Generally, the unit of randomization in a randomized clinical trial (RCT) is the individual participant (the patient). Occasionally, the unit of randomization can be a cluster, such as hospitals, schools, villages, etc. When randomizing large groups, the concept of cluster randomized trials (CRTs) has become increasingly common.

Patient clusters are allocated to the intervention arms of a trial using a randomly generated allocation sequence, as opposed to allocating patients individually. As CRTs involve interventions at a cluster level, informed consent from participants may not always be possible to obtain. Although CRTs may seem simple to manage, informed consent poses a serious challenge for CRTs.

As of November 2014, The George Institute for Global Health have designed a study to provide evidence on the optimum head position for people who have suffered stroke. The trial is called HeadPost.1 The results will determine whether lying flat with a head position of zero degrees is more effective than sitting up with a head position greater than or equal to a 30-degree angle (see image).

The study has been designed as a CRT and clustered by sites. This study is the largest nursing care trial conducted to date, involving over 19,000 patients in 140 hospitals in the U.K., South America, Australia, China, Taiwan and Mongolia.2 As of June, the study has over 17,000 recruited patients and 116 activated sites, of which 113 are currently recruiting. As the study consists of such a large population, certain challenges have occurred throughout the course of the trial.

One challenge that presented itself was obtaining informed consent. Informed consent can, at times, be mistakenly viewed as obtaining a participant’s signature on a consent form. Although this is part of the process, informed consent involves a number of components such as providing adequate information to the participant on the trial; facilitating the participant’s understanding of the information; providing an appropriate amount of time to ask questions and discuss the matter with close relatives; providing information as the study progresses; and obtaining participants’ voluntary agreement to participate.

With the study intervention delivered in large patient pools, participants within the cluster may find it difficult to avoid the intervention. For example, if the intervention was delivering health advice over the radio, participants would be unaware the broadcast was an intervention. As such, their ability to provide consent would be diminished. Obtaining consent for large clusters is challenging. As such, innovative solutions are needed when consent cannot be obtained.

One recommendation that was implemented within the HeadPost study was to appoint a. “gatekeeper” or “cluster representation mechanism.” This representative was a person or body in charge with making the final decision on behalf of the entire cluster. The informed consent was signed off by the site’s head of department and approved by the respective country’s ethics committee. Ethics committees were agreeable to the intervention, as either requiring a patient to lie flat or to sit up are both considered “usual care.” This helped reduce delays in the intervention, reduced pressure on site staff, reduced selection bias (to a degree), and allowed for further enrolment into the study.

Australia has also introduced the “opt-out” consent method. This means all patients (and their details) are given the option to opt-out of the study (or remove some details). The patient will then remain a “partial” participant at any time until the study has been finalized.4 This methodology has shown to reduce selection bias in registries and low-risk studies, with an opt-out rate typically between 2-3%.

CRTs are widely perceived as the recommended approach for evaluating large interventions. These interventions may include community-based health promotion initiatives or education interventions that target health professionals. Although the traditional approach of obtaining informed consent is used throughout clinical trials, this method is impractical for CRTs. Consent must be acquired and applied to the entire cluster, not just individual patients. Furthermore, the consent process must be obtained as early as possible from a representative or “gatekeeper,” who can make the final decision for the cluster. With the “opt-out” method now applicable in Australia, subjects are able to opt-out at any time throughout the study. As a result, ethics committee requirements were fulfilled and enabled the global study be conducted. With a growing demand to evaluate the delivery of health services, public education and policy on social care, the need for CRTs is likely to increase