Previous studies have suggested that the total number of first-time donors is decreasing¹ and that only 8% of those new donors continue donating.² Other reports suggest that younger and minority donors were less likely to donate again within 12 months,³ shorter intervals between a first and second donation increased the likelihood of subsequent donations,⁴ and first-year donation frequency is a strong predictor of long-term donor return.² Variables such as age, ethnicity, blood type, and education have also been shown to influence donor return behavior.⁴

The Retrovirus Epidemiology Donor Study-II (REDS-II) Donor Iron Status Evaluation (RISE) study was a prospective cohort study of hemoglobin (Hb) and iron metabolism in first-time and frequent blood donors. By the midpoint of the 15-month study it became clear that RISE subjects in both cohorts were presenting to donate blood more frequently than predicted. This analysis sought to determine if the reason(s) for higher-than-expected donor return rates involved the donor’s participation in a nonremunerated study.

Shared data on all blood donations were compiled into a centralized research database. REDS-II centers provided monthly donation and deferral records for visits made from January 1, 2006, through December 31, 2009. The database consists of routinely collected information that includes donation procedure, donation type, previous donation date, and other data. Medical history information and donor demographic data, including sex, race or ethnicity, country of birth, and date of birth were also collected from blood center records.

One of the major projects of the REDS-II program was a 24-month longitudinal study of iron status in two cohorts of blood donors. The RISE study enrolled a total of 2425 study subjects with the objective of evaluating the effects of blood donation intensity on iron and Hb status and assessing how these are modified as a function of baseline iron or Hb measures and demographic, reproductive, and behavioral factors. For this analysis, we compared the frequency of donations and other donation-related variables between RISE participants (enrolled subjects) and blood donors from the REDS-II donation database (nonenrolled subjects).

An approximately equal number of donors from each blood center was enrolled in the RISE study. Within the REDS-II donation database, however, centers C and F are the largest of the six blood centers and contribute over 50% of the donation records each year. Among the nonenrolled subjects, the proportion of donors from each blood center matched expected distribution within the larger REDS-II database (Table 1). Most RISE subjects were recruited at fixed locations (80%) while 67% of the index donations made by the nonenrolled donors were given at mobile drives. Roughly 12% of both enrolled and nonenrolled donors were giving blood for the first time at their initial visit. Among the remaining RISE-enrolled donors, 24% of them were reactivated donors and 63% were frequent repeat donors. The proportion of reactivated and frequent repeat within the nonenrolled donor cohort was approximately the same (33 and 52%, respectively).

Donors enrolled in RISE returned to donate significantly more often than nonenrolled donors (Table 2). This pattern was consistent for all six REDS-II blood centers (data not shown). Results from an unadjusted Poisson regression indicate that RISE-enrolled first-time donors gave 2.09 times more than nonenrolled first-time donors within the 15 months after the initial visit. The rate of donation was similar among reactivated donors (rate ratio [RR] = 2.13), while frequent repeat donors enrolled in RISE gave 1.19 times more often than their nonenrolled counterparts. After adjusting for age, race or ethnicity, sex, and other factors, the Poisson donation rates decreased but remained significantly higher in the enrolled donors, ranging from 1.12 to 1.91 times more frequently than nonenrolled donors (Table 2).

Frequent donors had a higher mean number of donations after the index visit (3.40 for enrolled donors and 2.85 for nonenrolled donors) than other donor groups. Among frequent repeat RISE-enrolled donors, a comparison of donation frequency before and after the initial visit showed significantly more donations were given in the 15 months before enrollment (5563 total; mean, 3.62; including 487 [8.8%] double-RBC donations) than were given in the 15 months after enrollment (5219 total; mean, 3.40; including 595 [11.4%] double-RBC donations). Donation rates for frequent nonenrolled donors also declined from 362,960 donations (including 27,138 [7.5%] double-RBC donations) in the 15 months before the index visit to 299,866 donations (including 25,430 [8.5%] double-RBC donations) in the subsequent 15 months.

As the study progressed from initial enrollment of subjects through the final stages, we noticed that enrolled donors seemed to be returning to donate at a rate much higher than expected and higher than that typically seen in blood donors. This informal observation held up in our detailed analysis of the donation database. In each of the donor categories (repeat, reactivated, and first time), the rates of donation for enrolled donors were significantly higher than for those not enrolled in the study, even after controlling for center, sex, race or ethnicity, age, education, country of birth, and donation location. Although study participation may have been causally related to increased donation rates, it is also possible that selection bias played a role in that the most highly motivated donors (in terms of donation frequency) were also the most likely to enroll in our study.

A preindex versus postindex comparison of the frequent donors revealed that both the enrolled and the non-enrolled groups had fewer total donation visits after the index visit than they did before index. Blood centers in the United States generally, and each of the six REDS-II blood centers specifically have been moving toward the conversion of whole blood donors into double-RBC donors to increase component collection overall and improve efficiency of operations. Part, but not all, of the observed decline in number of donations from the preenrollment to the study analysis period was explained by a trend toward more double-RBC donations and fewer whole blood donations in both groups. Regression toward the mean could also explain these data, namely, a group of donors selected because of high donation frequency over one time period tends to have lower (closer to the mean) donation rates over a subsequent time period. It is also possible that motivational factors related to participation in the study, as well as more frequent contact with enrolled donors played a role in lessening the rate of decline in donation visits compared to the nonenrolled subjects.

We also compared donation frequencies across blood centers to understand if any of the observed differences between subject groups could be explained by variation between centers. The amount of contact and methods for contacting enrolled donors did vary quite a bit between the six participating blood centers (see Materials and Methods). In fact our multivariable model found blood center to be a significant predictor of return rate. A cursory examination of the intensity of subject contact and donor return rate did not yield evidence that additional contact attempts resulted in higher donation frequency among enrolled subjects. We did not perform a quantitative analysis of the impact of subject contact methods, but perhaps this could be addressed in future research.

Studies performed by Nguyen and coworkers,⁵ Steele and coworkers,⁶ and Glynn and coworkers⁷ found that the primary factors motivating donors were altruism, medical testing, and the belief that donating blood would bring about a health benefit. It is possible that since our study is about the effects of donating on iron levels, donors were more motivated to return so that their iron levels could be tested. Furthermore, donors motivated by altruism may have been even more inspired to participate in the RISE study since in addition to helping patients who needed blood, they would also be contributing to the advancement of knowledge and safety related to the effect of frequent donation.

Schreiber and colleagues⁸ and Schlumpf and colleagues³ both found that convenience was an important motivational factor for first-time and repeat donors. In the study by Schlumpf and colleagues frequent repeat donors reported that fixed sites were more convenient due to the availability of hours for appointments, but first-time donors reported that mobile draws located close to work-places were more convenient. Although our multivariable analysis showed higher return rates at fixed sites than at mobile sites, controlling for this potential confounder did not eliminate differences we saw between enrolled and nonenrolled subjects.

Another significant factor that has been discussed in several articles pertaining to donor motivation is “intent to return.”^3,5,9 Donors who report in interviews that they intend to return in the future do return to donate at higher frequencies. In the RISE study, all subjects were asked to commit to donating at least once or twice during the following year as a requirement for participation in the study. Asking for this commitment at the time of enrollment in the study may have impacted our subjects’ return rates. None of the blood centers participating in our study ask donors to commit to a minimum number of donations per year, although most encourage donors to set up their next appointment before leaving the center after a donation, and all make telephone calls or send letters or e-mails to remind donors when they are eligible for another donation.

This study has a few limitations. It was not possible to quantify donor motivation, and thus we can only speculate about reasons for donor return since an interview with subjects was not within the scope of this research project. Also, in calculating the rates of return for our study, we did not include attempts at donation in which donors were deferred. Many centers anecdotally reported examples of enrolled subjects who made repeated attempts to donate but were deferred many times for low iron values. The donation frequencies we calculated are based on successful blood cell component visits alone and may not completely capture the full range of donor motivation. As mentioned above, we were unable to compare pre- versus postdonation rates for first-time and reactivated donors. Perhaps a follow-up study could be performed in which donation rates among subjects are analyzed to see if rates continue to remain higher or drop to more typical levels after the end of the study. The retrospective nature of our data did not allow us to perform a quantitative analysis of the impact of subject contact methods on donation frequencies, and future research on this subject should include a more controlled and thorough analysis of that issue. Finally, the specification of enrolled versus nonenrolled subject groups was not exactly comparable. Since there was no access to a donation history interview with nonenrolled donors as there was with the RISE-enrolled subjects, the reactivated donor group was defined slightly differently for enrolled subjects than for nonenrolled donors.