and referral program. Throughout its. 14-year history, the program was ..... Larochelle P, Bass MJ, Birkett NJ, De Chain-. plainJ, Myers MG. Recomnnenidationis ...
The Saskatchewan Heart
Foundation
Blood Pressure Survey Between 1975 and 1988, demographic data, weight and height measurements, and blood pressure readings were obtained for 77 890 residents of Saskatchewan (about 7.6% of the population). High readings were present in 7.8% of those surveyed, but prevalence fell over the lifetime of the survey. Subjects whose drug therapy was modified had a larger fall in blood pressure than those whose medical regimen was unchanged. R
E .
U1 .>;ll X,
Entre 1975 et 1988, 77 890 residants le la Saskatchewan, representant 7.6% de la population, firent l'objet d'une enquete permettant d'obtenir des donnees demographiques, des mensurations de taille et de poids et des lectures de tension arterielle. On a obtenu des resultats eleves chez 7.8% des participants, mais la prevalence de ces chiffres eleves s'est attenuee tout au long de l'enquete. Les suiets dont la therapie medicamenteuse avait fait l'objet d'une modification ont connu une plus forte diminution de leur tension arterielle comparativement a ceux dont le regime medicamenteux n'avait pas ete modifie. Can hIm Physkian 1991;37:623-628.
THOMAVS W WILSON, MD, VRCPC N,IERNE DUBOIS, RN, BScN VIVIAN RAMNISDEN, RN PARMINDER RAINA LEONARD K. TAN, MB, BS, PhD
suggested that high blood pressure, even a single casual reading, was predictive of early mortality. These results were confirmed in numerous epidemiologic surveys over the following two decades.2 The value of treating moderate or severe hypertension, ie, diastolic blood pressure of 105 mmHg or more, was established by 1970.3 The realization that hypertension is asymptomatic and affects a substantial portion of the population led to concerns that many hypertensive patients were unaware of their condition or inadequately treated. Indeed, a common teaching aphorism held that I10% of the population was hypertensive and that half that number was aware of having the condition. Of those, half were under treatment, and only half of those treated achieved goal blood pressure.
With these thoughts in mind, the Heart and Stroke Foundation of Saskatchewan (until July 1, 1989, the Heart and Stroke Foundation of Saskatchewan was known as the Saskatchewan Heart Foundation) launched its blood pressure survey in 1975. Its goals were to examine the distribution of blood pressure readings in Saskatchewan, estimate the prevalence of hypertension, and demonstrate the efficacy of a community-based hypertension detection and referral program. Throughout its 14-year history, the program was supported entirely through voluntary donations to the Heart and Stroke Foundation. In this paper, the main results ofthe survey are presented and compared with those of similar surveys. We believe that it is the largest undertaking of its kind in Canada and ranks third behind the Community Hypertension Evaluating Clinic (CHEC)4 and the Multiple Risk Factor Intervention Trial (MRFIT)5 studies in North America.
Dr Wilson is a Professor in the Departments of Phannacologv and Medicine, Universip of Saskatchewan, Saskatoon. Ms Dubois is a Research Associat, Clinical Phawnacologv, Royal Universip Hospital,
METHODS
REPORT PUBLISHED IN
19591
Saskatoon. Ms Ramsden is Education Coordinatot; Heart and Stroke Foundation of Saskatchewan. Mr Raina is a Research Offier, Department of Communiy Health and Epidemiolog), University of Saskatchewan. Dr Tan is a Professor, Department of Communty Health and Epidenlogv, Unwersit of Saskatchewan.
The program relied heavily on volunteers. Indeed, there was one full-time paid staff member - the provincial coordinator - and four to six part-time paid coordinators. Lay persons were used to collect initial data on subjects and to weigh and measure, while registered nurses or certified nursing assisCanadian Family Physician VOI 37: Alarch 1991
623
Table 1. SOME NORTH AMERICAN BLOOD PRESSURE SURVEYS
PREVALENCE (%) STUDY
NHES'2 NHANES'3 GARBUS14
SILVERBERG'5 CHEC4
BIRKETT ET AL" BIRKETT ET AL7 HEALTH AND WELFARE CANADA'8
NSHH8 SASKANTCHEWAN
HEALTH9 THIS STUDY
DATE(S) OF
SURVEY
LOCATION
1960-1962 USA 1971-1974 USA 1973 New Orleans, LA 1973 Edmonton, Alt. 1973-1975 USA 1981-1982 Ontario 1983 Ontario 1985 Canada 1986
Nova Scotia
1989
Saskatchewan
1975-1988 Saskatchewan
624
Canadian Family Physician VOL 37: March 1991
11%
N
Census Tract-based
6672
15.2
16
19
12
43
CensusTract-bseod Community-based
24513
17.0
--
--
--
55
30 329
33.0
13
22
15
50
Community-based
9591
12.1
27
27
19
25
Community-based 1 049 225 Census Troct-based 2735 Census Tract-based 6528 Census Tract-based 3092
21.7
45
17
11
28
11.5
72
17
6
5
11.0
70
17
6
6
18.0
43
16
5
36
Census Tract-based Census Tract-based
2108
21.0
55
13
15
17
2186
17.0
34
19
17
30
Community-based
77 890
17.7
56
13
21
13
recorded blood pressure readings. Each person taking blood pressure was certified to be accurate in preliminary testing. The project began in 1975 in Saskatoon and, in 1976, was expanded to the other larger cities in the province (Regina, Prince Albert, Moosejaw), a number of smaller cities (Weyburn, Estevan, Yorkton, Melville, Melfort, The Battlefords, Swift Current), and several smaller towns and villages. Virtually the entire populated southern half of the province and several centers in Northern Saskatchewan had access to the survey. In each geographic area, the same sequence was followed. Local physicians were informed of the program by letter and in person by Dr Wilson, the medical coordinator, or other teaching physicians. The aims, objectives, and methods ofthe survey were discussed. Public awareness of hypertension generally and the Blood Pressure Survey in particular grew through stories and interviews in the local media. Dates, times, and locations ofthe survey were publicized. Privacy was assured by using a stand-alone booth in public places (malls, church basements) or in separate examining rooms in the workplace. During registants
co,
TYPE
tration, demographic data were obtained and a unique survey number assigned. Information on previous diagnosis of high blood pressure and on previous and current treatment was recorded. Subjects were weighed and measured, then rested in the supine position for 5 minutes. Heart rate and blood pressure were recorded in the right arm using appropriately sized cuffs and a mercury sphygmomanometer. Techniques were consistent with those recommended by the Canadian Hypertension Society.6 For screening purposes, we considered the following values to be the upper limit of normal: age 1 to 39, 139/89; 40 to 59, 159/94; 60 and older, 164/99. Subjects with either systolic or diastolic values above these limits rested for a further 2 minutes. The readings were repeated, and the lower of the two values recorded. All subjects were informed oftheir blood pressure readings. Those with high readings were given written information and verbal counseling. In essence, they were informed that a single high blood pressure value did not make the diagnosis of hypertension and that a physician should determine its significance. For those without a
family physician, a list of physicians in the area was offered. On the basis of this initial reading, subjects were classified as normotensive or potentially hypertensive. The latter were subdivided into undetected (no previous high blood pressure readings), untreated (previously high readings, not on antihypertensive medication), or uncontrolled (on medication). Subjects who reported taking antihypertensive drugs and whose blood pressure reading was within normal limits were termed controlled hypertensives. All potential hypertensives were recalled at 3 months. They were sent a letter and contacted by telephone if possible. Those who did not return were sent a second follow-up letter. Returning subjects were given a definite date and appointment time for maximum convenience. They were asked whether they had seen a physician and whether the physician had initiated or changed drug treatment. Pulse rates and blood pressure were recorded as before. Subjects were discouraged from attending the survey more often than asked. Those who insisted had their blood pressure measured but not recorded. Until 1979, the survey was exclusively community-based. The Blood Pressure Survey booth was installed in a public place and the public invited to attend. After 1979, "targeted" workplace screening supplemented the public screenings. Data were recorded on an index card for each subject. Cards of potential hypertensives were stored separately from those of nornotensives, and all were kept under lock and key to ensure confidentiality. The data were stored on floppy disks using an IBM PC and the Lotus 1-2-3 software program. Data were transferred to the University of Saskatchewan VAX mainframe computer, where they were checked for completeness, edited for errors, and stored on magnetic tape. Statistical analysis was performed using the BMDP statistics package.
RESULTS We screened 84 135 persons from 1975 to 1988. Data cards were incomplete for a portion, so that data for 77 890, or about 7.6% of the provincial population, were entered for analyses. Figure I shows the percentage
of each age-sex group surveyed relative to the provincial population.7 Male and female subjects were proportionally represented, those aged 40 to 69 over-represented, and those aged 1 to 19 under-represented. Figure 2 shows the mean and standard deviation of systolic and diastolic blood pressure in subjects screened. Note that systolic blood pressure increases with age, while diastolic pressure levels out after age 50. These findings can almost be superimposed on data from other surveys.4 There was a weak but significant correlation between the body mass index (weight in kg - [height in in]2) and mean arterial pressure ([2 diastolic pressure + systolic pressure] + 3) (r = 0.26, P < 0.001 for male subjects; r = 0.25, P< 0.001 for female subjects). These correlations were weaker in the older age groups. While the correlations are highly statistically significant in a group this size, only 7% of the variance in blood pressure (0.262) is explained by differences in body mass index. About 7.8% of subjects, or 6078, had systolic or diastolic readings (or both) that were above the upper limit of normal for their age group (Figure 3). Controlled hypertensives are included and represent about 56% of the total hypertensive group (ie, potential hypertensives and controlled hypertensives). If controlled hypertension is included, our estimate of the total prevalence of hypertension increases to 17.7% of the population. The prevalence of potential hypertension increased with age, and the three categories were about equal. Undetected hypertension was more common in the youngest age group screened, while untreated potential hypertension was more common in older age groups. Figure 4 shows the prevalence of potential hypertension in the various age groups over the lifetime of the program. For each age group, prevalence tended to fall from the earlier to the later years ofthe program. These decreases were found to include all three categories (undetected, untreated, and uncontrolled data not shown). Of the 6078 potential hypertensives recalled, about 57%, or 3464, returned to the survey, most (about nine out of 10) in response to the first reminder. Males younger than 50 years of age were less likely to return (about 45%), but severity of hypertension Canadian Famiy Physician VOL 37: March 1991 625
SURVEYED:
Percentage
of each
age group
(male,female,
and
total
Fi' re 1. PERCENTAGE OF POPUIATION SURVEYED: Percentage of each age group (male, fernale, and total subjects) surveyed (N = 77 890) compared with the total population of Saskatchewan (N 1 029 950).
20-
* MALE
U ° -I
a.
1
FEMALE
B TOTAL
^
10
0 U.
0. 0-19
20-39
40-49
50-59
60-69
70+
TOTAL
AGE (YEARS) and category of hypertension were not predictors. Of those that returned to the survey, most (76%) had seen a physician, and treatment had been initiated or changed in about one third. The treatment of older subjects was more likely to have changed (Figure 5). On repeat measurement at 3 months, the average blood pressure of all returning subjects fell by 9/5 mmHg. This allowed 23% of subjects to achieve normal blood pressure readings on the second screen, regardless of treatment status or change in
Figure 2. MEAN SYSTOLIC AND DIASTOLIC BLOOD PRESSURE IN EACH AGE GROUP: Error bars are standard deviation. 170 165 160 150145
T
140I
E
135 125 120 115-
J
I
L
8110
to Ln
.-Female
9
90 130 wI 60E~~
~~~~~G
55-
50'5 10-19
626
20 39
40 49
50-59 AGE
Canadian Family Physician VOI. 37: March 1991
60-69
70.
medication. Those whose drug treatment had been changed by the physician showed a greater fall in blood pressure than those whose treatment was unchanged, 17/8 mmHg compared with 7/4 mmHg (P < 0.001).
DISCUSSION These data show that the blood pressure distribution and the prevalence of high blood pressure in Saskatchewan is similar to that in other modern western societies.4 It must be admitted that a study of this type has certain limitations. First, as a community-based survey, readings were obtained only on volunteer subjects so that some preselection bias could have occurred. Census tract surveys (where a defined representative sample is recorded) may reduce such bias. However, in two such surveys conducted recently,8,'9 up to one third of those chosen refused to participate, so that selection bias remained. Second, we performed a single reading in normotensives and only two in potential hypertensives. In our pilot study, we recalled around 300 normotensives at 3 months and found that 86% of readings were within 10 mmHg.'0 Moreover, even a greater number of readings at a single sitting may not improve prediction of future read-
ings. Fully half of placebo-treated subjects in the Australian National Blood Pressure Trial reached goal blood pressure in 6 to 12 months." Third, about 10% of the data cards could not be analyzed, and the tedious task of data reduction could have led to errors. In spite of these limitations, the data appear consistent with several other North American community-based and census tract surveys (Table 14,8,9,12-18). In these, the average prevalence of hypertension (ie, high blood pressure readings or on treatment) was 17.5% ± 6.2 (SD). Areas of the southern United States seem to show higher prevalence. In general, the proportion of controlled hypertension seems to have increased and that of undetected hypertension to have decreased during the past three decades (Figure 4). Over the lifetime of our program, there was a 40% reduction in potential hypertension in the over 70 age group and smaller, but significant, drops in all other age groups. Thus, the proportion of subjects whose blood pressure is controlled has risen from between 15% and 30% during the 1960s to between 50% and 70% in the late 1980s. Presumably, this is in part due to the intense public and professional education efforts of government, industry, and voluntary agencies. Similar programs in the United States have been credited with improving control of hypertension there.'9 Another criticism of studies such as ours is the lack of linkage to clinical care for subjects with high readings. It is argued that such subjects are exposed to the deleterious effects of being labelled hypertensive without receiving any possible benefit from treatment.20 About one quarter of our subjects who returned to the survey did not seek medical advice. In our pilot study, we found that a similar fraction who did not return had not sought medical advice. Whether these figures would be improved if blood pressure screening were carried out elsewhere (in a physician's office, for example) is uncertain. Of those that returned to the survey and had seen a physician, about one third had started or had a change in drug therapy. Interestingly, the elderly and middle-aged were more likely to have a medication change (Figure 5). These figures might
-
Figure 3. PREVALENCE OF HTPERTENSION IN EACH AGE G ROUP4 See M.ethodfor'definitio. 60.
55 ~~~undetectd untreated ~~~~uncontrolled controlled
so1,
6, E
40
230
1-19
20-39
50-59
40-49 AGE
60-69
70-
Figure 4. DECREASING PREVALENCE OF POTENTIAL HYPERTENSION: Prevalence ofpotential hypertension declined in each age group during three consecutive peraods.
z 30 Ln z
25-
AGE
Li-I.-
o cr ww
c)>
20 -
70+
WI _ _lo 10
if
60-69
~~~~~~~~20-39
1--
QL
50-59 1-1i9 40-49 1984-1987
__ __
_
__
_
1976-1979
_
__
_
_
__
_
_
__
1980-1983 TIME (years)
_
_
Figure 5. PROBABILITY OF A CHANGE IN MEDICATION: Probability that persons screened high received a change or initiation of drug therapy by thefirstfollow-up reading at 3 months. z
0.45-
0
0.40-
I,.
--
0.35 a,
"I
0.30-
z z
0.250.20-
LL 0
m
0.15 @ /
0.12-
*
*
fe~~~~~~~~mcles
0.10
m 0
0.05-
a.
0
1
1-19
20-39
40
-
49
50- 59
60-69
70+
AGE
Canadian
Family
Physician
VOl
37:
Alarch 1991
627
reflect appropriate caution on the part of many practisiilg physicians who obtain blood pressure readings oni several occasionis over 2 to 6 moinths before initiating or changing drug therapy2' and the increasiing recognition that the elderly derive equal or greater beniefit froIn antihypertenisive therapy.6 Nevertheless, subjects whose diug therapy had been changed showed substantially greater blood pressure change than those who had not, attestiing to the eflicacy of modem antihypertenisive drug therapy. In suminary, we have reported findinigs of oine of the largest blood pressure suiveys on record. Our data should serve as a useful resource for documentinig progress in coIntrolling hypertension in the future. U * - - 0 0 0 - 0 - - 0 0 - 0 0 0 0 0 0 - 0 0 0 - - 0 0 0 - - 0 - - 0 - - 0 - 0
Acknowledgment We are indebted to the following persons, without whom the survey would have been impossible: the provincial coordinators, Ms CM. Dilts (1975 to 1977), the late Ms W Shaw (1977 to 1979), Ms S Tegart (1979 to 1980), Ms_J. McCulloch (1980 to 1981), and Ms M. Schmaltz (1981 to 1983); members oJ the Hypertension Subcommittee, Ms M. Clhappell, Dr R.j7. Heimtuan, Dr L. Horlick, Dr D. A'olbinson, Ms R. AMalik, Profjssor M. MacDonald, Dr B. Reedei; and Ms P Zipchen; the inany volunteers of the Heart and Stroke Foundation oJ Saskatchewan and its Executive Directors, Mr D. (uInningham (1975 to 1985) and Ms D. [Vateter (1985 to presentt). Margaret Matheson and Elsie Habbick provided expert secretarial assistance.
Requests for reprints to: Dr TW Wsion, Departnmentt of Pharmiacology, Universiy ofjSaskatchewan, Saskatoon, SK S7N QV)O
References 1. Anioniymiious. Build atid blood pressure study. Chicago, Ill: Society of' Actuaries, 1959. 2. Pooling Pr-oject Reseaclh GIroup. Relationiship of blood pressure, steirum clholester-ol, simioking habit, relative weight anid ECG abitoninualities to ixicideice of' imajor coronl)ir! events: finial report of' the poolinig project. J Chionuti Mis 1 978;3 1:201-306. 3. Veterans Admiinlistrationi Cooperative Study
Group oln Anitihypertenisive Ageints. Eflects of' treatmiecnit oni imiorbidity in hypertension. I1. Results in patienits wzith diastolic blood pressure iveraginig 90 through 114 iuiiiiHg. _7,ILIA 1970;213:l 143-52. 4. StanilerJ, Stamiler R, Reidlinger WVF, Algerla G, Roberts RH. Hypertenisioni scireeninig of' 1 million Americans. Communiity Hyperteinsioni Evaluatinig Clinic (CHEC) program, 1973 through 1975. AMIAA 1976;235:2299-36.
5. Rutanl GH, Kuller LH.H NeatotiJI), X\enothl\Ol D)H, McD)onald RH, Smlith \WMS. MNortality arso-
628
Canadian Family Physician VOI, 37: AIarch 1991
ciated with diastolic hypertenlsioni aiid isolated systolic hypertenisioni amiionig meni screcened for the Multiple Risk Factor Intenrention l'rial. Circulation 1988;77:504- 14. 6. Larochelle P, Bass MJ, Birkett NJ, De ChainplainJ, Myers MG. Recomnnenidationis from the coInsenisus coiif'erenlce oni hypertenisionl in the elderly. Can MIed AoocJ 1986;1 35:741-5. 7. Anonymous. Report uj the Medical Caie Inoutante (onrnmission. Regicca, Sask. Saskatchewaci Health Insurance Coiniissioni, 1988. 8. Nova Scotia Departmiienit of Health. Report oJfthe Nova Scotia Heart Health Survey (1J986). Halifax, NS: Departmenit of Natiocial Health anid Welfare, 1987. 9. Saskatchew,,,ani Health, City of Reginia Health Departmenit, Saskatooni Coiiimuniity Health Uiiit, Heart and Stroke Foundationi of Saskatchewan, University of Saskatchewan. Report of the Saskatchewan Heart Health Sunrey. Regina, Sask: Saskatchewvan Health, 1990. 10. Wilsoni 1WV, Dilts CMI. Eflicacy of a blood pressure survey. Ann R (oll Physicians Surg Can 1978; 1:324. 11. Ilihe Maniagemincit Comnmittee. 'Ihe Australian therapeutic trial in milld hypertensioni. Lancet 1980;i: 1261-7. 12. Nationial Ceniter lor Health Statistics. Bloodpressure by age and sex. Untited States. aVashinigtoni, DC: Nationial Ceiiter for Halith Statistics, 1964; publicatioin no. (PHS)64-1000. (Vital aiid health statistics; series I 1:11o 4). 13. Nationial Ceniter ftOI Health Statistics. Blood pressure levels of persons 6-74yeasn oJ age in the United States. Rockvillc, Md: Natioical Ceinter foI Health Statistics, 1976; DHEW publicatioin ino. HRA 78-1648. (Vital aiid health statistics; sIie'S I 1,11 0 203). 14. GCarbus SB, Garbus SB. Anialysis of nIiass hypertemision screcniimig. Prmey Aled 1976;5:48-59. 15. Silverberg DS, SIIlith EJO,Juchli B, Vani Dor sser E. Use of' shoppilig Centels ill screent'ling for hypcrtecisioni. (Can Ued Asso.wJ 1974;1 11:769-74. 16. Birkett NJ, DoL)micir AlP Maynmcard NM. Prevaleince acid (olitrol of hypertecision ill an Oiitac-io Coult\'. Can Aled AsocJ 1985; 1 32:1019-24. 17. Birkett NJ, Evacis CE, Hlynces RB, et al. Hypertenlsioci coiitrol in twvo Caciadiaci coIIIIciucities. J Hypertens 1986:4:369-74. 18. Anioccyiccous. MainlJindings epuorl oJ the Canadian bloodpressure.survey. Ottawa, Oiit: Health acid 1Velf:are C'anada, 1989. 19. DaminenbergAL, Dmiizd IT, HoIami MIJ, Haynes SC, Leaverton PE. Progiess ill the bsattle agaimist hypertension. Cliaclige ill blood pressure lcvels in the Ulcited States Flomc 1960 to 1t980. H,petlension 1987; 10:226-33. 20. Aniociyiiious. Hypertensioni. 'lorocito, Ouit: Omitario CouIncil of Health, 1977. 21. Logan AG. Report of the Caciadiani Hyperten-
siomi Soc'iety's ('omiseclisu cociferet'ltcic Omi the' m1anavgemicecilt of mIcild by perl tellisoci. Canlt .\led Aoo ,.s(7 1984;131:1033-7.
