The effect of timing of incentive payments on response rates for cohort study telephone interviews in primary care setting with cost-minimization analysis, a randomized controlled trial

Patients are frequently approached for surveys because they are of considerable research interest. Using surveys to collect patient self-reported information is very common in health services research, such as epidemiological studies [1], clinical trials [2] and patient satisfaction studies [3].

Telephone surveys are commonly used for collecting data in longitudinal studies of patient populations. There are significant advantages to using telephone interviews to administer surveys: it reduces the amount of time a patient needs to spend at the doctor’s clinic; it removes the need for subjects or interviewers to travel; and interviews can be conducted outside of business hours when respondents are more available. Overall, it is more convenient for the subject and can potentially help to enhance response rates for longitudinal studies that require patient self-reported information. Telephone interviews are often preferable to postal or online surveys as the interviewer can help to ensure the validity and reliability of responses by reducing missing data and providing item clarification to the respondents [4]. This is particularly important in patient population settings, where subjects are often elderly who may have a visual impairment or in populations with poor literacy. Furthermore, an essential component for most survey studies is an adequate response rate to reduce non-response or self-selection bias [5, 6].

Many studies have found that incentive payments can enhance the response rates for postal and telephone surveys [7, 8]. However, findings about the effect of timing of incentive payments on response rates have been conflicting. A meta-analysis on the effects of incentive payment in cross-sectional postal surveys found that compared to delayed incentive payments, up-front incentive payments can increase the response rates of 11.9 percentage [7]. A Cochrane review which included 24 trials evaluating the timing of incentive payments for postal questionnaires found that the odds of receiving a response increased by more than a half for up-front payments, compared with delayed payments (odd ratio: 1.61; 95 % confidence interval 1.36 to 1.89) [9]. Previous studies have adopted the social exchange theory to explain up-front incentive payments can increase response rates [6, 10, 11]. It was suggested that “the norm of reciprocity” plays a role in compliance to requests [12, 13]. Requesters can trigger the reciprocity by offering an unsolicited favor. Recipients of a gift or incentive are more likely to agree to a subsequent request than someone who did not receive any favors [12]. There are two explanations. First, people return favors because they are concerned what the other people think of them they do not want to be perceived as ‘free riders’ [14]. Second, people feel good about themselves when they return a favor [12]. Up-front incentive payments can thus initiate an exchange that influences people to feel obligated to complete the survey. Accepting an up-front incentive payment without answering the subsequent telephone survey might make people feel guilty, according to the norm of reciprocity in the social exchange theory [6, 10, 11, 15]. On the contrary, delayed incentive payment imposes little social obligation on people. The respondent with a delayed payment option is the one who holds controls over the initiation of an exchange. Before a respondent answers a survey, no obligation exists on either requester or respondent [6, 11].

On the contrary, the Cochrane review found that the timing of incentive payments had no effect for electronic questionnaire response rates (odd ratio: 1.08; 95 % confidence interval 0.77 to 1.50) [9]. Furthermore, another meta-analysis found that there was no difference in response rates between up-front and delayed incentive payments for interviewer-mediated surveys, including face-to-face and telephone interviews [8].

However, there are still some knowledge gaps. First, in previous studies, the mode of questionnaire administration has been mainly by self-administration [9]. The effect of timing of incentive payments on the response rate of interviewer-administered questionnaire, especially for telephone surveys, is still unknown. Second, previous studies have been predominantly conducted in non-patient populations, such as in the general population, students, technicians, corporate settings and medical personnel [7, 9]. Third, most studies to date have been cross-sectional. Little is known about the effect of timing of incentive payments on the response rates to longitudinal telephone surveys in primary care patient settings.

In order to understand the impact of timing of incentive payments on response rates, we conducted supplementary and parallel work on our naturalistic observation study on Chinese primary care patients with lower urinary tract symptoms [16]. The aim and objectives, and the study methodology of the main study have been reported elsewhere [16–19]. One of the advantages to conducting the investigation in a primary care population was that we could approach the notion of general population in primary care while avoiding some of the complexities of having to sample the general population. Second, telephone survey is preferable in studies in primary care where the doctor’s consultation schedules are tight with many patients attending the doctor in a short period of time. Compared with face-to face interviews, telephone surveys can save the patient’s time at the clinic; enhance the number of patients recruited; and is less disruptive on the delivery of medical services.

The present study aimed to examine whether up-front or delayed incentive payment was associated with a higher response rate for telephone interviews in a longitudinal cohort study on primary care patients, and to compare the cost between the two incentive payment methods.