The safety of the nation s blood supply depends primarily

Transcription

1 B L O O D D O N O R S A N D B L O O D C O L L E C T I O N COMPUTER-ASSISTED HEALTH INTERVIEWING Computer-assisted audiovisual health history self-interviewing Results of the pilot study of the Hoxworth Quality Donor System Thomas F. Zuck, Paul D. Cumming, and Edward L. Wallace BACKGROUND: The safety of blood for transfusion depends, in part, on the reliability of the health history given by volunteer blood donors. To improve reliability, a pilot study evaluated the use of an interactive computer-based audiovisual donor interviewing system at a typical midwestern blood center in the United States. STUDY DESIGN AND METHODS: An interactive video screening system was tested in a community donor center environment on 395 volunteer blood donors. Of the donors using the system, 277 completed surveys regarding their acceptance of and opinions about the system. RESULTS: The study showed that an interactive computer-based audiovisual donor screening system was an effective means of conducting the donor health history. The majority of donors found the system understandable and favored the system over a face-to-face interview. Further, most donors indicated that they would be more likely to return if they were to be screened by such a system. CONCLUSION: Interactive computer-based audiovisual blood donor screening is useful and well accepted by donors; it may prevent a majority of errors and accidents that are reportable to the FDA; and it may contribute to increased safety and availability of the blood supply. ABBREVIATIONS: HQDS = Hoxworth Quality Donor System; NAIS = nurse-administered interview system. From the Hoxworth Blood Center, University of Cincinnati, Cincinnati, Ohio; Center for Management Systems, Naples, Florida; and Talisman Ltd, Vienna, Virginia. Address reprint requests to: Paul D. Cumming, PhD, Talisman Ltd., 1740 Burning Tree Drive, Vienna, VA 22180; paul.cumming@talmed.com. Supported in part by Grant #HL from the National Heart, Lung, and Blood Institute. Received for publication March 23, 2001; revision received July 16, 2001, and accepted July 19, TRANSFUSION 2001;41: The safety of the nation s blood supply depends primarily on the recruitment of low-risk donors and the effective testing of donated blood. Although estimated transfusion risks from known pathogens are extremely small, 1,2 risks remain when donations are made within the window period and if pathogens are present for which no tests exist. Minimizing these risks depends on the selection of donors whose demographic and personal behavior characteristics place them at low risk for blood-borne diseases. The effectiveness of current blood donor recruitment is evidenced by the low prevalence of known pathogens among first-time donors. Rates of confirmed-positive markers of blood-borne diseases among first-time donors range from 0.3 percent for HIV to 1.8 percent for HCV; these rates are much lower than those among the general population. 3-6, selective recruitment, and donor screening have been responsible for these low rates of infection, which have the additional benefit of reducing transfusion costs. Three major improvements in transfusion safety have been made over the last 30 years: recruitment only of volunteer donors in the 1970s; verbal screening and testing of donors in the 1980s and 1990s; and imposition by the FDA in 1992 of strict adherence to good manufacturing practices in licensed blood establishments, including standardized questioning of blood donors regarding health-related risk factors. Clearly, the effectiveness of current screening methods depends upon the willingness of donors to respond truthfully to questions. Yet, when queried, a small percentage of donors lie about their own risky behavior. 7,8 Thus, improving donor understanding of the substance of questions and increasing their willingness to respond truthfully remain unfinished tasks. In 1992, the FDA sponsored a test of a computer-assisted, interactive, donor-screening procedure. 9 In spite of a flawed design, the study provided evidence of the benefits obtainable from computer-assisted systems for interactive donor screening. Other recent health-related surveys of risky personal behavior substantiated the hypothesis that, properly utilized, computer-assisted systems improve subjects understanding of questions and increase the truthfulness of their responses Volume 41, December 2001 TRANSFUSION 1469

2 ZUCK ET AL. In 1992, one of us (PDC) began developing a computerassisted system for health interviews of blood donors, to improve consistency and reduce or eliminate screening errors. In cooperation with a moderate-sized community blood center (Hoxworth Blood Center), a pilot study to test the effectiveness of the system and its acceptance by donors was initiated. The results of that study are presented here. MATERIALS AND METHODS The screening system The system tested, designated the Hoxworth Quality Donor System (HQDS), consisted of software installed on either stand-alone desktop computers or portable personal computers (PCs) for use at fixed sites or on mobiles. The PCs were fitted with audio and video capabilities, including earphones and touch screens, respectively. Questions were presented in both audio and color video formats, and each question was accompanied by a picture of a Hoxworth staff member and depictions of the types of behavior or other factors related to each question. Figure 1 shows a typical screen. A series of touch boxes located beneath each screen permitted the donor to answer a question, to advance to the next question, or to return to the previous question; there was also a help box to summon a nurse. The system visually presented results of each donor s response to every question for a nurse s review and printed a copy of the final responses, including free-form comments. The system also maintained a continuous record, called the Log, of every action taken by each donor and staff reviewer at every interview; this record included the time taken for each action. The system was installed in the main donor room of the blood center at a single station at which 397 random donors were requested to test the system; 395 chose to participate the two refusals were due to the additional time required to complete the study process. Fig. 1. Medical history interview screen from the HQDS used in the pilot study at Hoxworth Blood Center, Reprinted with permission from Talisman Ltd TRANSFUSION Volume 41, December 2001 The postdonation survey While 395 donors used the HQDS, because of staff omissions, only 277 were asked to respond to the written postdonation survey. Included in the survey questionnaire were four queries seeking the donors opinions as to the clarity and understandability of the questions posed by the system, the time required by and acceptability of the system, and the likelihood of truthful responses to sensitive issues in a computer interview. Repeat donors were asked additional questions about their preference for the HQDS or the conventional nurse-administered interview system (NAIS), the understandability of questions posed by the two systems, and the likelihood that the donor will return if screened by one system or the other. Donors were asked to answer each question by using a system of scores from 1 to 5. Responses of 1 or 2 were considered favorable to HQDS, those of 4 or 5 to be favorable to NAIS (i.e., unfavorable to HQDS), and those of 3 to be neutral. In addition, all donors surveyed were invited to comment in writing on their experiences with the HQDS. A copy of the survey questionnaire is shown in Fig. 2. Fig. 2. Donor survey form used in the pilot study of the HQDS at Hoxworth Blood Center, 1999.

3 COMPUTER-ASSISTED HEALTH INTERVIEWING TABLE 1. Numbers and percentages of donors scores by survey question Scores Question No number response Totals Part A 1 Number Percentage Number Percentage Number Percentage Number Percentage Part B 1 Number Percentage Number Percentage Number Percentage The donor group Respondents were predominately white (92.5%), college- or postgraduate-educated (91.3%), and more or less equally split between male and female; 94 percent were repeat donors. They were divided into three groups by age: 17 to 30 years old, 31 to 44 years old, and 45 or more. Written comments were received from 39 percent of respondents. Log data were available on the actions taken and times required by all donors. Statistical evaluation The statistical method used to analyze survey data was the chi-square procedure with 1 df. Chi-square and p values were determined mathematically by using the chi-square formula and statistical tables. 16 RESULTS The time required for a donor to complete the HQDS medical history, with full audio on all questions, as well as the consent form, and to wait for review, averaged 12 minutes, as compared with 5 minutes for the NAIS. A nurse s review of the HQDS required another 2 minutes. Results of the surveys, including analyses by demographic groups, are presented in Tables 1 and 2. Part A Part A was to be completed by all donors. Although 95 percent of survey respondents were repeat donors, their answers reflected their opinions after their initial exposure to HQDS. Ninety-two percent found the system to be clear and understandable, and 95 percent felt comfortable with the process and the privacy provided. It was surprising that, although completion of the HDQS was more time-consuming than that of the NAIS, 64 percent were satisfied with the time required to complete the former. As shown in Table 2, donors who owned or used computers found the time required to complete the HQDS to be somewhat less acceptable (61.4% satisfied) than did those who did not own or use computers (83.9% satisfied). Donors with a high school education were more satisfied with the time required (91.7%) than were those with a college (74%) or postgraduate (53.6%) education. Donors with a high school education (91.7%) were more convinced than were college graduates (68.3%) or postgraduates (61.8%) that donor responses to sensitive questions posed by computers were more TABLE 2. Percentages of responses to survey questions by donor demographic characteristics* Part A Questions Part B Questions Demographic characteristics Computers Uses Does not use Sex Male Female (years) Age * Scores grouped by favorable (1+2), neutral (3), and unfavorable (4+5), expressed as percentages. Volume 41, December 2001 TRANSFUSION 1471

4 ZUCK ET AL. TABLE 3. Chi-square analysis of differences in frequencies of favorable (1+2) and unfavorable (4+5) scores on survey questions, with neutral (3) responses deleted Question number (1+2) (4+5) Totals Chi-square p value Part A < < < <0.001 Part B < < <0.001 likely to be truthful. Younger donors (77.8%) were more convinced of donor truthfulness in responses to a computer than were the two groups of older donors (64.0% and 64.5%). Chi-square analyses of differences in frequencies of donor responses to the four survey questions on Part A are shown in Table 3. The frequencies of responses favorable to HQDS were significantly greater than those of unfavorable responses (p<0.001). In spite of relatively large percentage differences in favorable responses to the questions about truthfulness and time, only those between-group differences in responses to questions about time were significant. Computer nonusers found the additional time required by HQDS more acceptable than computer users, but the difference was just barely significant. Largely because of the small numbers TABLE 4. Comparative analyses of favorable and unfavorable responses to selected donor survey questions by demographic groups with 1 df Demographic Chi-square p value Question groups value <0.05 A2 Computers Owns/uses versus does not own/use 4.0 Yes 8-12 vs No 8-12 vs Yes vs Yes A vs No 8-12 vs No Age group vs No vs No B vs No 8-12 vs No Age group vs No vs Yes vs No B vs No 8-12 vs No B vs No 8-12 vs No vs Yes in the groups (Table 4), differences among donors with high school (8-12 years of school), college (13-16), and postgraduate (17+) education were significant. Part B Part B was to be completed by repeat donors only. The numbers of neutral responses by repeat donors to the three questions in Part B indicated that substantial numbers were indifferent to the relative merits of the two interviewing systems. However, 64 percent of responding repeat donors who stated a preference for one of the systems preferred the HQDS, and 90 percent found the HQDS more understandable (Table 1). While most repeat donors stated they were equally likely to return whether or not a change from NAIS to HQDS was made, 80 percent of those with a preference indicated they were more likely to return after having been screened by HQDS. A few donors, primarily in the oldest and postgraduate-educated groups, stated they would be less likely to return, given a change from NAIS to HQDS (Table 2). Significant differences in favorable and unfavorable responses by repeat donors to the questions in Part B of HQDS were found between the youngest and the oldest donors on Question 1 (preference) and between college- and postgraduate-educated donors on Question 3 (likelihood of return), as shown on Table 4. There was a significant difference in the frequencies of favorable and neutral responses on Question 3 by the college- and postgraduate-educated groups, with favorable frequencies of the latter group being significantly fewer and more neutral (analysis not shown). All other differences between favorable and unfavorable and favorable and neutral responses among the groups were not significant. Information from the Log Detailed information about the actions of the 395 donors who used the HQDS was captured automatically on the HQDS Log, and it reflected the following: the system was used mainly by repeat donors (94%) who changed only 82 (0.2%) of their initial answers during the interview. In the nurse-administered review that followed the interview, nurses changed 980 (6%) of the donors initial answers. Questions accounting for most of the nurse-administered changes were those on medications; recent surgery, injections, or medications; prior refusal as a blood donor; international travel; and effectiveness of AIDS tests. A nurse s review of donors answers required 539 referrals by nurses to 24 deferral codes. Physician care, medications, vaccinations or immunizations, malarial area travel or malaria medication, and prior deferral accounted for 85 percent of the donor deferrals. DISCUSSION The effectiveness of a health history interviewing system depends on its acceptance by and assistance to donors and 1472 TRANSFUSION Volume 41, December 2001

5 COMPUTER-ASSISTED HEALTH INTERVIEWING its ability to protect both the safety and availability of blood. While the number of participants in the pilot study was small and while their characteristics were atypical of donors in general, the results of the study indicated that computerassisted donor interviewing could be effective. Study findings were consistent with those of other recent studies of the effectiveness of computer-assisted interviewing of subjects in situations where truthfulness and accuracy regarding risky behavior were of paramount importance. Donor acceptance and assistance Virtually all donors were comfortable with the HQDS and the privacy it provided. This is important, for privacy has been identified by many investigators as key to the retention of donors and to improving the truthfulness of their responses. 10,11,17-20 The HQDS also assisted donor understanding of the interview questions in several ways: by video presentation reinforced with simultaneous private audio, a capability important to less literate donors 21,22 ; a help option to summon a nurse for assistance; and a final comprehensive review of donor responses. While a majority of repeat donors found the HQDS and NAIS equally understandable, 43 percent found the questions more easily understood when presented by the HQDS. To be effective, an interview system should be comfortable and easy to use without undue delay. 17 While the majority of repeat donors either preferred HQDS (41%) or were indifferent (36%), among the demographic groups, younger (48%) and less educated (64%) donors were more likely to prefer HQDS. The latter findings are reasonably consistent with those of a recent survey by Watanabe et al. 19 Almost twothirds of all donors studied were satisfied with the 14 minutes required to complete the interview; however, almost half of all written survey comments stated that the HQDS took an unnecessarily long time to complete. A 2- to 3-second audio version of the system is available and permitted by the FDA, which would reduce the time required almost to that needed for the NAIS. Safety and availability Donor health history interview systems affect the safety of the blood supply by promoting donor truthfulness, improving donor understanding of the questions, and reducing staff errors and omissions, with truthfulness being of paramount importance. Donors lie for various reasons, principally because they are seeking testing. Direct evidence of donor lying has been provided by Williams et al., 7 who found that approximately 2 percent of all donors withheld information on disqualifying behavioral and other risk factors. Mahl et al., 8 in a study of young German blood donors, reported similar findings, as did Locke et al. 11 in a 1992 study of American Red Cross donors, and an American Institute for Research study 9 in which computer-assisted interviewing was used first. Most donors surveyed in our study believed that donors would be more likely to answer personal and sensitive questions truthfully if asked by a computer. Two recent donor surveys by Watanabe et al. 10,19 reported that donors believe computer-assisted interviewing would elicit more truthful responses as to their risky behaviors, especially from young donors, test seekers, and first-time donors. Turner et al., 12 in a survey of male teenagers that employed risk-behavior questions similar to those asked blood donors, found answers obtained by computer-assisted interviewing to be 2 to 14 times more truthful than those obtained by a self-administered paper questionnaire. Other investigators have reported similar findings. 14,15 The ability of computer-assisted interviewing to improve donor understanding of the questions was evidenced by the fact that 43 percent of repeat donors in the study found questions posed by the HQDS more understandable than those posed by the NAIS, whereas 5 percent found them less so. The HQDS was designed to reduce staff errors and omissions. While the pilot study was too small to test this goal, no such events occurred during the study. Electronic discipline imposed by a computer-assisted system such as the HQDS does not permit staff errors and omissions to go uncorrected. A donor health history interview system affects blood availability through its effect on the frequency of donor return. In the pilot study, while most repeat donors stated they were likely to donate again, whatever the system, 80 percent of the remaining one-third (27.2% of all repeat donors) stated they were more likely to donate again if the HQDS were used. This opinion was most strongly held by the younger and less educated repeat donors. The pilot study demonstrated that a computer-assisted audiovisual self-interviewing system such as HQDS is an effective means of conducting the donor health history interview. While results of the pilot study were reassuring, a much larger study involving a more typical mix of donors is required to validate study findings. Such a study is now underway at three major blood centers supported by the National Heart, Lung, and Blood Institute. ACKNOWLEDGMENTS The authors thank the staff of the Hoxworth Blood Center for their assistance in conduct of the pilot study, in particular, Patricia M. Carey, MD, Terri Neff, MS, and Daniel Meurer. They also thank Matthew Brashears, Talisman Ltd., for his assistance in analyzing study data and reviewing the manuscript. REFERENCES 1. Kleinman S, Busch MP, Korelitz JJ, Schreiber GB. The incidence/window period model and its uses to assess the risk of transfusion-transmitted human immunodefi- Volume 41, December 2001 TRANSFUSION 1473

6 ZUCK ET AL. ciency virus and hepatitis C virus infection. Transfus Med Rev 1997;11: Schreiber GB, Busch MP, Kleinman KH, Korelitz JJ. The risk of transfusion-transmitted viral infections. The Retrovirus Epidemiology Donor Study. N Engl J Med 1996;334: Karon JM, Rosenberg PS, McQuillan G, et al. Prevalence of HIV infection in the United States, 1984 to JAMA 1996;276: Coleman PJ, McQuillan GM, Moyers LA, et al. Incidence of hepatitis B infection in the United States, : estimates from the National Health and Nutrition Examination Surveys. J Infect Dis 1998;178: Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1998 through N Engl J Med 1999;341: Glynn SA, Kleinman SH, Schreiber GB, et al. Trends in incidence and prevalence of major transfusion-transmissible viral infections in US blood donors, 1991 to Retrovirus Epidemiology Donor Study (REDS). JAMA 2000;284: Williams AE, Thomson RA, Schreiber GB, et al. Estimates of infectious disease risk factors in US blood donors. Retrovirus Epidemiology Donor Study. JAMA 1997;277: Mahl MA, Hirsch M, Sugg U. Verification of the drug history given by potential blood donors: results of drug screening that combines hair and urine analysis. Transfusion 2000;40: Increasing the safety of the blood supply by screening donors more effectively: executive summary, final report. Rockville, MD: FDA, Watanabe KK, Schreiber GB, Hayes A, Williams AE. Factors influencing the recruitment and retention of young donors (abstract). Transfusion 1999;39(Suppl):125S. 11. Locke SE, Kowaloff HB, Hoff RG, et al. Computer-based interview for screening blood donors for risk of HIV transmission. JAMA 1992;268: Turner CF, Ku L, Rogers SM, et al. Adolescent sexual behavior, drug use, and violence: increased reporting with computer survey technology. Science 1998;280: Bloom DE. Technology, experimentation, and the quality of survey data. Science 1998;280: Kissinger P, Rice J, Farley T, et al. Application of computerassisted interviews to sexual behavior research. Am J Epidemiol 1999;149: Des Jarlais DC, Paone D, Milliken J, et al. Audio-computer interviewing to measure risk behaviour for HIV among injecting drug users: a quasi-randomised trial. Lancet 1999;353: Agresti A. Categorical data analysis. New York: John Wiley & Sons, Thomson RA, Bethel J, Lo AY, et al. Retention of safe blood donors. Transfusion 1998;38: Kline L, Friedman LI, Dempesy D, et al. Assessment of blood donor privacy during health history interviews. Transfusion 1996;36: Watanabe KK, Zuck T, Schreiber GB, et al. Computerized donor screening: perceived advantages and disadvantages among blood donors (abstract). Transfusion 2000;40(Suppl):4S. 20. Boekeloo BO, Schiavo L, Rabin DL, et al. Self-reports of HIV risk factors by patients at a sexually transmitted disease clinic: audio vs written questionnaires. Am J Public Health 1994;85: Williams MV, Parker RM, Baker DW, et al. Inadequate functional health literacy among patients at two public hospitals. JAMA 1995;274: Miles D, Davis T. Patients who can t read. Implications for the health care system. JAMA 1995;274: TRANSFUSION Volume 41, December 2001