Participant Recruitment in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)

Transcription

1 Participant Recruitment in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) Sara Pressel, MS, Barry R. Davis, MD, PhD, Gail T. Louis, RN, Paul Whelton, MD, MSc, Horacio Adrogue, MD, Debra Egan, MS, MPH, Michael Farber, MD, Gerald Payne, MD, Jeffrey Probstfield, MD, and Harry Ward, MD, for the ALLHAT Research Group The University of Texas Health Science Center School of Public Health, Houston, Texas (S.P., B.R.D.); Tulane University, New Orleans, Louisiana (G.T.L., P.W.); Veterans Affairs Medical Center, Houston, Texas (H.A.); National Heart, Lung, and Blood Institute, Bethesda, Maryland (D.E., G.P.); Pitman Internal Medicine Associates, Pitman, New Jersey (M.F.); University of Washington, Seattle, Washington (J.P.); and King/Drew Medical Center, Los Angeles, California (H.W.) ABSTRACT: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) is a practice-based, randomized, multicenter clinical trial in 42,419 high-risk hypertensive patients aged 55 years and older; 10,356 of these patients are also in a lipid-lowering trial component. The purpose of the antihypertensive component is to determine whether the occurrence of fatal coronary heart disease and/or nonfatal myocardial infarction differs between patients randomized to diuretic (chlorthalidone) and those randomized to either calcium antagonist (amlodipine), angiotensin-converting enzyme inhibitor (lisinopril), or alpha-adrenergic blocker (doxazosin) therapy. (The doxazosin arm has been discontinued.) The purpose of the lipid-lowering component is to determine whether lowering low-density lipoprotein cholesterol with a 3-hydroxymethyl-glutaryl coenzyme A reductase inhibitor (pravastatin) in moderately hypercholesterolemic patients will reduce all-cause mortality compared to a control group receiving usual care. ALLHAT recruited patients from a variety of practice settings from February 1994 through January Sites were paid for randomizations and are paid for Address reprint requests to: Sara Pressel, MS, ALLHAT Project Manager, The University of Texas Health Science Center School of Public Health, 1200 Herman Pressler St., Suite E801, Houston, TX (pressel@utsph.sph.uth.tmc.edu). Received August 9, 2000; accepted June 25, The ALLHAT Collaborative Research Group is listed in JAMA 2000;283: Controlled Clinical Trials 22: (2001) Elsevier Science Inc /01/$ see front matter 655 Avenue of the Americas, New York, NY PII S (01)

2 Participant Recruitment in ALLHAT 675 completed follow-up visits and documented study events. Communication and monitoring were facilitated by nine regional coordinator teams. It was recognized from the outset that patient recruitment would be a very large task because of the number of participants ( 40,000) needed, the ambitious nature of the goal for recruitment of African- Americans ( 55%), and the knowledge that many investigators had limited experience recruiting participants for clinical trials. Multiple adjustments in the initial ALLHAT overall recruitment plan facilitated achievement of sample size goals for both components of the trial. The experience obtained from this large trial should be valuable for the planning and implementation of successful recruitment in future trials. Control Clin Trials 2001;22: Elsevier Science Inc KEY WORDS: Clinical trial, patient recruitment, antihypertensive drug treatment, lipid-lowering treatment INTRODUCTION Recruitment of a sufficient number of motivated participants who meet the study inclusion and exclusion criteria is among the most important requirements for success in a clinical trial [1 8]. The aim of this article is to describe and evaluate the participant recruitment process for the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). The rationale and design of this clinical trial have been previously reported [9]. Briefly, the main purpose of the ALLHAT antihypertensive component is to determine whether the occurrence of fatal coronary heart disease (CHD) and/or nonfatal myocardial infarction differs between patients randomized to diuretic (chlorthalidone) and those randomized to either calcium antagonist (amlodipine), angiotensin-converting enzyme inhibitor (lisinopril), or alpha-adrenergic blocker (doxazosin) therapy. (The doxazosin arm was discontinued early in 2000 [10].) Because of the established benefit of antihypertensive therapy in reducing fatal and nonfatal stroke, morbidity and mortality from cardiovascular disease, and all-cause mortality, the antihypertensive trial component does not include a no treatment or placebo control group. The main purpose of the ALL- HAT lipid-lowering component is to determine whether lowering low-density lipoprotein (LDL) cholesterol with a 3-hydroxymethyl-glutaryl coenzyme A reductase inhibitor (pravastatin) in moderately hypercholesterolemic patients will reduce all-cause mortality compared to a control group receiving usual care. These patients are also in the ALLHAT antihypertensive component. It was recognized from the outset that patient recruitment would be a very large task because of the number of participants ( 40,000) needed, the ambitious nature of the goal for recruitment of African-Americans ( 55%), and the knowledge that many of the investigators had limited experience recruiting participants for clinical trials. Consequently, a Recruitment and Eligibility Subcommittee (a subcommittee of the ALLHAT Steering Committee) developed a recruitment strategy designed to meet the trial s enrollment goals. Regional coordinator teams closely monitored enrollment experience at clinical sites. We describe the multiple adjustments in the initial ALLHAT overall recruitment plan that were necessary to achieve the estimated sample size goals for both arms of the trial. Outlined in subsequent sections are the details of how those recruitment goals were satisfied. The experience obtained from this large simple trial requiring these several and varied adjustments should be valuable

3 676 S. Pressel et al. for the planning and implementation of successful recruitment in similar future trials. METHODS ALLHAT, sponsored by the National Heart, Lung, and Blood Institute (NHLBI) in cooperation with the Department of Veterans Affairs (VA), is a practice-based, randomized clinical trial with 42,419 high-risk hypertensive patients aged 55 years and older; 10,356 of these patients are also in a lipid-lowering trial component. Patient Eligibility Criteria The eligibility criteria for the antihypertensive component were intended to be as inclusive as possible [8]. Hypertensive men and women aged 55 years and older were eligible if they had at least one other risk factor for CHD (e.g., evidence of atherosclerotic disease, type II diabetes). Antihypertensive trial participants were selected for participation in the lipid-lowering component of ALLHAT based on the presence or absence of CHD and LDL-cholesterol levels. Patients without a history of CHD with LDL-cholesterol mg/dl were eligible, as well as patients with CHD who had a baseline LDL-cholesterol mg/dl. Study Organization and Procedures ALLHAT s structure, that of a large, simple trial, differs from most previous NHLBI-sponsored hypertension trials in that most participants are recruited from existing practice-based settings, few procedures are required that are not part of routine care, visit frequency is as expected for routine hypertension care, and few forms are required per participant. The baseline form is four pages long and the routine follow-up visit form is three pages long. The endpoint documentation form, required for all deaths, CHD, stroke, and specific other endpoints, is two pages long. The ancillary forms used are one-page forms and include those for study drug code unblinding, reporting a refusal or loss to follow-up, submitting forms to the clinical trials center, ordering study medications, endpoint quality control, reporting of serious adverse effect, ordering study supplies, requesting patient transfer to another clinical site, and requisitioning blood tests. Sites are compensated for each patient randomized and for follow-up according to a fixed payment schedule. The clinical trials center serves as the data coordinating center and also has additional responsibilities (Figure 1). At the outset of the trial, the clinical trials center was responsible for identifying and paying the clinical sites, for hiring physician investigators and study coordinators (regional coordinators) to oversee recruitment and adherence, and for awarding and supervising subcontracts for a drug distribution center, central laboratory, and electrocardiogram reading center. To oversee the trial and to monitor patient safety and overall study progress, a steering committee of experts in relevant subject areas and a data and safety monitoring board (DSMB) were appointed by the NHLBI. In addition, investigators meetings were held

4 Participant Recruitment in ALLHAT 677 Figure 1 Organizational structure of ALLHAT. every year to 18 months during the recruitment phase to focus attention and provide motivation and education regarding the protocol and trial issues. Weekly reports were generated regarding recruitment for the antihypertensive and lipid-lowering components at each site. Each site was assigned a goal

5 678 S. Pressel et al. for the antihypertensive trial, and weekly recruitment statistics were compared to the sites goals. Based on these goals, the number actually recruited at the site, and the number of weeks left for recruitment, each site was made aware of the number of participants needed per week to achieve their goal. RECRUITMENT EXPERIENCE A variety of recruitment methods evolved in the 4-year enrollment period. Originally, 270 sites were to be recruited, each of which would enroll 150 patients for a total of 40,000 participants. In the original design, 20,000 would be eligible and enrolled for the lipid component. This sample size goal was later reduced to 10,000 based on newly published results of recent lipid-lowering trials [10 12]. In addition, the lipid level inclusion criterion for patients with CHD was changed. ALLHAT has a broad mixture of prototypical clinical sites, including university- or medical center-based, Veterans Hospital-based, and practice-based primary care as well as specialty clinics. Several health maintenance organizations and military clinics also participate. Recruitment efforts varied considerably among them. The practice-based clinics were encouraged to use clinic and hospital chart reviews to identify patients who could be screened for ALLHAT at their next scheduled visit. Some practice-based clinics were able to get referrals from colleagues or other clinics in the same organization. Some implemented community screening projects. University or medical center clinics used chart reviews, referrals, mass mailings, and even community screening efforts and health fairs. Recruitment, which included a 6-month internal pilot phase (Vanguard), started February 1994 and was projected to end July The goal of the Vanguard phase was to recruit 600 patients from February 1, 1994 through July 31, During that time, 19 clinics randomized 483 participants (80.5% of expected), and the DSMB recommended to the NHLBI that the trial should proceed to full-scale. By October 1994, 270 sites were approved and over 200 had initiated recruitment. Only 1218 participants had been enrolled. Study posters and brochures depicting age-appropriate and racially diverse individuals were provided for display in the ALLHAT clinics. Clinical site recruitment was continued with a new goal of 400 clinics and an expected average of 100 participants per clinic. It was emphasized that clinics that could recruit more should do so. In March 1995, the lower age eligibility cutoff was changed from age 60 to 55 years and current cigarette smoking was added as a risk factor to enhance recruitment. About 19% of ALLHAT s randomized cohort was age at randomization, and about 9% qualified as high risk based on current cigarette smoking only. The first studywide investigators meeting was held in April Subsequent investigators meetings were held every year to 18 months during the recruitment phase, and every site was encouraged to have study personnel attend. Sites that were more successful in recruiting ALLHAT participants could bring additional staff to the meetings at the study s expense. National and regional recruitment and quality control awards were presented at these meetings in the form of colored ribbons and plaques.

6 Participant Recruitment in ALLHAT 679 By August 1995 it was still apparent that the goal of 40,000 enrollees would not be achieved by August 1996, so recruitment was extended through July To improve recruitment the reimbursement rate was increased to retroactively reward clinics that were randomizing well and to encourage those just starting or lagging behind. For sites that began randomization prior to April 1995 (several months prior to the reimbursement increase), there was a temporary increase in randomizations that lasted about a month. Although this does not support a sustained relationship between an increased reimbursement rate and increased recruitment efforts, increased reimbursement could have affected a new potential site s willingness to come into the trial as well as their enthusiasm, at least initially, for recruitment. During 1995, a number of clinics requested permission from their institutional review boards to advertise locally to recruit hypertensive patients from outside the practice or clinic. The regional offices assisted clusters of clinics in advertising locally for participants using kit materials prepared by the NHLBI Office of Prevention, Education, and Control. Also, each of the regions carried out special incentive projects, such as contests with prizes going to the clinics recruiting the largest number of patients within certain time limits. By the end of December 1995 (100 weeks into the trial), the ALLHAT leadership decided to recruit a total of 600 sites, and the number of recruiting sites started to increase once again (Figure 2). Sites that were accepted but did not recruit did so because they either withdrew after acceptance, gave up after a period of time after acceptance, or were asked to withdraw because they had not randomized any participants. By July 1996, the end of the second year of recruitment for the full-scale study, there were 564 clinical sites, but recruitment lagged with approximately 19,000 participants enrolled. The recruitment period was extended to January 31, 1998 and other changes were made. Inclusion of clinical sites in Puerto Rico was approved. Additionally, the study received funds from Pfizer, Inc. to fund staff at selected clinical sites to aid recruitment the Field Personnel Program. Figure 2 ALLHAT clinical site selection and participant randomization.

7 680 S. Pressel et al. Interested sites were invited to submit a short application indicating their potential for recruitment, current staffing for ALLHAT, and a plan and budget for additional staff should they be selected to receive funding through this program. Eighty-two sites were selected for the program based on their potential for recruitment and evident need for the extra staff. Funding was provided for 36 full-time equivalents at a total cost of $1,440,000. It was expected that each funded site would be able to recruit two patients per week per full-time equivalent, for a total of 8012 expected participants. The final result was 7273 participants or 91% of expected. Additional changes were made in Inclusion of clinical sites in Canada was approved. Also, a national advertising campaign was undertaken to include radio, print ads, and direct mail letters to patients known to be hypertensive and interested in participating in research studies. In spring 1997, intense advertising efforts were carried out in three waves of 2 weeks each, in different but overlapping areas of the country. Interested patients were advised to call a toll-free telephone number to learn more about ALLHAT and, after basic screening questions, to be referred to the nearest ALLHAT clinical site able to accept referrals. The same toll-free number was also provided on fliers to the ALLHAT clinics and as inserts in information packets sent out from the NHLBI Information Office to persons requesting information on high blood pressure, as well as on posters sent to pharmacies and displayed at community high blood pressure screening sites in communities with ALLHAT clinics. Evaluation of the ALLHAT publicity campaign is presented in Table 1. The campaign started the week of April 11, 1997 and ended 5 weeks later. To estimate the effect of the publicity campaign, we examined the number of participants recruited from the campaign s beginning until 4 weeks after its end. Prior to the campaign, randomizations averaged per week. To evaluate the campaign, it was assumed that randomizations above 250 or 300 per week during the evaluation period were likely due to the publicity campaign. Based on these assumptions, the publicity campaign yielded an additional 1125 to 1575 Table 1 Source ALLHAT Publicity Campaign Evaluation Number of Referrals Cost ($) Cost per Referral ($) Additional Randomized Participants Cost per Additional Randomized Participant ($) Radio 10,917 1,005, , Direct mail 7, , , Newspaper 3, , , Other referral 493 7, Other advertisement 127 1, Poster 92 1, Penny Saver Press 72 1, Doctor referral Weeklies Total 22,904 1,724, a 1, a 1575 is estimated number of additional randomizations due to the publicity campaign over a 9-week period. A more conservative estimate of 1125 was also examined, yielding a cost of $ per additional randomized participant.

8 Participant Recruitment in ALLHAT 681 participants with the resulting costs per participant randomized ranging from $1095 to $1533, respectively (Table 1). The cumulative accrual of ALLHAT sites and patients are displayed in Figure 2. The accrual of ALLHAT sites shows two lines, one for accepted sites and one for sites that remained active over time. Cumulative participant recruitment for both components of the trial is displayed in Figures 3 and 4. The antihypertensive trial patient accrual rate approximated 126 per week for the first year of the full-scale trial and just over 250 per week for the next 2.5 years, with the last month averaging 748 per week. The trial goal of 40,000 was exceeded within that last month by over The lipid-lowering trial patient accrual rate approximated over 25/ week for the first 1.5 years of the full-scale trial, over 50 per week for the next year, and over 75 per week for the last year with the last month averaging 105 per week. The revised study goal of 10,000 was exceed by over 300. Patient Characteristics Patient characteristics are displayed in Table 2. The average age of the randomized patients is 67 years, with 6.5% of the patients aged 80 years and older. About 35% are Black (32% Black non-hispanic); 16% are Hispanic. Women comprise 47% of the cohort. The risk criteria are as follows (not mutually exclusive): atherosclerotic cardiovascular disease 45.0%, ST-T wave changes 10.0%, diabetes 35.7%, low high-density lipoprotein (HDL) cholesterol 12.2%, left ventricular hypertrophy by electrocardiogram 16.1%, left ventricular hypertrophy by echocardiogram 4.1%, cigarette smoking 21.9%. The mean visit 1 blood pressure is 145/83 mm Hg, and the mean visit 2 blood pressure is 146/84 mm Hg. The mean baseline serum cholesterol is 216 mg/dl, LDL-cholesterol 136 mg/dl, HDL-cholesterol 47 mg/dl, and fasting triglycerides 173 mg/dl. Given the relative lack of clinical trial information in African-Americans, it was ALLHAT s original goal that at least 55% of the randomized participants Figure 3 Cumulative recruitment for the antihypertensive trial.

9 682 S. Pressel et al. Figure 4 Cumulative recruitment for the lipid-lowering trial. would be African-American. About 32% of ALLHAT s randomized cohort is Black non-hispanic. Although the 55% goal was not met, 32% is still very good and considerably higher than proportional sampling of the community prevalence would have provided. Still, over half of ALLHAT s randomized cohort is comprised of minority participants. Since ALLHAT was practice-based and most patients were expected to come from existing patient groups already in these practices, successful minority recruitment depended mostly on selection of sites with large minority patient groups. DISCUSSION ALLHAT recruitment initially progressed slowly. Even though ALLHAT had a Vanguard phase that met 80.5% of its recruitment goal, the full-scale enrollment phase presented many challenges and lasted 1.5 years longer than expected. An important factor in the recruitment success of the Vanguard phase was that 58% of the Vanguard clinical sites (11/19) were located at VA facilities or community health centers. These practice types were able to randomize the most patients per site during the Vanguard (22.8 and 34.2 patients per site, respectively, in the Vanguard, compared to an average 11.6 patients per site for other practice types). In the full-scale trial, 119 sites (19%) were VA or community health center facilities. The VA sites continued to be the top recruiters overall (median number of randomizations 77.0) compared to other types of sites including community health centers (range of median randomizations ). There were several major changes made during the course of the enrollment period to meet the overall goal, some of which were changes to the study protocol and the rest intended to enhance site operations. Protocol changes included: (1) increasing the number of participating sites from 270 to 400 and then to 600; (2) adding cigarette smoking as a risk factor and decreasing the

10 Participant Recruitment in ALLHAT 683 Table 2 Baseline Characteristics of ALLHAT Antihypertensive Component Patients Number randomized Age mean (sd) 66.9 (7.7) (%) (%) (%) (%) 6.5 Race/ethnicity White, non-hispanic (%) 47.1 Black, non-hispanic (%) 32.2 White Hispanic (%) 12.5 Black Hispanic (%) 3.3 American Indian/Alaskan (%) 0.2 Asian/Pacific Islander (%) 1.1 Other (%) 3.5 Women (%) 46.8 Years of education mean (sd) 11.0 (4.0) Cigarette smoking Current (%) 21.8 Past (%) 40.3 Never (%) 37.9 Antihypertensive treatment Treated 2 months (%) 86.8 Treated 2 months (%) 3.4 Untreated (%) 9.8 Atherosclerotic cardiovascular disease (%) 45.1 ST-T wave (%) 10.0 Type II diabetes (%) 35.6 Low high-density lipoprotein (HDL) cholesterol (%) 12.1 Left ventricular hypertrophy (LVH) by electrocardiogram (%) 16.1 Left ventricular hypertrophy (LVH) by echocardiogram (%) 4.1 Blood pressure (BP) mean (sd) Visit 1 systolic BP (13.9) Visit 1 diastolic BP 83.3 (9.9) Visit 2 systolic BP (15.6) Visit 2 diastolic BP 84.0 (10.1) Serum potassium mean (sd) meq/dl 4.4 (0.7) Fasting serum glucose mean (sd) mg/dl (57.1) Serum creatinine mean (sd) mg/dl 1.0 (0.3) Serum cholesterol mean (sd) mg/dl (43.3) LDL-cholesterol mean (sd) mg/dl (37.0) HDL-cholesterol mean (sd) mg/dl 46.8 (14.7) Fasting triglycerides mean (sd) mg/dl) (134.8) lower age eligibility cutoff from 60 to 55 years; (3) extending recruitment by 1 year through July 1997 and then by 6 more months through January 1998; and (4) expanding site recruitment to Puerto Rico and Canada. Changes intended to affect site operations included: (1) increasing clinic reimbursements for enrollment of more than 25 patients; (2) initiating a Field Personnel Program to assist selected clinics; and (3) instituting a nationwide advertising campaign. Two major changes were made to the lipid-lowering trial. These were changing the lipid level inclusion criterion for persons with coronary heart disease and decreasing the sample size from 20,000 to 10,000.

11 684 S. Pressel et al. Increasing the number of participating sites. It was clear that our original estimate of 270 clinical sites recruiting 40,000 participants (about 150 participants each) was optimistic. Our interim goal of 400 sites required 100 participants per site and that, too, was deemed to be unrealistic as recruitment progressed. Although ALLHAT s recruitment goal was met, it took 623 sites recruiting an average of 68 patients each to do so. Seventy-two sites reached the original study goal of 150 or more participants. Adding cigarette smoking as a risk factor, and decreasing the lower age eligibility cutoff from 60 to 55 years. About 9% of the recruited cohort qualified as high risk based on current cigarette smoking only. About 19% were aged at randomization. Evaluating the eligibility criteria and making justifiable changes without harming the protocol could be an effective tool to enhance patient recruitment. Extending recruitment twice, eventually through January Although these recruitment extensions were necessary to achieve ALLHAT s recruitment goals, it was also necessary to extend follow-up 1 year to preserve study power. Expanding site recruitment to Puerto Rico and Canada. Thirty Canadian sites randomized 700 participants (about 23 participants each). Although it appears that expansion to Canada did not have a large impact on recruitment, those sites began randomizing just 11 months prior to the end of recruitment. The specific effort to bring sites from Puerto Rico into the trial began in 1995, and the first participant from this major group of Puerto Rican sites was randomized about 2 years prior to the end of recruitment. Thirty-three Puerto Rican sites randomized 6022 participants (about 182 participants each). Expanding the geographic area for recruiting clinical sites, while requiring careful attention to regulatory and language challenges, can be a valuable tool for enhancing patient recruitment. Increasing clinic reimbursements for enrollment of more than 25 patients. The original reimbursement scheme was designed to cover the effort devoted to identifying and consenting patients, as well as time spent completing forms. It was anticipated that increased reimbursements would cover more of this time and positively affect recruitment long term. There was some short-term improvement, but this was not sustained. It is difficult to assess what the effect of higher reimbursement rates might have been from the beginning of the trial or what effect this increase actually had on subsequent site recruitment and recruitment rates at those sites. Consideration of any reimbursement scheme, or changes to a reimbursement scheme, must consider the complexity of the protocol, estimated effort required per patient, the number of required patients, the time frame allowed for the protocol, and the impact on total study costs. Initiating a Field Personnel Program to assist selected clinics. Eighty-two sites were funded for extra staff based on recruitment potential and evident staffing needs. This effort was quite successful, as these sites randomized 91% of their expected number of participants during the program. The additional funding enabled these sites to dedicate significant staff time (up to one full-time equivalent) to recruitment efforts. If planned reimbursements are relatively low, this type of program can encourage sites with significant recruitment potential to concentrate their efforts during a crucial trial phase. However, consideration also must be given to providing some support after the recruitment period ends for follow-up

12 Participant Recruitment in ALLHAT 685 of the extra randomized participants. The Field Personnel Program in ALLHAT continued through the end of the trial. Instituting a nationwide advertising campaign. A centrally managed advertising campaign can be very costly. It is estimated that ALLHAT recruited an additional 1575 participants through such a campaign (about $1100 per participant in addition to the routine cost of determination of eligibility and randomization). A decision whether or not to include an advertising campaign in a trial s recruitment efforts depends on other patient sources and their relative costs per patient, as well as the sites preparedness for handling these types of recruitment referrals. The two changes to the lipid trial protocol (narrowing the LDL-cholesterol inclusion range for participants with CHD and reducing the necessary sample size from 20,000 to 10,000) balanced each other. The first change reduced the number of eligible participants for the lipid trial, and the second reduced the need for eligible participants. It is difficult to compare ALLHAT s experience with that of other trials since ALLHAT is in several ways unlike any other trial conducted in the United States and Canada. First, ALLHAT is the largest hypertension trial ever conducted and has hundreds of clinical sites; only a few trials have had this many clinics. Further, ALLHAT has a long follow-up period and includes two separate treatment protocols. Finally, many clinics with little or no research experience are involved and following an ethnically diverse patient population. In spite of these challenges, recruitment in ALLHAT was successful. Tremendous effort and support were required from the project office, clinical trials center, regional coordinators, and the clinical sites. In ALLHAT, recruitment took longer and required more resources than planned, and expectations were revised several times. The push was continual and demanding, requiring a very creative and constant effort as well as several major protocol changes for the antihypertensive trial and for the lipid trial. Despite the special challenges of clinic and patient recruitment in the trial design, the ALLHAT investigators accomplished their task, with a final enrollment of 42,419 participants with hypertension plus an additional risk factor. The changes intended to affect clinic operations helped at least moderately with recruitment by adding to the numbers of patients but also provided an intangible boost to study morale needed to reach the final recruitment goals. The changes to the study protocol were much more effective, resulting in desired increases to the pace of patient recruitment. Recruitment is a challenge in any clinical trial, regardless of size, scope, or funding. If it becomes clear that recruitment goals will not be met in the prescribed time frame, consideration should be given to any of the above strategies, particularly protocol-related changes such as expanding the number of clinical sites, extending the time frame for patient recruitment, and making appropriate changes to patient eligibility criteria. This study is supported by a contract with the National Heart, Lung, and Blood Institute (NHLBI) federal contract N01-HC The ALLHAT investigators acknowledge contributions of study medications supplied by Pfizer Inc. (amlodipine and doxazosin), AstraZeneca (atenolol and lisinopril), and Bristol-Myers Squibb (pravastatin), and financial support provided by Pfizer to the NHLBI.

13 686 S. Pressel et al. REFERENCES 1. Hollis JA, Satterfield S, Smith F, et al. Recruitment for a Phase II trial of hypertension prevention. Effective strategies and predictors of randomization. Ann Epidemiol 1995;5: Whelton PK, Bahnson J, Appel LJ, et al. Recruitment in the Trial of Nonpharmacologic Intervention in the Elderly (TONE). JAGS 1997;45: Whelton PK, Lee JY, Kusek JW, et al. Recruitment experience in the African-American Study of Kidney Disease and Hypertension (AASK) Pilot Study. Control Clin Trials 1996;16:17S 33S. 4. Petrovitch H, Byington R, Bailey G, et al. Part 2: Screening and recruitment. Hypertension 1991;17(Suppl 2):II-16 II Satterfield S, Borhani NO, Whelton P, et al. Recruitment for Phase I of the Trials of Hypertension Prevention. Am J Prev Med 1993;9: Lovato L, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruitment for controlled clinical trials: Summary and annotated bibliography. Control Clin Trials 1997;18: Anonymous. Rationale, design, implementation, and baseline characteristics of patients in the DIG trial: A large, simple, long-term trial to evaluate the effect of digitalis on mortality in heart failure. Control Clin Trials 1996;17: Carew B, Ahn S, Boichot H, et al. Recruitment in the Studies of Left Ventricular Dysfunction (SOLVD): Strategies for screening and enrollment in two concurrent but different trials. Control Clin Trials 1992;13: Davis BR, Cutler JA, Gordon DG, et al. Rationale and design for the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Amer J Hypertens 1996;9: Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: The Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344: Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. West of Scotland Coronary Prevention Study Group. N Engl J Med 1995;333: Sacks FM, Pfeffer MA, Moye LA, et al. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med 1996;335: