Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Schmerber v. California wikipedia , lookup
Hemorheology wikipedia , lookup
Jehovah's Witnesses and blood transfusions wikipedia , lookup
Blood transfusion wikipedia , lookup
Plateletpheresis wikipedia , lookup
Blood donation wikipedia , lookup
Men who have sex with men blood donor controversy wikipedia , lookup
Bone Marrow Transplantation, (1999) 24, 313–319 1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt Factors associated with attrition from a national bone marrow registry GE Switzer1,2, MA Dew2,3,4, AA Stukas2, JM Goycoolea2, J Hegland5 and RG Simmons2,6 Departments of 1Medicine, 2Psychiatry, 3Psychology, 4Epidemiology and 6Sociology, University of Pittsburgh; and 5Director of Scientific and Technical Services, National Marrow Donor Program, Pittsburgh, PA, USA Summary: During its 10-year existence, the National Marrow Donor Program (NMDP) has been extremely successful at recruiting potential bone marrow donors to join the volunteer registry. Due in part to successful recruitment and the longevity of the registry, the focus of the NMDP has now shifted to decreasing potential attrition when volunteers are recontacted for additional testing to determine whether they would be the optimal donor for a specific patient. Our own interest in the bone marrow donation process led us to examine four domains of variables – demographic characteristics, volunteer history, recruitment-related characteristics and donationrelated concerns – that we hypothesized would be associated with increased likelihood of donor attrition at a key donor decision-point (DR-stage blood typing). Questionnaires were mailed to potential donors after they were contacted at the DR-stage, and had made the decision of whether or not to continue with blood typing. Our final sample included 756 volunteers who decided to continue with typing, and 258 individuals who declined further participation in the registry. In the bivariate analyses, factors in three of the four domains (all except demographic characteristics) were found to be substantially correlated with likelihood of attrition. Logistic regression indicated that nine central variables across the three domains produced the majority of increased attrition likelihood. Finally, a dose-response analysis suggested that as the number of attrition-related factors endorsed by an individual increased, his/her likelihood of dropping out of the registry also increased. Implications for future research and interventions to reduce potential donor attrition are discussed. Keywords: bone marrow donors; attrition; retention The National Marrow Donor Program (NMDP) was founded in 1987 to develop a nationwide registry of potential bone marrow donors. Because many patients ill with blood-related disorders cannot find an adequate marrow match among family members, the NMDP established a target goal of registering 100 000 unrelated potential donors who would be available to donate marrow as needed. Now, Correspondence: Dr G Switzer, University of Pittsburgh, Department of Medicine, 3811 O’Hara Street, Pittsburgh, PA 15213, USA Received 6 July 1998; accepted 23 December 1998 10 years later, the NMDP registry has exceed all expectations and is growing at a rate of 20 000–30 000 new volunteers per month, with total registry size of more than 3 million.1 Although scientists and biostatisticians working with the NMDP have recognized that a relatively large number of registered volunteers is necessary to adequately represent the full range of human leukocyte antigen types (HLA: protein markers used for matching two individuals’ bone marrow) found in the US population, the continued rapid growth of the registry has shifted attention from emphasis on donor recruitment to the retention of potential donors who are already on the registry. Donor retention has become increasingly important as the registry matures, especially given the fact that more than a quarter of those registered may, for a variety of reasons (eg declined further testing, medically deferred, unreachable, temporarily unavailable), be unavailable when actually called upon to undergo further blood testing toward donation.1 Although the exact impact of donor attrition is difficult to quantify because patients who search the registry in a preliminary manner and find only a few potential matches may seek other forms of therapy and because there are multiple points at which potential donors may decline further testing, both the NMDP and its funding agencies agree that improving donor retention is crucial. The critical importance of ensuring that potential donors, when contacted, will agree to undergo further blood testing leading to donation led us to examine registrant characteristics that might be related to this decision. Specifically, we were interested in whether those who continued with testing (volunteers) differed from those who declined further testing (nonvolunteers) along four key dimensions: demographics, history as a volunteer in general, registry recruitment issues and donation-related concerns. Although there is a growing body of literature concerning the physical and psychological consequences of unrelated donation,2–6 as well as research examining individuals’ willingness to join the registry and to (hypothetically) agree to donate,7–9 there have been no studies that examine psychosocial issues associated with demonstrated willingness to continue with testing when volunteers are actually recontacted for a specific patient. Most potential donors join the registry by donating a small blood sample that is then partially HLA typed and listed on the computerized registry. A patient searching the registry may identify several potentially matched registrants and request that they undergo further blood typing (DRstage testing). Potential donors are then contacted, informed that they are a potential match, and asked whether they are Unrelated bone marrow volunteers GE Switzer et al 314 still interested in donating. This contact point is important for several reasons, including the fact that it is often the first contact that donors have had with the NMDP since joining the registry (sometimes many years earlier), as well as the fact that most potential donor attrition occurs at this stage. Only the more committed individuals reach the later typing stages. Thus, the goal of current effort was to determine which potential donor characteristics were related to the decision of whether or not to continue when contacted for DR-stage blood typing. In addition to examining potential donor demographic characteristics, we gathered information on three classes of psychosocial variables that previous research and anecdotal evidence suggest may be related to continued registry participation; the evidence for each class of variables is summarized below. Volunteer history In general, we expected that individuals who volunteered in settings other than bone marrow donation would be more likely to continue with blood testing. Studies conducted with blood donors, as well as studies with bone marrow donors indicate that individuals with a history of volunteerism, may develop an internal self-image consistent with helping others, and thus may be more committed, and/or less ambivalent toward volunteering. For example, in a classic set of investigations, Piliavin and colleagues10,11 demonstrated that first-time blood donors who did not have predonation altruistic self-images, often developed an intrinsic commitment to blood donation after subsequent donations. Our own research with bone marrow donors suggests that commitment to blood donation may generalize to other forms of medical volunteerism; we have found blood donors to be less ambivalent about the prospect of their upcoming bone marrow donation than were their non-blood donor counterparts.6 Thus, we anticipated that registry members with any history of volunteering, and more specifically, with a history of blood donation would be less likely to drop out of the registry at DR-stage typing. In addition, we hypothesized that individuals who had joined the bone marrow registry more recently – because their decision to help by donating marrow was more salient – would be less likely to drop out of the registry than those who had joined in the more distant past. Recruitment issues The theme of commitment to the donation process demonstrated by previous volunteerism may also may be central to the context in which individuals are recruited to the registry and, in turn, to later decisions about donating. For example, we expected that individuals who exhibited early ambivalence by delaying the decision to join the registry, or who were discouraged by others from joining, would be more likely to drop out later. Conversely, those who demonstrated some further interest in the process by consulting friends and family or medical professionals, or who were spontaneously encouraged by others to join the registry should be less likely to drop out. We expected that individuals recruited in contexts that appeared to involve social pressure to join the registry – in contrast to contexts that evoked more intrinsic commitment – would be more likely to later drop out than their counterparts. The NMDP recruits prospective donors in a variety of manners, including community, minority, university and corporate drives, drives focused on a specific patient in need of a transplant, and individually, at blood donation centers. In the context of marrow donation, potential donors who joined with other people instead of individually, who joined for a specific ill patient (typically at a community drive in conjunction with friends and neighbors), or who joined because of their membership in a particular ethnic group, might be expected to be less committed to the process than those who initiated registry membership individually. Studies of volunteerism and altruistic behavior in a variety of contexts have demonstrated that external social pressures to help others can undermine intrinsic commitment to help, and cause volunteers to discontinue their participation sooner, or at higher rates than volunteers who did not experience such pressures.11–13 We also expected that, in addition to the conditions under which individuals joined the registry, their stated motives for joining would also be associated with the likelihood they would drop out of the registry later. Previous research has demonstrated that motives are related to continued volunteerism or to feelings about participation in volunteer activities among volunteer ‘buddies’ for AIDS patients,14,15 4-H volunteers,16 social service volunteers,17 elderly volunteers,18 kidney donors4,19 and bone marrow donors.20 Volunteers who are intrinsically motivated to join the marrow registry (eg by their commitment to religious or ethical precepts), should be less likely to drop out as compared to those who are extrinsically motivated (eg by their empathic focus on a specific patient). Donation-related concerns Our final domain of interest included medical, as well as more general, concerns about the donation process. We anticipated that potential donors with greater numbers of concerns would be more likely to drop out of the registry then their less concerned counterparts. Anecdotal accounts from bone marrow donor coordinators suggest that potential donor fears and concerns are often cited as the reason for declining to continue with blood testing. Similar factors are also associated with nondonation among potential blood donors.11,21 Methods Study participants and procedure Forty of the approximately 100 US donor centers were initially selected to participate in the study based on their geographic location, and the total number and ethnic background of volunteers on their registries. Our goal in sel- Unrelated bone marrow volunteers GE Switzer et al ecting centers was to ensure that our sample represented the total registry network on each of these three variables. We began with the 40 donor centers that had previously participated in other research and demonstration projects for the NMDP. Using a stratified random sampling procedure, we added and/or deleted centers until we achieved a total sample of 40 centers that matched the overall registry on the three stratifying variables. Four centers declined to participate when they were initially approached and were replaced with centers with similar characteristics. Two other centers merged several months into the study and, thus, we ended with a final sample of 39 centers. Potential study participants included all registry members at the 39 donor centers who (1) preliminarily matched an ill patient between 1992 and 1996; and (2) either agreed (volunteers) or declined (nonvolunteers) to undergo further blood testing (DR-stage typing) to confirm the match. Other potential donor groups (eg medically deferred, unreachable, temporarily unavailable) – approximately 15–20% of the total number of potential donors requested for DR-stage typing each year – were not included in the sampling frame. Although our sample included only potential donors who specifically declined further testing – because they represent the clearest example of refusal to participate – anecdotal evidence from donor coordinators suggests that (1) potential donors who do not return telephone calls about testing; (2) those who indicate some medical difficulty that would preclude donation; and (3) those who indicate that they are currently unavailable, may also be expressing reluctance to participate and thus be similar in many ways to those who specifically refuse. During the course of the study, the number of potential donors who were contacted to undergo further blood testing and met criteria for our sampling frame averaged approximately 1000 per month. On average, 90% of the total monthly sampling frame were volunteers and 10% were nonvolunteers. To select individuals to receive our questionnaire, we used a random sampling procedure, stratified to ensure adequate representation of potential donors from smaller centers and ethnic minority groups. Nonvolunteers were sampled at a much higher rate than were volunteers to ensure that the data collection period would be similar for both groups. Individuals sampled using this procedure fell into two groups based on whether or not they had given a priori consent (via the informed consent form for joining the registry) to be contacted to participate in other related research projects such as ours. When individuals joined the registry, they could check a box signifying that they would be willing to be contacted about other scientific projects related to bone marrow donation. Only those who had agreed to be contacted became part of our initial sampling frame. Approximately 10% of the sampling frame had not given permission to be contacted for other projects, and were thus excluded from our sampling frame. Questionnaires, consent forms, and a cover letter explaining the study were mailed to donor centers where the sampled registry members (identified to us only by center ID numbers) were registered. Donor coordinators at each center then affixed an address label and mailed the packet to the potential donor. This relatively complex mailing procedure enabled us to adhere to NMDP confiden- tially conventions which suggest that only local donor centers should have access to the names and addresses of potential donors. Subjects who did not respond to the initial mailing within 2 weeks were sent a reminder postcard, and after 2 more weeks, received a second full packet. As we anticipated at the outset of the study, response rates for volunteers (62%) and nonvolunteers (16%) were quite different. Because nonvolunteers had made a decision not to continue participation in the registry, we had expected that it would be difficult to secure their participation in a study related to their past roles as NMDP registry members. Despite our use of the state-of-the-art mailed survey techniques demonstrated to improve response rates (eg repeated follow-up mailings, persuasive cover letter, and monetary and nonmonetary incentives), the response rate for nonvolunteers remained relatively low. We consider the interesting difficulties of securing participation of resistant populations further in the Discussion section. However, we were able to determine that among nonvolunteers, response rates differed only slightly across ethnic groups (ranging from 9% for Hispanics to 17% for Caucasians), length of time on the registry, and by donor center (for those centers that sent at least 10 nonvolunteer – packets, response rates ranged from 10% to 30%, X = 17%, interquartile range = 12%–20%). Measures The central study outcome was whether potential donors decided to continue with testing or to decline further participation in the registry when contacted for DR-stage blood typing. Other measures are described in detail below. Volunteer history: Volunteer history was assessed with a combination of questionnaire items and data extracted from the NMDP database. Current volunteerism in general settings, excluding blood donation, was assessed with a single item asking respondents if they were currently active in any volunteer work (0 = no, 1 = yes). History of blood donation was also assessed with a single item asking whether they had ever been a blood or apheresis donor (0 = no, 1 = yes). The length of time that respondents had been members of the NMDP registry was extracted from the NMDP database and dichotomized (0 = member for less than 4 years, 1 = member for 4 or more years). The midpoint of 4 years was chosen as the cutpoint because at the time the study started, the registry had been in existence for 8 years. Recruitment-related issues: Five items were used to assess respondents’ decisions to initially join the registry. We asked whether they had delayed the decision to join (0 = no, 1 = yes), whether they had consulted family/friends or professionals before deciding to join (0 = no, 1 = yes), and whether they had been spontaneously encouraged (0 = no, 1 = yes) or discouraged (0 = no, 1 = yes) from joining. Three items assessed characteristics of the recruitment setting, including whether other potential registrants were present when respondents joined, whether they joined at a drive focused on a specific patient, and whether their own ethnicity played a role in their decision to join (all coded 0 = no, 1 = yes). 315 Unrelated bone marrow volunteers GE Switzer et al 316 Finally, motives for joining the registry were assessed using an open-ended question asking respondents to describe their reasons for originally volunteering to be bone marrow donors. Responses to this question were assigned to one of five motive categories based on Simmons’ prior work with kidney donors,4,19 and utilized in our own previous studies of bone marrow donors.20 Thus, motives were coded as (1) empathy-related if they focused on respondents’ feelings toward the recipient; (2) idealized helping if they indicated ‘automatic’ responding without deep thought about the potential costs of donation; (3) social or religious obligation if they focused on feelings that donating was a social or religious/moral obligation; (4) exchange-related if they stated or implied an awareness of the relative costs (to themselves) and benefits (to themselves or the marrow recipient) of donating; (5) expected positive feeling if they focused on the good feeling it gave them to join the registry (see Ref. 20 for a complete description and examples of each motive category). Although some respondents indicated multiple motives for joining the registry, assignment to motive categories was based on the first motive listed (0 = motive not endorsed, 1 = motive endorsed). Donation-related concerns: Medical and general concerns about the donation process were assessed by asking respondents to endorse any of a list of possible concerns. Medical concerns included physical pain, damage to the donor’s own health, anesthesia side-effects, and fear of needles (0 = item not endorsed, 1 = item endorsed). General nonmedical concerns included missing time from work, that family would worry, that the patient’s chances of survival were low even with donation, that the donor would have no control over who received his/her marrow, and that marrow donation was against the donor’s religious beliefs (0 = item not endorsed, 1 = item endorsed). Analytic strategy The analysis was conducted in three phases. First, we examined the proportions of volunteers and nonvolunteers who possessed each of the predictor characteristics across four domains: demographic, volunteer history, recruitmentrelated and donation-related. We computed odds-ratios to determine whether having a given characteristic increased the likelihood of registry attrition. Next, we used multivariate logistic regression to examine the unique impact of each factor on registry attrition, controlling for other factors. In order to be appropriately conservative and limit the risk of type I error, attrition-related factors were allowed to enter this regression analysis only if their bivariate relationship with attrition was at least modest in size (effect size, Cohen’s h ⬎ 0.2022), and if the attrition-related factor was sufficiently prevalent to include in multivariate analyses (ie occurred in at least 15% of the sample). Finally, we conducted a dose-response analysis to ascertain whether possessing an increasing total number of attrition-related factors compounded an individual’s likelihood of attrition from the registry. Results Factors associated with donor attrition Percentages of volunteers and nonvolunteers who possessed each of the characteristics within each of the four study domains and odds-ratios indicating which donor characteristics led to increased likelihood of attrition are presented in Table 1. None of the potential donor demographic characteristics were reliably associated with increased likelihood of registry attrition. However, variables from each of the other four domains were associated with higher attrition. In terms of volunteer history, blood donors were less likely to drop out of the registry, and those who had been members of the registry for longer than 4 years were significantly more likely to discontinue participation. Virtually all the recruitment-related issues were associated with attrition. Those who delayed the decision or were discouraged from joining were more likely to drop out of the registry, while those who consulted relatives or professionals were less likely. Individuals who joined with others, joined at a drive for a specific patient, or reported that their ethnic group mattered in their decision of whether to join the registry were all more likely to discontinue registry participation. Those who reported empathy motives were more likely to drop out, while those who reported social/religious, exchangerelated, or expected positive feeling motives were less likely to drop out of the registry. Finally, all medical and general concerns about the donation process – except that one’s family would worry – were associated with increased likelihood of discontinuing registry membership. Fourteen variables that met our a priori criteria of showing at least modestly-sized associations with attrition, and occurring in more than 15% of the total sample were entered into a logistic regression analysis to examine their unique effects on potential donors’ DR-stage status (volunteer/nonvolunteer). Predictors were entered hierarchically in three blocks corresponding to their hypothesized temporal relation to the outcome; volunteer history, recruitment-related issues, and donation-related concerns. Results of the logistic regression are presented in Table 2. Both block I attrition-related factors remained strongly associated with attrition; blood donors were less likely, and those registered longer were more likely, to drop out of the registry. All block II attrition-related factors except joining at a drive for a specific patient and having empathy or social/religious motives for joining predicted increased attrition above and beyond block I predictors. Finally, among block III factors, concerns about damage to the donor’s health and time missed from work for the donation predicted attrition even after block I and II predictors were accounted for. As shown in Table 2, each block of factors produced a significant improvement in model fit over the preceding model (ie the model without the justadded block of variables). The final model that included all variables provided a good overall fit to the data. Finally, we conducted a dose-response analysis to determine if the presence of increasing numbers of attritionrelated factors within a given respondent was associated with enhanced likelihood of discontinuing registry member- Unrelated bone marrow volunteers GE Switzer et al Table 1 317 Factors associated with donor attrition Nonvolunteer (n = 258) Volunteer (n = 756) Demographic characteristics % Women Mean age % Not married % No children % College degree % Not employed % Non-white 59 36 37 36 45 18 18 57 37 36 34 40 14 23 1.09 1.88 (t-value) 1.02 1.05 1.27 1.37 0.73 Volunteer history % Volunteer elsewhere % Blood donor % ⬎4 years on registry 42 58 41 40 77 10 1.09 0.42*** 6.25*** 61 58 23 23 47 46 64 33 30 31 1.85*** 0.78 0.59** 0.68* 1.94*** 58 68 36 42 52 17 1.85*** 1.98*** 2.70*** 34 19 12 4 2 24 20 19 14 5 1.39** 0.97 0.62** 0.31*** 0.36* 68 59 47 35 59 30 36 20 1.48*** 3.28*** 1.58*** 2.22*** 56 41 12 14 2 34 35 7 3 1 2.53*** 1.29 1.84* 4.79*** 6.03*** Donor factor Recruitement-related characteristics Deciding to join % Delayed decision % Consulted friends/family % Consulted professional % Encouraged to join % Discouraged from joining Recruitment setting % Joined with others % Drive for patient % Ethnic group mattered Motives for joining % Empathy % Idealized helping % Social/religious % Exchange % Expect positive feelings Donation-related concerns Donor medical concerns % Pain % Damage own health % Anesthesia % Needles Donor general concerns % Time off work % Family would worry % Patient’s chances low % Who gets marrow % Against religion Odds-ratio *P ⭐ 0.05; **P ⭐ 0.01; ***P ⭐ 0.001. ship. Study participants were grouped into four ordinal categories based on the number of attrition-related factors that each person had (0–1, 2–3, 4–6 and 7–9 factors). Only factors that significantly predicted attrition in the logistic regression analysis were used for this analysis. As illustrated in Figure 1, the logistic model indicates that a strong dose effect exists (B = 1.25; standard error = 0.11; goodness of fit chi-square (3, n = 1014) = 0.65, P = 0.72; improvement over the null hypothesis of no effect: chi-square (2, n = 1014) = 150.78, P ⬍ 0.001). Thus, as the number of factors possessed by a given individual increased, the likelihood that he/she would drop out of the registry also increased dramatically. Discussion Our central aim was to examine potential attrition-related factors for attrition from a national bone marrow donor registry. One of the most striking findings was the modest association of demographic characteristics with attrition, and the relatively greater importance of psychosocial characteristics in determining whether potential donors continued with blood testing. Several of the variables in each of the four psychosocial domains were associated with increased likelihood of attrition in both the univariate and multivariate analyses. Realistic commitment to volunteerism, and to bone marrow donation specifically, emerged as one of the unifying themes of our multivariate analyses. As we had expected, two aspects of volunteer history – whether or not the individual was a blood donor, and the length of time that he/she had been a bone marrow registry member – were highly predictive of donor attrition. Potential donors who responded that their ethnic group was an important factor in the decision to join the registry also dropped out of the registry at higher rates. We had hypothesized that citing one’s ethnic group as a reason for joining the registry might Unrelated bone marrow volunteers GE Switzer et al Table 2 Unique effects of potential attrition-related factors on donor attrition: odds-ratios from logistic regression analysis Odds-ratio Block I predictorsa Blood donor ⬎4 years on registry 0.42*** 5.56*** Chi-square improvement over null (df = 1) 95.78*** Block II predictors Delayed decision Consulted professional Discouraged from joining Joined with others Drive for patient Ethnic group mattered Empathy motives Social/religious motives 1.59* 0.52** 2.27*** 1.56* 1.07 2.08*** 1.45 0.78 Chi-square improvement (df = 6) 55.37*** Block III predictors Damage own health Anesthesia Needles Time off work 2.70*** 1.02 1.33 2.56*** Chi-square improvement (df = 3) 58.99*** Chi-square for full model (df = 13) 844.52 a Variables included in the logistic regresssion had effect sizes (h) greater than 0.2, and occurred in more than 15% of the sample. 70 Percent with given number of factors 318 60 50 40 30 20 10 0 0–1 2–3 4–6 7–9 Total number of factors Figure 1 Likelihood of potential donor attrition by number of factors endorsed. indicate extrinsic social pressure, or even a lack of understanding of the nature of joining a registry of this sort, rather than internal commitment to donation. In addition, variables that might be related to ambivalence about donating – delaying the decision to join the registry and having been discouraged from joining – were associated with greater attrition likelihood. Finally, those who demonstrated commitment to the process by taking the additional step of consulting a health professional were less likely to drop out of the registry. Each of these psychosocial variables is an important indicator of some aspect of commitment to bone marrow donation: past behavior, reasons for joining the registry, others’ reactions, and current attitudes. In addition to variables related to commitment to donation, two ‘donor concerns’ variables were also associa- ted with attrition in the multivariate analysis. Individuals who were concerned that the process might damage their own health, and those who were concerned about the time that they would miss from work were more than twice as likely to drop out of the registry than individuals who did not express these concerns. Somewhat surprisingly, concern about undergoing anesthesia – which is probably the most risk-laden part of the donation process – did not produce elevated attrition likelihood in the multivariate analysis. These findings have several implications for the recruitment and retention of potential bone marrow donors, as well as other types of medical volunteers. First, the lack of significant demographic effects, and the substantial number of psychosocial/contextual effects suggests that there is ample room for interventions targeting donor retention. Given our findings, it seems important to design recruitment settings that (1) reduce ambivalence about joining, perhaps by providing additional informational material or opportunities for discussion; (2) shield potential donors from social pressure to join; and (3) foster intrinsic commitment to donating. Additionally, both at recruitment and when potential donors are contacted for DR-typing, special effort should be made to allay medical and work-related concerns. As was noted in the Methods section, the relatively low response rate among those who dropped out of the registry is one of the limitations of this study. We had anticipated, and did find, that individuals who declined to continue as registry members were less likely to complete a questionnaire about their experiences. Our ability to secure participation from nonvolunteers was further constrained by the fact that registry members – and nonvolunteers – were guaranteed anonymity to everyone except their local donor center. Thus, our questionnaire packets were mailed to nonvolunteers from their local center, perhaps causing misunderstandings about the sponsors and purpose of the questionnaire. Furthermore, we were unable to directly contact potential study participants by telephone either before or after mailing the packet. Prenotification and telephone follow-up are proven techniques for improving response rates.23–25 Once we had confirmed that we were indeed obtaining low response rates among the nonvolunteers, we implemented several procedures, including revising the cover letter to be more persuasive, and offering a variety of incentives, to encourage registry members to complete the questionnaires. These procedures elevated response rates by approximately 8%. Despite this improvement in return, we suspect that nonvolunteers who returned their questionnaires may be somewhat different than those who did not. The direction of such differences is difficult to specify, although it seems unlikely that nonvolunteer nonrespondents would have had higher levels of commitment, for example, than nonvolunteer respondents. Hence, it seems unlikely that our analyses have overestimated the impact of this category of psychosocial variables on the outcome. Nevertheless, because we cannot determine exactly how respondents might differ from nonrespondents, it will be important to confirm our findings in other samples of registry nonvolunteers. In conclusion, this study is the first to examine the relationship of potential marrow donor characteristics to the Unrelated bone marrow volunteers GE Switzer et al decision of whether or not to continue with the process leading toward possible donation. The findings indicate that key psychosocial variables are linked to increased likelihood of attrition from the registry and suggest that intervention strategies targeting these variables both at recruitment and at later contact points might be effective in enhancing donor retention. For example, interventions that encouraged newly recruited donors to participate in other types of medical donation (eg blood donation) might in turn foster enhanced commitment to the registry. More thorough education of new recruits, more selective recruiting procedures designed to provide reluctant volunteers a opportunity to ‘opt-out’ of joining the registry, and asking the potential donor a few questions about how his friends/family feel about donation may all be cost-effective ways of creating a highly committed registry of volunteers. Although we cannot provide specific cost-benefit figures for implementing such interventions rather than simply recruiting additional volunteers, it is clear that the monetary cost of recruiting a volunteer, drawing blood, sending the blood for HLA typing, and entering the typing on the NMDP registry will exceed the small extra costs involved in most of these interventions. In addition to the monetary savings that might be realized by enhancing retention, potential psychological costs to the patient for whom registered and matched volunteers may not be available will be reduced. Finally, potential psychological costs such as guilt, regret, or lowered self-esteem to potential donors who decline further testing may also be attenuated. 7 8 9 10 11 12 13 14 15 16 17 Acknowledgements This work was supported by grant HL48883 from the National Heart, Lung, and Blood Institute, Bethesda, MD, USA. 18 19 References 20 1 National Marrow Donor Program. CPI Statistics 1998. 2 Andrykowski MA. Psychiatric and psychosocial aspects of bone marrow transplantation. Psychosomatics 1994; 35: 13–24. 3 Butterworth VA, Simmons RG et al. Psychosocial effects of unrelated bone marrow donation: experiences of the national marrow donor program. Blood 1993; 81: 1947–1959. 4 Simmons RG, Marine S, Simmons RL. Gift of Life: The Effect of Organ Transplantation on Individual, Family, and Societal Dynamics. Transaction Books: New York, 1977. 5 Stroncek D, Strand R, Scott E et al. Attitudes and physical condition of unrelated bone marrow donors immediately after donation. Transfusion 1989; 29: 317–322. 6 Switzer GE, Simmons RG, Dew MA. Helping unrelated stran- 21 22 23 24 25 gers: physical and psychological reactions to the bone marrow donation process among anonymous donors. J Appl Soc Psychol 1996; 26: 469–490. Beatty PG, Atcher C, Hess E et al. Recruiting blood donors into a local bone marrow donor registry. Transfusion 1989; 29: 778–782. Briggs NC, Piliavin JA, Becker GA. On willingness to be a bone marrow donor. Transfusion 1986; 26: 324–330. Sarason IG, Sarason BR, Slichter SJ et al. Increasing participation of blood donors in a bone marrow registry. Health Psychol 1993; 12: 272–276. Callero PL, Piliavin JA. Developing a commitment to blood donation: the impact of one’s first experience. J Appl Soc Psychol 1983; 313: 1–16. Piliavin JA, Callero PL. Giving Blood: The Development of an Altruistic Identity. Johns Hopkins University Press: Baltimore, MD, 1991. Batson CD, Coke JS, Jasnoski ML, Hanson M. Buying kindness: effect of an extrinsic incentive for helping on perceived altruism. Pers Soc Psychol Bull 1978; 4: 86–91. Kunda Z, Schwartz S. Undermining intrinsic moral motivation: external reward and self-presentation. J Pers Soc Psychol 1983; 45: 763–771. Clary GE, Snyder M. A functional analysis of altruism and prosocial behavior: the case of volunteerism. Rev Pers Soc Psychol 1991; 12: 119–148. Snyder M, Omoto AM. Volunteerism and society’s response to the HIV epidemic. Curr Direct Psychol Sci 1992; 1: 113–116. Rohs FR. Social background, personality, and attitudinal factors influencing decision to volunteer and level of involvement among adult 4-H leaders. J Voluntary Act Res 1986; 15: 87–99. Rubin A, Thorelli IM. Egoistic motives and longevity of participation by social service volunteers. J Appl Behav Sci 1984; 20: 223–235. Okun MA, Eisenberg N. Motives and intent to continue organizational volunteering among residents of a retirement community area. J Community Psychol 1992; 20: 183–187. Simmons RG, Klein SD, Simmons RL. Gift of Life: The Social and Psychological Impact of Organ Transplantation. Wiley-Interscience: New York, 1977. Switzer GE, Dew MA, Butterworth VA et al. Understanding donors’ motivations: a study of unrelated bone marrow donors. Soc Sci Med 1997; 45: 137–147. Oswalt RM. A review of blood donor motivation and recruitment. Transfusion 1977; 17: 123–135. Cohen J. Statistical Power Analysis for the Behavioral Sciences (2nd edn). Lawrence Erlbaum: Hillsdale, NJ, 1988. Dillman DA. Mail and Telephone Surveys: The Total Design Method. John Wiley: New York, 1978. Fox RJ, Crask MR, Kim J. Mail survey response rate: a metaanalysis of selected techniques for inducing response. Public Opin Q 1988; 52: 467–491. Yammarino FJ, Skinner SJ, Childers TL. Understanding mail survey response behavior: a meta-analysis. Public Opin Q 1991; 55: 613–639. 319