Promoting the return of lapsed blood donors: A seven-arm randomized controlled trial of the question-behavior effect

Research Group on Behavior and Health, Laval University., Hema-Quebec., Institute of Psychological Sciences, University of Leeds., Hema-Quebec., Department of Psychology, University of Sheffield.

Health Psychology. 2014;33((7):):646-55.

Clinical Commentary

Prof. Barbara Masser - University of Queensland, St Lucia, Australia.

What is known?

The question-behaviour effect (QBE) defined as the increased probability of engaging in a behaviour as a result of simply measuring intentions to engage in that behaviour has been observed for many health behaviours. It has previously been observed in blood donation, but the results as to the efficacy of this as an intervention for blood donation are somewhat mixed. The efficacy of the QBE as a method of re-activating lapsed donors has not been previously explored.

What did this paper set out to examine?

The authors conducted a randomized controlled trial with 7,000 lapsed donors (donors who had not registered to give blood in the last 2 years). The aim was to explore whether the QBE could be enhanced by i) measuring intentions in an interrogative (‘Do I have the intention..’) rather than declarative (‘I intend..’) form; ii) whether assessing additional constructs such as anticipated regret at not donating in the future, moral norm/obligation to donate or positive self-image as a function of donating prior to declarative intention would enhance the QBE in comparison to the intention only conditions and in comparison to an established behavioural change intervention (implementation intentions); iii) whether these effects would vary as a function of donors’ age, gender, first time vs. repeat donor status and whether they returned a completed questionnaire or not.

Participants were randomly assigned to one of the six intervention conditions (declarative or interrogative intention, moral norm, anticipated regret, positive self-image each paired with declarative intention or implementation intentions) or the control (business as usual) condition. The outcome variables were registrations at blood drives during a 6 and 15-month period.

What did they show?

Intention to treat analyses showed that while implementation intentions increased registrations to donate in comparison to the control condition at both 6 and 15 months, the effects of the QBE interventions were less consistent. At 6 months, those allocated to the positive self-image condition registered at a significantly higher rate than donors in the control condition. At 15 months, donors in the anticipated regret, interrogative intention and declarative intention conditions registered at a significantly higher rate than those in the control condition. These effects did not differ by donor gender or age. However, only first-time (and not repeat) donors in the moral norm and positive self -image conditions registered at a significantly higher rate than donors in the control condition at 6 and 15 months. Donors in all the intervention conditions who returned the questionnaires registered at significantly higher rates than those who did not and at a rate significantly greater than those in the control condition at both time points.

Although response rates were uniformly low and thus these are analyses somewhat statistically underpowered, donors in the implementation intention condition who returned their questionnaire registered at a significantly higher rate than responders in all other conditions at 15 months, and at a significantly higher rate than responders in the moral norm, anticipated regret and positive self-image condition at 6 months. At both 6 and 15 months the standard declarative intention questions performed as well as the interrogative intention or enhanced intention questions in increasing registrations. Among those who did not return the questionnaire, there were no differences between conditions in registrations at 6 months. At 15 months, those in the moral norm condition registered at a significantly lower rate than those in the control condition.

What are the implications for practice and for future work?

The practical implications of this study are - somewhat ironically -- that if you want to reactivate lapsed donors then you should distribute an implementation intention intervention. This intervention increased registrations in comparison to the control condition at 6 and 15 months. Although the observed effect size was statistically small if this intervention could be incorporated into practice in a cost neutral way then it would be a valuable tool in reactivating lapsed donors. Acting to increase registrations using QBE type effects depends on whether lapsed donors can be motivated to return the materials, and thus guarantee that their intentions have been measured. If they can then the simple declarative intention items seem to work best although it should be noted that which intervention results in the highest registration rate could change as response rates increase. If measurement cannot be guaranteed then more options open up. If an increase in registrations is needed in the short term (6 months) then positive self-image and intention items should be administered to first-time donors. If a longer-term (15 months) increase is desired then Intention or anticipated regret and intention items results in an increase in registrations to donate in comparison to business as usual practices. Theoretically, in terms of the QBE, the analyses for donors who returned the questionnaires are the most informative. These show that for responding lapsed donors for whom there is evidence that intentions were measured (and who experienced the QBE as theoretically defined) the basic QBE is not enhanced by adding in measures of additional constructs or assessing intentions in alternative forms. Although outperformed by implementation intentions at 15 months, simply measuring declarative intentions significantly increased registrations at 6 and at 15 months in comparison to the business as usual control condition.

This analysis provides many opportunities for future research. In addition to determining how the QBE works and when it will be observed, the current analysis suggests opportunities in exploring ‘mere exposure’ effects as a means to reactivation of lapsed donors. That is, how simply sending a lapsed donor a questionnaire asking them their intentions to donate in the next 6 months results in an increase in registrations to donate not at 6 but at 15 months. For those specifically interested in the QBE as theoretically defined an additional opportunity arises in considering how to improve response rates. Traditionally this is achieved through including low monetary value tokens with questionnaires. However, how this may effect or interact with the QBE in a blood donation context needs to be empirically explored.

Objective: This study tested key variations in the question-behavior effect against a control condition or an implementation intention condition on returning to give blood among lapsed donors (individuals who had not given blood in the past 2 years). Design: At baseline, 7,000 lapsed donors were randomized to 1 of 6 experimental conditions or to a control condition. Participants in the experimental conditions were asked to complete a 6-item postal questionnaire assessing intentions only, interrogative intention, moral norm plus intention, anticipated regret plus intention, positive self-image plus intention, or implementation intentions. Objective measures of behavior were obtained 6 and 15 months later. The frequency of registrations to give blood over the next 6 and 15 months was measured. Results: Intention-to-treat analysis of the frequency of registrations (GENMOD procedure, Poisson distribution) indicated main effects for condition (experimental vs. control) at both 6 months, chi2(1) = 4.64, p < .05, and 15 months, chi2(1) = 5.88, p < .05. Positive self-image and implementation intention interventions outperformed the control condition at 6 months. At 15 months, standard intention, interrogative intention, and regret plus intention conditions showed more frequent registrations compared with control and were just as effective as implementation intention formation. Moderation analysis showed that the moral norm and positive self-image conditions were significant for first-time (1 previous donation) but not repeat (2 or more previous donations) donors. Conclusion: The question-behavior effect can be used to reinvigorate blood donation among lapsed donors, and can be as effective as forming implementation intentions. (PsycINFO Database Record (c) 2014 APA, all rights reserved).
Study details
Language : English
Credits : Bibliographic data from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine