Recruitment in the Hypertension Detection and Follow-up Program

5 Recruitment in the Hypertension Detection and Follow-up Program Charles E. Ford, Herbert G. Langford, and Melanie J. Palmer

ABSTRACT: The Hypertension Detection and Follow-up Program (HDFP), through a twostage population-based screen for high blood pressure involving over 159,000 persons in 14 U.S. communities, recruited and randomized 11,386 participants. The design initially called for the recruitment of 800-1000 hypertensive individuals per clinic over a 9-month period. The size of the population to be screened was based on an estimated 15% of the general population being age and blood pressure (BP) eligible. All adults 30-69 years of age in defined populations were to be screened, with efforts to enroll into the trial all who met the BP eligibility criteria. Recruitment was carried out by interviewers who enumerated households and screened age-eligible individuals for elevated BP; those meeting BP criteria were invited to a clinic screen for confirmation. Of those persons enumerated, 179,239 were age eligible, and 159,566 of these had home BP measurements. The home screening yielded 22,994 persons who had a diastolic BP at or above 95 mm Hg, and 17,499 attended the clinic screen. At the clinic screen, 11,386 had diastolic BP at or above 90 mm Hg and were randomized into one of two treatment groups. The enumeration goal, and the recruitment, yield and cost varied by clinic, primarily due to different characteristics of the defined populations. Overall, the target population approximated U.S. demographic characteristics except that a larger proportion of blacks was included. Because of generally good response rates, the randomized study population was representative of similarly aged hypertensive persons in the target populations. The recruitment results suggest that a population-based screening program for a high prevalence condition such as hypertension is an effective mode of recruitment, particularly when a population-based sample is required. KEY WORDS: recruitment, population-based sample, two-stage screen, hypertension

INTRODUCTION The H y p e r t e n s i o n Detection a n d Follow-up P r o g r a m (HDFP) was a mult i c e n t e r r a n d o m i z e d 5-year clinical trial designed to assess the effectiveness of rigorous systematic antihypertensive d r u g t h e r a p y in reducing morbidity and mortality a m o n g p e r s o n s with high blood pressure. The s t u d y was initiated early in 1971 with a request for proposal (RFP) from the National Heart, Lung, a n d Blood Institute (NHLBI). It was f u n d e d u n d e r contract b y the NHLBI a n d was c o n d u c t e d in two phases, as specified in the RFP. The first

54S 0197-2456/87/$3.50

Controlled Clinical Trials 8:54S-67S (1987) © Elsevier Science Publishing Co., Inc. 1987 52 Vanderbilt Ave., New York, New York 10017

Recruitment in the HDFP

55S

phase (1971-1972) involved pilot projects and planning efforts for the HDFP clinical trial, conducted in the second phase (1973-1979). To facilitate the extrapolation of study findings to the general hypertensive population, the HDFP design incorporated a community-based screen for high blood pressure to identify and recruit participants for the clinical trial.

OBJECTIVES The HDFP was designed to answer several questions: 1. Could persons with high blood pressure, detected in a community screening program, be brought under pharmacologic management and kept under such management long term? 2. Would systematic antihypertensive drug treatment result in a significant reduction in all-cause mortality over a 5-year period? 3. Would such treatment reduce the occurrence of cardiovascular events such as stroke or coronary heart disease incidence? 4. Would pharmacologic therapy be as effective in the whole spectrum of those with elevated BP, including women, young, middle-aged, and older adults, and persons with and without concomitant target organ disease. In support of the main HDFP objectives, the screening and recruitment efforts had several specific goals: 1. The collection of demographic data on the target population. 2. The determination of the prevalence of hypertension and the status of high-blood-pressure detection, treatment, and control in the screened communities. 3. The assessment of blood pressure screening results in a variety of community settings. 4. The enrollment in the trial of as many persons as possible who were found to be eligible.

METHODS Pilot Study The pilot study phase of the HDFP involved nine clinical centers and the coordinating center. Its major purposes were: 1. To work out trial planning details. 2. To test alternative approaches to recruitment. 3. To determine the yield of hypertensive persons from various population bases. 4. To develop procedures for conducting the screen. 5. To develop a common protocol for the clinical trial phase. Although the RFP indicated that clinic applicants should be capable of supplying at least 800-1000 hypertensive persons from community-based populations, it did not detail how this should be done. Thus, the clinics partic-

56S

c.E. Ford, H. G. Langford, and M. J. Palmer ipating in the pilot study proposed and tested various methods of recruitment and ways of selecting populations to be screened. The pilot study was carried out between May 1971 and May 1972. It did not involve randomization of participants nor were any of the pilot study participants--or data collected on them--included in the HDFP clinical trial. As a result of the pilot study, the HDFP Steering Committee in cooperation with the NHLBI project office and the coordinating center staff determined that: 1. The trial sample size would be 10,500 persons. 2. There would be a 9-month screening and recruitment period. 3. Additional clinics would be required to supply a population base large enough to recruit the sample in this timeframe. Subsequently, five additional clinics, a central biochemistry laboratory, and an electrocardiogram reading center were contracted to participate in the HDFP.

Design of the Clinical Trial The design, methods, and baseline findings of the HDFP have been reported previously [1,2]. Briefly, the HDFP was planned as a multicenter clinical trial with a sample size requirement of 10,500 participants. The trial used a method of organization similar to that recommended by a committee of the

SaltLake City

Chicago CentralLaboratory Minneapolis \\ , t Lansing EGGCenter ~, ' ~ ! ~ B

c_ i

Davis 1

I " ~ "" l ~ " ' N e w York t./~ ~ f ~ ~ Baltimore ~._ ~ "Georgetown f----~ ; ~-. Y

LosAngeles

~

(Coordinating Center)

oston

Atlanta EvansCounty

Jackson Birmingham

Figure 1. Geographic location of Hypertension Detection and Follow-up Program clinics and support centers.

Recruitment in the HDFP

57S

National Advisory Heart Council [3,4]. Recruitment for the study was accomplished through blood pressure screening of all persons 30-69 years of age in defined populations of varied composition in 14 U.S. communities, shown in Figure 1. Screening for the clinical trial was conducted from January 1973 through May 1974 [5]. Most persons were first screened at home for elevated blood pressure. (One clinic conducted its first screen in the work place of selected industries.) Persons with suspected hypertension then under went a second screen in HDFP clinics. Based on random allocation to one of two treatment groups, subsequent treatment for hypertension was given either at HDFP clinics (stepped care) or by referral to an existing source of medical care in these communities (referred care). Trial follow-up terminated in November 1979, when all randomized participants had completed a full 5 years in the trial and the results of the study were reported [6].

METHODS OF ENUMERATION A N D SCREENING

Three recruitment methods were refined from the pilot study for use in the clinical trial phase. The first method called for the screening of an entire geographical area, for example, entire census tracts, a housing project, or an enumeration district, and was used by 9 of the 14 clinics. The second method involved the selection of a random sample of all census tracts in a defined geographical area or of the households in a city or county (used by three clinics). The third method involved screening the entire employee roster of selected industries (used by one clinic) [7]. Each clinic conducted a census of the population selected for screening. The purpose of the census enumeration was to obtain demographic data on household residents for description of the target population and to identify persons w h o were between 30 and 69 years of age. A screening of all ageeligible persons by 13 of the 14 clinics then followed. One clinic screened only the age-eligible employees of the target industries. (Persons who were terminally ill or institutionalized---correctional institutions, nursing homes; hospitals, or homes for the aged, infirm, and needy, within the target area-were not screened.) The screen included blood pressure measurements and a brief interview covering demographic, social, and health status questions [2]. Interviewers were trained to conduct the screening procedures in a standardized manner [8]. All individuals contacted by the HDFP interviewers were given some information about the program and its purpose. Persons with elevated blood pressure at the first screen were further advised of the implications of a "single elevated reading," were invited to attend a clinic screen, and were informed of what additional procedures would be offered them if they were found to have high blood pressure at that time. Only if the participant (screening participant) asked whether the program would provide treatment if he had high blood pressure was the subject of treatment discussed--and then the participant was told that the program would encourage all persons with high blood pressure to get treatment but would only provide such treatment for

58S

c.E. Ford, H. G. Langford, and M. J. Palmer half of those found to have high blood pressure. Further explanations of the objectives of the study were then made. When blood pressure was measured, if the mean of the second and third fifth-phase diastolic blood pressure (DBP) readings was 95 mm Hg or higher, the individual was invited to the clinic for a confirmation blood pressure screen. If the mean DBP reading at the clinic was 90 mm Hg or higher, the individual was randomly assigned to one of the two treatment groups: stepped care or referred care [2].

Randomization Randomization was stratified by clinic and by entry DBP stratum. The coordinating center prepared and sent to each clinic three sets of sequentially labeled and sealed opaque envelopes containing the randomization assignments; the sets corresponded to the three entry DBP strata. When a person was confirmed to be hypertensive, the next sequential envelope from the appropriate blood pressure stratum was stapled to the participant's chart. The envelope's number was recorded on the chart, which was later forwarded to the coordinating center. Once an envelope was assigned, persons were considered participants in the clinical trial for purposes of analysis, regardless of any further participation in the program. At this first clinic visit, all participants were asked to sign a release of information form so that the clinics could send examination results to their primary providers of medical care. It was explained that the participant should come back to the clinic in 1 week for a physical examination and at that time the results of the blood and urine tests, the electrocardiogram, and the chest x-ray would be explained. It was further explained that the participant then would be advised what to do in terms of getting further treatment. The randomization envelope was not opened until the end of the second clinic visit, w h e n all baseline data had been collected. If the envelope assignment was to be stepped care, the participant was invited to participate in a special stepped-care clinic in which all hypertensive care, including medicine, was provided free of charge. Informed consent was obtained from steppedcare participants at the first treatment visit, approximately 1 week later. For those stepped-care participants who did not attend the scheduled first treatment visit, efforts were made throughout the trial to gain their active participation in the stepped-care program. If the envelope assignment was to referred care, the participant was referred to his o w n physician (or an existing source of appropriate medical care if the participant had no care source). Referred-care participants were advised that they were entitled to free physical examinations, biochemical tests, electrocardiograms, and chest radiographs provided by the program at years 2 and 5 and that the study investigators were interested in keeping in contact with them throughout the program. Referred-care participants were further encouraged to participate in annual home visits (place of work in one clinic) to assess their blood pressure levels, their treatment status, and the occurrence of trial outcomes.

59S

Recruitment in the HDFP

RESULTS Screening Experience Screening efforts for the clinical trial began in January 1973. It became obvious early in the trial that the screening could not be completed within 9 months, as originally proposed. As is evident in Figure 2, the cumulative number screened during this period remained well below the number required to meet the screening schedule. Home screening actually took 16 months and was completed in May 1974. At the close of recruitment, the actual cumulative number of persons screened was less than the total number of enumerated age-eligible persons because some individuals declined to be screened. (Also, the Chicago clinic screened employees of selected industries rather than members of households and therefore did not attempt to screen household members not on the employment roster.) The numbers of persons screened and randomized are shown by month in Figure 3. Randomization rates lagged behind the initial screening rates because treatment assignment was made only after BP confirmation at the clinic screen 1-2 weeks later. The recruitment rate peaked in the summer and fall of 1973. Table 1 shows the location of the 14 HDFP clinics, their sampling frames,

18°t "~

/

150 ~..........................................................................................................................................................................................................................

;;E

8

...................

a0 //.// o!~--~'~.~-c/......

1

3

:

5

,

,

7

,

,

,

9

,

11

,

~

1

=

~

3

,

=

5

MONTH Figure 2. Cumulative number of persons screened for high blood pressure between January 1973 and May 1974 as compared with the cumulative monthly screening goals proposed before recruitment began (original goals) and after the screening period was extended (final goals).

60S

c . E . Ford, H. G. Langford, and M. J. Palmer

A. Screened Population 18000 a LU

15000

Z

LU LIJ re"

12000

O0 n" L~J rn

9000 6000

Z

3001 ........... 1

3

5

7

9

11

1

3

5

3

5

MONTH

B. Randomized Participants 1800 C3

1500

N 0 O Z

E rrL~

Z

1200 900 600 300 0

1

3

5

7

9

11

1

MONTH Figure 3. Monthly number of participants screened (A) and randomized (B) in the HDFP screening for high blood pressure from January 1973 to May 1974.

61S

Recruitment in the HDFP

Table I Location of the HDFP Clinics, Their Sampling Frames, and the N u m b e r of Persons Enumerated, Screened, and Randomized (1973-1974) Number (%)a Number Screened Number Clinic Enumerated (Ages 30-69 (%)b Location Sampling Frame (All Ages) years) Randomized Atlanta, GA Census tracts 19,648 7,183 (85.7) 1,136 (15.8) Baltimore, MD Random sample of 20,800 7,498 (90.4) 978 (13.0) 17 census tracts Birmingham, Random sample of 28,295 10,956 (94.3) 730 (6.7) AL the city Boston, MA Chicago, IL Davis, CA

Census tracts Industrial employees Random sample of a county

37,038 22,688 40,523

14,775 (92.0) 7,153 (57.9) 13,510 (94.7)

750 (5.1) 515 (7.2) 604 (4.5)

East Lansing, MI Evans County, GA Georgetown, DC

Census tracts

21,531

5,933 (90.9)

586 (9.9)

Census tracts of three counties Housing project

15,083

5,693 (95.1)

688 (12.1)

27,804

7,851 (93.0)

760 (9.7)

Jackson, MS

Enumeration districts Census tracts

25,324

7,055 (87.5)

891 (12.6)

38,537

18,465 (93.8)

880 (4.8)

Census tracts

43,410

17,517 (94.6)

845 (4.8)

Housing project Census tracts

31,007 70,638

14,522 (92.0) 21,455 (86.6)

926 (6.4) 1,097 (5.1)

442,056

159,566(89.0)

11,386 (7.1)

Los Angeles, CA Minneapolis, MN New York, NY Salt Lake City, UT Total Centers

aPercent of enumerated age-eligible persons. ~Percent of screened persons.

and the number of persons enumerated, screened, and randomized by each clinic. With the exception of Chicago, the clinics screened 85.7%-95.1% of their respective age-eligible populations, with ten centers screening over 90%. The Chicago clinic screened 73.7% of the employees of selected industries (57.9% of the enumerated age-eligible persons). Altogether, nearly 90% of those persons eligible participated in the first screen. The number of persons screened varied by clinic (5,693 in Evans County, Georgia, to 21,455 in Salt Lake City). This was due, in part, to variations in the estimated prevalence of hypertension in the target population----clinics with lower estimates of high-blood pressure prevalence targeted larger populations for screening. Of the 159,000 persons w h o participated in the initial screen, 22,994 (14.4%) had mean DBPs of 95 mm Hg or greater. These individuals were invited to undergo a second screen at HDFP clinics, and 17,499

C. E. Ford, H. G. Langford, and M. J. Palmer

62S

Attended 2nd Screen

100T C

J

•~

80

•

60 T

"'

81.8 73.6

75.0

72.1

Black

White

Black

40

o e--

~

20 0

White

Men

Women

Figure 4. Rate of participation in the second screen by race and sex. The percentage is based on the number of persons with elevated blood pressure (DBP ~ 95 mm Hg) at the first screen. (76.1%) attended. Figure 4 presents the percentage attendance at the clinic screen by race and sex group. More detailed presentations of the HDFP screening results can be found elsewhere [2,5,9,10]. Recruitment Results All participants w h o attended the second screen and were confirmed as having elevated DBP (t> 90 mm Hg) were randomized. Table 1 shows the number and percentage of screened persons that each clinic randomized. The relative proportions randomized by clinic reflect the prevalence of high blood pressure in their respective target populations. These proportions ranged from 4.5% in Davis, California, to 15.8% in Atlanta, Georgia. Table 2 gives the percent race-sex distribution of those who were ageeligible (30 to 69 years) in each clinic. Although there was no apparent pattern in screening yield with respect to sampling frame (Table 1), the influence of the markedly higher prevalence of high blood pressure in blacks was apparent. The six clinics that screened almost entirely white populations had the lowest percentages randomized. On the other hand, the four clinics that screened almost entirely black populations had the highest percentages randomized. The four clinics that screened mixed black and white populations had medium percentages randomized. Overall, the age-eligible population was 74% white (including less than 1% other races) and 26% black, while the randomized participants were 44% black. The cumulative intake rate during screening is shown in Figure 5.

63S

Recruitment in the HDFP T a b l e 2 P e r c e n t R a c e - S e x D i s t r i b u t i o n a of A g e Eligibles b y HDFP Clinic Percent White Male b

Percent White Female b

Percent Black Male

Percent Black Female

Atlanta, G A Baltimore, MD Birmingham, AL

5.3 0.2 26.2

8.0 0.5 32.7

33.1 41.0 16.1

53.4 58.0 24.9

Boston, M A Chicago, IL Davis, CA

42.8 39.0 48.4

51.6 47.0 50.4

2.4 5.8 0.6

3.1 8.3 0.5

East Lansing, MI Evans County, GA Georgetown, DC

31.1 35.4 0.6

35.0 40.4 0.8

14.8 10.6 31.6

19.1 13.6 67.0

Jackson, MS Los Angeles, CA Minneapolis, M N

0.4 46.7 44.6

0.6 52.8 51.1

39.4 0.2 2.1

59.6 0.3 2.2

N e w York, NY Salt Lake City, UT Total Centers

36.5 45.2 33.9

45.8 53.8 39.6

7.6 0.4 10.3

10.0 0.5 15.9

Clinic

=Percents do not always sum to 100 due to missing race-sex codes. blncludes less than 1% "other" races.

12.

9

..........................................................................................

1

3

5

7

9

..............................................................

11

1

3

5

MONTH Figure 5. Cumulative m o n t h l y intake (randomization) of participants in the HDFP as c o m p a r e d with the original a n d final cumulative m o n t h l y intake goals.

64S

c.E. Ford, H. G. Langford, and M. J. Palmer Although 11,386 persons were randomized, 446 participants were subsequently excluded from analysis due to randomization errors that occurred in one clinic. (Because the errors in randomization occurred after October 12, 1973, all participants from this clinic who were randomized on or before this date were included in analyses.) This left 10,940 participants for follow-up: 5,485 assigned to stepped care and 5,455 assigned to referred care. Over 71% had baseline DBPs of 90-104 m m Hg, 19% had DBPs of 105-114 m m Hg, and 10% had DBPs greater than or equal to 115 mm Hg. Approximately 15% showed evidence of major target organ damage on entry. Participants were 56% white, 54% male, 44% aged 30-49 years, 34% aged 50-59 years, and 22% aged 60 to 69 years. The two treatment groups were comparable across a wide spectrum of baseline characteristics overall and for each of the major subgroups [9]. When those randomized into stepped care were offered treatment for hypertension, a uniformly high percentage accepted; overall, 97% attended the first treatment visit.

DISCUSSION The 14 clinics of the HDFP carried out a systematic blood pressure screening of more than 159,000 persons in defined populations, with a second screen for those having elevated blood pressure. There was a consistently high level of participation at each stage of the recruitment process. A total of 84% of the targeted households were enumerated; there was an 89% acceptance of screening by the age-eligible population; more than 76% of those invited to participate in the second screen attended; nearly 97% of those randomized to the stepped-care group initially accepted treatment; and more than 80% of the stepped-care group remained active through most of the 5-year followup. Of those randomized to the referred-care group, the majority participated in the annual home visits and the two follow-up clinic visits to which they were entitled.

Screening Efforts The HDFP screening efforts were community-based, and therefore, it was important to gain the cooperation of the medical and general community. The directors of clinics or their representatives met with the local medical societies and explained the objectives of the HDFP. Most clinic directors also met with individuals or groups that they considered influential in the community and instituted publicity campaigns involving the press, radio, and television. Screening and recruitment required approximately 16 months to complete. This was nearly twice the time originally projected. The extended screening period was partly a result of underestimating interview personnel needs in the clinics but was also affected by the identification of a larger number of age-eligibles than anticipated prior to the enumeration of the target population and by the quality-control efforts of the clinics' recruitment coordinators. In a few clinics, "curbstoning," the creation of study data while seated at a curbstone, by one or two interviewers necessitated the rescreening of affected

Recruitment in the HDFP

655

households. Curbstoning was detected by clinic coordinators, who systematically telephoned a sample of the supposedly screened households to verify contact. Such quality-control checks were carried out over the course of recruitment [4]. Nearly all clinics completed recruitment at about the same time.

Screening Yield The demographic characteristics of the screened population were, to a large extent, determined by NHLBI's selection of clinical centers through the contract process. The higher prevalence of hypertension among blacks than whites resulted in a proportionately higher screening yield among the age-eligible black individuals. This more than any other factor influenced the number of people a clinic had to screen to meet its recruitment goal. As a direct result of this higher prevalence, the HDFP study population included a higher percentage of blacks than the U.S. population; on the other hand, it closely resembled the racial distribution of hypertension in the target population [2,5,10,12]. Recruitment and Randomization Although more than 75% of those found to have DBPs of 95 mm Hg or higher attended the second screen, this rate was not uniform across clinics and varied by race, sex, age, and education level [10]. With one exception, the percentage of eligible persons attending the clinic screen ranged from 71.7% for the Birmingham clinic to 91.1% for the Chicago clinic, with the clinics that screened predominantly white populations having the higher attendance rates. The exception was an attendance rate of 55.4% at a clinic that had multiple staffing and facility problems. These problems were corrected only after repeated site visit reviews, but not before a serious recruitment incident. Starting in October 1973, a staff member began opening the sealed randomization envelopes and selectively assigning to stepped care persons w h o m she knew and felt could be helped by the free hypertension treatment. Although this was not done maliciously, the randomization sequence for this clinic was compromised and the situation was not corrected until after recruitment was completed. The coordinating center staff discovered the problematic randomizations through a planned examination of the envelope sequence numbers, which were compared with the date and time sequence of randomizations within each clinic. Although problems with randomization at this clinic were suspected as early as March 1974, the randomization sequence errors were not confirmed until late 1974, due in part to a backlog of forms processing in the clinics and at the coordina "ti:ng center [4]. In February 1975, the HDFP Steering Committee directed that of the 760 participants randomized at this clinic, the data on all 446 persons randomized after October 12, 1973, were to be excluded from analyses. However, the 446 excluded participants continued to receive the study treatment to which they were randomized. The HDFP sample size goal was 10,500. After exclusion of the improper randomizations, 10,940 participants were followed in the clinical trial. The sample size had been estimated on the basis of an anticipated 5-year mortality

66S

c.E. Ford, H. G. Langford, and M. J. Palmer of 96 per 1000 in the referred-care group and a 50% dropout rate in the steppedcare group over 5 years. Actually, the referred-care group experienced a mortality of 74 per 1000. In the stepped-care group as a whole, blood pressure control and trial participation were better than had been estimated for sample size calculation, with only about 20% of the group dropping out of the trial over the 5-year period [9]. The randomized participants included a substantial number from each sex group, the black and white races, each of the four age decades between 30 and 69 years, groups with and without concomitant target organ disease, and the three major DBP strata, with ~lfild hypertension (DBP 90-104 mm Hg) accounting for 75.1% of study participants [2,9]. The cost of recruitment, per randomized participant, depended on several factors that varied across clinics. The New York clinic, in a screen of a densely populated housing project, spent approximately $235 per randomization (1973 dollars), with only 6.4% of those screened being randomized. In Atlanta, which screened households of entire census tracts, each randomization cost about $250, with 15.8% being randomized. The Jackson, Mississippi, clinic screened entire enumeration districts, with interviewers going door to door; each randomization averaged about $470, with 12.6% randomized. Finally, the Davis, California, clinic spent nearly $836 per randomized participant, screening a stratified random sample of a county where a 70-mile round trip was at times required for the enumeration and where only 4.5% of those screened were eventually randomized.

SUMMARY

Variations in the characteristics of the 14 communities screened by the HDFP permitted the assessment of "whether intense community efforts to identify and treat hypertensive persons in special program (could) improve control of hypertension for those in the general population with previously undetected, untreated, or uncontrolled hypertension" [5]. Variations also permitted the enrollment of a broad spectrum of the U.S. hypertensive population into the HDFP clinical trial [10-12]. The recruitment results suggest that a population-based screening program for a high prevalence condition such as hypertension is an effective mode of recruitment, particularly when a population-based sample is required. Strong coordinating center leadership is necessary when multiple clinics participate in such endeavors. Consideration should be given to hiring junior professional employees, responsible to the Steering Committee and the coordinating center, who would continually visit the participating centers during recruitment. Their responsibilities would be to advise on techniques, training and handling of personnel, and onsite inspection to determine the quality of the primary data. ACKNOWLEDGMENT Research upon which this publicationis based was performed under contracts NO1-HV-12433 through 42; NO1-HV-22931, 37-39, and 45; and NO1-HV-32933, NO1-HV-72915, and NO1-HV82915 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services.

Recruitment in the HDFP

67S

REFERENCES

1. The Hypertension Detection and Follow-up Program Cooperative Group: The Hypertension Detection and Follow-up Program. Prev Med 5:207-215, 1976 2. Daugherty SA, Entwisle G, Eds: Hypertension Detection and Follow-up Program: Baseline characteristics of the enumerated, screened, and hypertensive participants. Hypertension 5(Suppl 4):IV-l-IV-205, 1983 3. Ederer F: Practical problems in collaborative clinical trials. Am J Epidemio1102:111118, 1975 4. Curb JD, Ford C, Hawkins CM, Smith EO, Zimbaldi N, Carter B, Cooper CJ: A coordinating center in a clinical trial: The Hypertension Detection and Follow-up Program. Controlled Clin Trials 4:171-186, 1983 5. Hypertension Detection and Follow-up Program Cooperative Group: Blood pressure studies in 14 communities: A two-stage screen for hypertension. JAMA 237:23852391, 1977. 6. Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the Hypertension Detection and Follow-up Program: I. Reduction in mortality of persons with high blood pressure, including mild hypertension. JAMA 242:2562-2571, 1979 7. Stamler R, Gosch FC, Stamler J, Lindberg HA, Hilker RJ: A hypertension control program based on the workplace. Report on the Chicago center. J Occup Med 20:618-625, 1978 8. Curb JD, Labarthe DR, Cooper SP, Cutter GR, Hawkins CM: Training and certification of blood pressure observers. Hypertension 5:610-614, 1983 9. Davis BR, Ford CE, Remington RD, Stamler R, Hawkins CM: The Hypertension Detection and Follow-up Program design, methods, and baseline characteristics and blood pressure response of the study population. Prog Cardiovasc Dis 29(Suppl 1):1!-28, 1986 10. Hypertension Detection and Follow-up Program Cooperative Group: Patient participation in a hypertension control program. JAMA 239:1507-1514, 1978 11. The Hypertension Detection and Follow-up Program Cooperative Group: Therapeutic control of blood pressure in the Hypertension Detection and Follow-up Program. Prev Med 8:2-13, 1979 12. Wassertheil-SmoUer S, Apostolides A, Miller M, Oberman A, Thom T: Recent status of detection, treatment, and control of hypertension in the community. J Community Health 5:82-93, 1979