Fewer people than before take part in health surveys. Many researchers seem to experience that potential participants must be offered more than an appeal to solidarity and their sense of responsibility.
Voluntary communal work is a well-known phenomenon: a joint effort in which the individual contributes to the common good. As researchers we have often experienced that volunteers state that the reason for participating is that their contribution may benefit others. However, participation in epidemiological studies has been decreasing over the last 30 – 40 years (1). If this is due to a decline in community spirit in the population, measures that increase participants’ external motivation – for example, remuneration in the form of gifts or money – should boost overall participation.
In this issue of the Journal of the Norwegian Medical Association, Gjøstein and colleagues publish the results of a study of incentives in a medical questionnaire survey of risk factors for colorectal cancer (2). They recruited participants from two groups: one group including people who had failed to reply to an invitation to undergo colonoscopy, and a group that was drawn from the population at random. The groups were then further divided into the following three sub-groups: lottery scratch cards were enclosed with the invitation (unconditional incentive), a prize draw for an iPad if they answered the questionnaire (conditional incentive), or no incentives at all. Among the 6795 participants, the response rate was lower for those who had not answered previously as had been anticipated. However, there were no differences between the different incentive groups.
Does this mean that we should stop rewarding participants in studies? Not necessarily. The researchers chose a Flax lottery ticket as an unconditional incentive. The results of previous studies, also in Norway, have revealed that gifts of money are probably a better choice (3, 4). This was also the conclusion of the authors of a new Canadian study, in which enclosing $2 with the invitation gave an odds ratio of participation of 1.4 (95 % CI 1.2 – 1.7), while a lottery scratch card did not influence the response rate (5). This is not altogether surprising. For most people a lottery scratch card will end in disappointment – while the few who win may become so excited that they completely forget about the questionnaire.
The prospect of benefitting one’s own health can also be an incentive for participating in health surveys. Some studies offer clinical examinations that participants would not otherwise receive. Therefore, it is surprising that almost half of those who had not responded earlier to an offer of colonoscopy screening nevertheless answered the relevant questionnaire survey. In their case, expectations about the time element and discomfort associated with the examination may have carried more weight than any possible medical benefit.
Our experience indicates that the problem of recruiting to health surveys is not that people actively decline but that many do not answer at all. Consequently, it may be wise to use resources to ensure that potential participants have actually received the request and have made up their minds about it. Gjøstein and colleagues called everyone they had not heard from after two weeks, and offered them the option of responding by telephone. This increased participation by 15 percentage points. It is an interesting additional finding. However, most researchers already make use of some kind of reminder – usually a new request in writing. Telephoning people can be regarded as inappropriate pressure, it is probably more expensive than a letter and it is uncertain whether a telephone interview produces the same responses as a questionnaire that is returned in the post. A Danish survey showed that people who took part in telephone interviews used extreme response categories more often, and alternatives such as «Don’t know» less often (6). It would be of great interest if Gjøstein and colleagues further developed their findings and compared the effect and cost of postal reminders in comparison with telephone requests.
The absence of effects for incentives, and possibly an improved response rate in the case of personal requests may indicate that the problem is not a decline in community spirit. There are many other good reasons for the fall in overall participation – and opportunities to remedy this (1): The population are swamped by market surveys – online, by post and telephone. The volume of research projects has also increased while confidence in the research results has probably declined. At times, questionnaire surveys can be so extensive and detailed that this factor alone can be a deterrent. If the study is designed in such a way that it implies repeated requests, it should come as no surprise if the drop-out rate gradually increases as time passes (7).
Perhaps a communal effort among researchers is what we need. Together we can augment information about the importance of recruiting as many participants as possible if we are to acquire robust indicators on the health of the population and find the causes of common illnesses. Together we can also channel our efforts to promote fewer, better and more targeted health surveys through increased cooperation and improved utilisation of data.