into place—from design to recruitment to execution to follow-up. Even in situations where resources are optimal, trials commonly fail to inform care because one or more trial elements defy expectations. Executing a trial, then, in a novel environment is an especially daunting prospect—as is the case for the <1% of trials conducted in Africa.1 In novel environments, all of the pieces have to come together, yet prior data, infrastructure, and resource limitations provide limited assurances that they will. In this issue of Circulation: Cardiovascular Quality and Outcomes, Owolabi et al2 present the result of just such an effort in Nigeria, the THRIVES trial (Tailored Hospital-based Risk reduction to Impede Vascular Events after Stroke). Although one of those pieces did not fall into place in THRIVES, this work teaches important lessons about executing trials in novel environments.

THRIVES was a randomized comparison of an implementation strategy to improve blood pressure (BP) control amongst stroke survivors recruited from 4 hospitals in Nigeria. The THRIVES intervention was a lightweight behavioral intervention designed to be scalable and applicable to a large target population. The intervention was delivered to stroke survivors within 1 year of stroke and consisted of previsit reminders, an educational video, a report card assessing risk factor control, and postvisit mobile phone texts emphasizing optimal risk factor control. Among the 400 randomized patients, there was no difference in the primary outcome—change in BP control at 1 year between the intervention (baseline BP 139/83, 12-month systolic BP 137) and control (baseline BP 138/84, 12-month systolic BP 136) groups. A potentially surprising reason for this null finding was the relatively good baseline BP control in both groups. This was unexpected as limited prior data suggested that severe hypertension was more prevalent in these groups. In a post hoc exploratory subgroup analysis, among patients with baseline hypertension (BP >140/90), systolic BPs decreased in both groups, but without a difference between groups, albeit with very modest statistical power to detect differences. The authors reasonably speculate that the failure of a difference to emerge between groups may be a reflection of the higher intensity care received by the control group, including regular appointment reminders and modest financial incentives for study visits.

Every trial has to make difficult value judgments that weigh the potential benefits of collecting additional information against the costs of capturing it. The same logic applies before the trial—is the value of getting high-quality preliminary data worth the time and resources needed to acquire it? Owolabi et al2 judged that it was better to push forward with uncertain assumptions about baseline BP rather than wait to collect the data. It is possible that judgment resulted in a false negative trial. However, in executing the trial, they also collected just the sort of high-quality preliminary data to enable subsequent trials. Assuming that study execution required a moderate number of resources, it is hard to argue with their judgment. Executing trials in trial-naive environments comes with some downside, but if a trial is seen as a pathway to subsequent trials, the magnitude of that downside should not be disqualifying.

Moreover, THRIVES is an interesting example of some of the virtues of conducting trials in low- to middle-income countries. First, and most importantly, the need for studies in these novel environments is great and growing over time. Stroke and cardiovascular diseases are emerging as leading causes of death and disability in low- and middle-income countries, and this trend is projected to continue to rapidly increase.3,4 Second, the costs of trial execution are vastly lower. Given the lower costs of labor and inputs, it is entirely plausible that low costs need not imply low quality. Taken together, these 2 points argue that the value of trials in low- to middle-income countries may be considerably greater than in higher income countries. Finally, the different clinical environments in low- to middle resource countries sometimes open up scientific questions that are inaccessible in wealthier countries. If a new technology has already disseminated in a wealthy country, for example, patient and provider expectations may create barriers to randomizing to control/placebo. Conversely, in a predissemination region, these barriers may be less steep.

Owolabi et al are to be congratulated for executing a high-quality trial—meeting the enrollment target in a short period of time, maintaining excellent outcome follow-up, and, we suspect, doing so with limited resources—in an environment where prior trial infrastructure was limited and preliminary data was lacking. Although it may be disappointing that the trial was negative, if this work is built upon in subsequent trials, the value of this negative study may easily outpace the value of many positive studies in more richly resourced regions.

Footnotes

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

James F. Burke, MD, Department of Neurology, University of Michigan, 2800 Plymouth Rd, Bldg 14, Rm G105, Ann Arbor MI 48109. Email

References

  • 1. luo J, Wu M, Chen W. Geographical distribution and trends of clinical trial recruitment sites in developing and developed countries.J Health Inform Dev Ctries. 2017; 11:1–18.Google Scholar
  • 2. Owolabi MO, Gebregziabher M, Akinyemi RO, Akinyemi JO, Akpa O, Olaniyan O, Salako BL, Arulogun O, Tagge R, Uvere E, Fakunle A, Ovbiagele B. Randomized trial of an intervention to improve blood pressure control in stroke survivors.Circ Cardiovasc Qual Outcomes. 2019; 12:e005904. doi: 10.1161/CIRCOUTCOMES.119.005904LinkGoogle Scholar
  • 3. Gaziano TA, Bitton A, Anand S, Abrahams-Gessel S, Murphy A. Growing epidemic of coronary heart disease in low- and middle-income countries.Curr Probl Cardiol. 2010; 35:72–115. doi: 10.1016/j.cpcardiol.2009.10.002CrossrefMedlineGoogle Scholar
  • 4. Nowbar AN, Gitto M, Howard JP, Francis DP, Al-Lamee R. Mortality from ischemic heart disease.Circ Cardiovasc Qual Outcomes. 2019; 12:e005375. doi: 10.1161/CIRCOUTCOMES.118.005375LinkGoogle Scholar