and in the community - NCBI

Use of acetylsalicylic acid by physicians and in the community Jeffrey Mahon,*t MD, FRCPC; Kathleen Steel,t MSc; Brian G. Feagan,*t MD, FRCPC; Andreas Laupacis,*t MSc, MD, FRCPC; Linda L. Pederson,* PhD Objective: To determine physicians' attitudes toward prescribing acetylsalicylic acid (ASA), physicians' own use of ASA and the prevalence of ASA use in the community following the trials of ASA for primary prevention of coronary heart disease. Design: Random sample surveys of physicians and the general public by mail and telephone respectively and a mail survey of a selected panel of expert cardiologists and neurologists. Setting: London, Ont., and surrounding Middlesex County. Participants: A total of 210 physicians (77% of eligible subjects), including family practitioners and most types of specialists, with an active medical licence and 666 English-speaking people (75% of eligible subjects) aged 18 years or more living in a household with active, listed telephone service. Main outcome measure: Long-term ASA use (at least 80 mg on alternate days for 4 or more consecutive weeks) for the treatment of atherosclerosis. Main results: Sampled physicians and experts agreed that long-term ASA therapy was indicated in patients with unstable angina, a transient ischemic episode or recent myocardial infarction but not for primary prevention in healthy middle-aged men and women at low risk for ischemic vascular disease. Both groups were uncertain about the role of ASA in primary prevention in asymptomatic people with risk factors for atherosclerosis. Nine (16%) of the 55 male physicians aged 50 years or more took ASA routinely for primary prevention. In the community survey almost all those who used ASA routinely were 50 years or older. The proportions of men and women in this age group who used ASA routinely for any reason were 19% (95% confidence limits [CLs] 11 and 28) and 14% (95% CLs 8 and 19) respectively; the proportions of men and women who used ASA routinely and apparently for primary prevention were 8% and 1% respectively. A total of 43% (95% CLs 30 and 57) of those with apparent ischemic vascular disease took ASA routinely. Medically unsupervised long-term ASA use for primary or secondary prevention of ischemic vascular disease was uncommon (reported by 2% of those who used the drug routinely). Conclusions: Physicians generally agree on a role for long-term ASA therapy in the secondary prevention of ischemic vascular disease. However, the prevalence of long-term ASA use in people with overt atherosclerosis in the community may be less than optimal. The role of the drug in the primary prevention of ischemic vascular disease is less accepted. Long-term ASA use in the community for primary prevention is uncommon but detectable.

Objectif: Determiner l'attitude des medecins envers la prescription de l'acide acetylsalicylique (AAS), leur propre utilisation de I'AAS et la prevalence de l'utilisation de I'AAS dans la collectivite a la suite des etudes sur I'AAS pour la prevention primaire de la coronaropathie. Conception: Sondages aleatoires par courrier chez les medecins et par telephone dans le public en general et enquete postale aupres d'un groupe. restreint de cardiologues et de

neurologues. Contexte: London (Ont.) et environs du comte de Middlesex. Participants: Un total de 210 medecins (77 % des sujets admissibles), y compris des medecins de famille et la plupart des specialites, detenant un permis d'exercice valide et From the departments of *Medicine and tEpidemiology and Biostatistics, University of Western Ontario, London, Ont.

Reprint requests to: Dr. Jeffrey Mahon, University Hospital, POBox 5339, London, ON N6A SA5 -

For prescribing information see page 1163

CAN MED ASSOC J 199 1; 145 (9)

1107

666 personnes d'expression anglaise (75 % des sujets admissibles) ages de 18 ans ou plus et dont le domicile est pourvu d'une ligne telephonique active et repertoriee. Mesure des resultats: Utilisation prolongee de l'AAS (au moins 80 mg, aux 2 jours, pendant 4 semaines consecutives ou plus) pour le traitement de l'atherosclerose. Principaux resultats: Les medecins et les specialistes faisant partic de 1'echantillon etaient d'accord pour dire que le traitement prolonge a l'AAS etait indique chez les malades atteints d'angine instable, d'un accident ischemique transitoire ou d'un infarctus du myocarde recent, mais non pas pour la prevention primaire chez des hommes et des femmes d'age moyen, en bonne sante et a faible risque de cardiopathie ischemique. Les deux groupes etaient indecis au sujet du r6le de 1'AAS dans la prevention primaire chez les personnes asymptomatiques qui presentent des facteurs de risque d'atherosclerose. Neuf (16 %) des 55 medecins de sexe masculin ages de 50 ans ou plus prenaient couramment de 1'AAS a titre de prevention primaire. Dans l'enquete communautaire, presque tous ceux qui utilisaient systematiquement de 1'AAS etaient ages de 50 ans ou plus. Les proportions d'hommes et de femmes dans ce groupe d'age qui utilisaient regulierement de l'AAS pour une raison quelconque etaient de 19 % (limites de confiance [LC] a 95 %: 11 et 28) et de 14 % (LC a 95 %: 8 et 19) respectivement; les proportions qui utilisaient systematiquement de I'AAS et apparemment pour la prevention primaire etaient de 8 % et de 1 % respectivement. Un total de 43 % (LC a 95 %: 30 et 57) de ceux qui souffraient manifestement de cardiopathie ischemique prenaient systematiquement de l'AAS. L'utilisation prolongee de I'AAS sans surveillance medicale pour la prevention primaire et secondaire de la cardiopathie ischemique etait rare (signalee par 2 % de ceux qui utilisaient regulierement le

medicament). Conclusions: En general, les medecins sont d'accord sur le r6le du traitement prolonge a l'AAS dans la prevention secondaire de la cardiopathie ischemique. La prevalence de l'utilisation prolongee de l'AAS chez les personnes atteintes d'atherosclerose patent dans la collectivite peut etre indesirable. Le r6le de ce medicament dans la prevention primaire de la cardiopathie ischemique est moins bien accepte. L'utilisation prolong6e de l'AAS dans la collectivite pour la prevention primaire est peu frequente mais verifiable. umerous studies indicate that acetylsalicylic acid (ASA) is useful for the primary and secondary treatment of ischemic cerebrovascular and cardiovascular disease (CVD). 1- These reports have provided a basis for published conclusions about long-term ASA therapy for CVD5-9 and may be influencing prescription of the drug by physicians. As well, the combination of ASA's effectiveness, the response of the lay press to current studies (Toronto Star, Jan. 27, 1988: A12; New York Times, Jan. 27, 1988: 1), pharmaceutical advertising'0 and unrestricted access to the drug makes it possible that medically unsupervised long-term ASA use is common. To our knowledge these issues have not been addressed since the results of recent large trials of ASA in asymptomatic subjects were pubN

lished.4"1

veyed to ascertain their recommendations to patients about the use of ASA for CVD, the relation between these recommendations and those of panels of experts, and the proportion of physicians who use ASA, including those who do so for the treatment and prevention of CVD. We carried out a population-based survey to determine the proportion of the general public who use ASA routinely for any reason, the proportion who do so for the treatment and prevention of CVD and the source of advice among those who do so.

Methods Middlesex County is located in southwestern Ontario, about 200 km from Toronto and Detroit. The region has a population of 332 000, of whom over 80% live in London.'2 London provides primary and tertiary health care for the county and is the site of a medical school. The data for both surveys were collected between October and December 1989.

This information may be important in at least two ways. First, the degree to which physicians are putting into practice the results of current trials of ASA for CVD may have implications for medical education in atherosclerosis. Second, the risks and benefits associated with ASA use may have an Physician survey and expert panel important effect on the health of the community if For the physician survey the target population use is widespread. of all physicians in active practice in of consisted Therefore, we conducted sample surveys Middlesex County. The sampling frame was the surand physicians and residents of London, Ont., Medical Directory, ' 3 which contains Ontario sur1988 were rounding Middlesex County. Physicians 1108

CAN MED ASSOC J 1991; 145 (9)

LE I er NOVEMBRE I199 1

..M

the names, addresses and areas of specialization of physicians who maintain a licence. We selected communities within the county that contained at least one licensed physician; from these a total of 1195 physicians were identified. The physicians were classified into one of three groups according to type of practice: general or family practitioners (GPs) (581 physicians), internists, cardiologists, neurosurgeons or neurologists (ICNNs) (173) and other (441). A systematic sample of 314 physicians, with an equal sampling fraction from each group, was selected. This sample size was adequate to provide 95% confidence limits (CLs) within 4% of the true proportion of physicians who use ASA routinely for the primary and secondary treatment of CVD, assuming that 10% do so and assuming a response rate of 70%. Long-term ASA use was defined before data collection as the use of at least 80 mg of ASA every other day for 4 weeks or more. This minimum dose was chosen on the basis of in-vitro evidence that it may be adequate to alter platelet aggregation'4 and protect against atherosclerosis. The period of 4 weeks or more was specified to eliminate most self-limiting conditions for which ASA is taken. If a respondent reported using an ASA-containing product at least every other day for 4 weeks or more but could not give the usual dose taken, the dose was assumed to be at least 80 mg. A survey questionnaire consisting of three sections was developed and pretested with 10 physicians. The first part concerned respondents' own use of ASA, including preparation, indication, dose, frequency and duration of use. The presence of CVD, as assessed through a history of a transient ischemic episode, nonhemorrhagic stroke or coronary heart disease, was also determined. The second section described seven patients for whom long-term ASA therapy could be considered: (1) a 50-year-old woman with unstable angina, (2) a 55-year-old man 4 weeks after an uncomplicated myocardial infarction, (3) a 62-year-old man with a transient ischemic episode in a carotid artery distribution and essential hypertension controlled with diuretic therapy, (4) a 45-year-old asymptomatic man with three vascular risk factors (non-insulin-dependent diabetes mellitus, obesity and smoking), (5) a 57-year-old asymptomatic woman with several risk factors (noninsulin-dependent diabetes, a family history of coronary heart disease, smoking and obesity), (6) a healthy 55-year-old man and (7) a healthy 65-year-old woman. Respondents were asked whether they would prescribe ASA and, if so, in what form and for how long. The third section collected demographic information. The survey package containing the questionnaire, a covering letter and a return envelope was NOVEMBER 1, 1991

mailed to the physicians. Follow-up consisted of two mailed and one telephone reminder. The physician was classified as a nonrespondent if these measures failed to elicit a response. Two panels of Canadian experts were established to determine the appropriateness of ASA therapy for the seven hypothetical patients. The first group, comprising four cardiologists and one expert in thromboembolism, had expertise in CVD, and the second comprised three neurologists with an interest in cerebrovascular disease. The descriptions of the patients were circulated to the panellists, and a modified Delphi technique was used in an attempt to reach a consensus.'5

Community survey The target population consisted of all Englishspeaking residents of Middlesex County who were at

least 18 years old and who were not living in a health care facility at the time of the survey. The population of those aged more than 19 years is 240 500, of whom 82% live in London.'2 The survey was conducted by telephone, with calls being preceded by a mailed letter of introduction. Therefore, the sampled population was limited to people living in households with active telephone service (estimated to include over 97% of London households'6) and those whose number was listed with an address in the directory (excluding a further 7.7% of households with unlisted numbers [Corporate Economics Department, Bell Canada: personal communication,

1990]).

The sampling frame was the telephone directory for London and region. Blocks of commercial listings were removed, as were listings for acute and chronic health care facilities. Directory sections containing at least one listing within the county were included. Sampling occurred in two stages. First, after randomly selecting the initial listing, we systematically chose 993 listings, according to a predetermined sampling fraction. Second, once an eligible household was reached by telephone, the adult member with the most recent birthday was selected to be interviewed.'7 The sample size was adequate to provide 95% CLs within 2% of the true proportion of the general public who use ASA routinely for the primary and secondary treatment of CVD (assuming that 6% do so and assuming a response rate of 70%, as determined on the basis of the pilot study results mentioned later). The representativeness of the sample with respect to age, sex, education and location of the home (urban v. rural) was assessed by comparison with 1986 Canada census data for Middlesex County.'2 Chronic ASA use was defined as in the physician survey. About 7 days before telephoning we mailed a list CAN MED ASSOC J 1991; 145 (9)

1109

-1

of 45 of the most frequently used ASA-containing preparations available in Canada to sampled households. The list was derived from one of almost 150 ASA-containing productsl8 and was provided to reduce the possibility that respondents would not recognize that they were using an ASA preparation. This approach precluded the use of random-digit dialling to collect the sample. The telephone questionnaire was pilot-tested in 95 randomly selected London households before the survey. The results provided information for sample size estimation but are not included in the current report because of subsequent revisions to the questionnaire (copies of the questionnaires used in the physician and community surveys are available from the authors on request). The interviews for the survey were conducted by six trained interviewers. Most respondents were contacted in the early evening. Eight attempts to reach eligible subjects were made at various times of the day before a listing was termed "no contact." The interviewer asked questions about the use of ASA-containing products (reasons for using them, preparation, dose, frequency and duration), the source of advice to use ASA and the knowledge of side effects. CVD was deemed present if the respondent gave a history of "heart attack," stroke, transient ischemic episode, amputation of an extremity because of impaired circulation or use of medication for angina. Demographic information was also collected.

Statistical analysis The characteristics of the samples were compared with those of reference populations by means of the x2 test. Proportions in the community survey were determined under the assumption that the sample was randomly selected. CLs for proportions

were based on the normal approximation to the binomial distribution'9 unless the event of interest was rare, in which case exact CLs were determined.20 We used the x2 test or Fisher's exact test (two-tailed) to compare proportions. In both surveys ASA use was examined in people less than 50 years old and in those aged 50 years or more. This age was chosen before data collection because benefit from taking ASA for the primary prevention of myocardial infarction may be limited to those who are least 50 years old.4

Results Physician survey and expert panel Of the 314 physicians sampled 42 were excluded because they were not living in Middlesex County (in 20 cases), were not engaged in work directly related to medicine (retired in 12 cases, disabled in 3 and other in 4) or had died (in 3). Of the remaining 272 eligible respondents 210 (77%) completed the questionnaire, 52 (19%) did not respond and 10 (4%) refused to participate. The demographic characteristics of the respondents are given in Table 1. The age distribution of the entire sample did not differ significantly from that of physicians in the province (data supplied by the College of Physicians and Surgeons of Ontario). Comprehensive data on sex distribution among Ontario physicians were not available. The response rates were more than 95% for ICNNs and about 70% for each of the two other groups. A number of physicians originally classified as GPs were reclassified as "other" because they indicated a specialty interest on the returned questionnaire. The prevalence of routine ASA use for any reason by the physicians is shown in Table 2. Routine use varied by age (less than 50 years v. 50

Table 1 Demographic characteristics of 210 physicians in Middlesex County. Ont.

Group Internists, cardiologists, Characteristic Mean age (and standard deviation [SD]), yr iSex, % Male

Female

Location, 0/ London Outside London 1110


General practitioners (n - 65)

neurosurgeons

and

neurologists (n

=

49)

Other (n= 96)

43.7 (15.2)

44.6 (11.1)

46.5 (12.6)

71 29

96 4

88 12

82 18

96

85

4 LE Ier NOVEMBRE 1991

1

more) (X2 = 26.3, p < 0.0001) but not practice type or sex (although the number of female physicians aged 50 years or more was verv low). Eighteen physicians used ASA routinely and expressly for the prevention of myocardial infarction or stroke, of whom 17 specified the daily- dose: 11

years or

Table 2: Prevalence of long-term use of acetylsalicylic acid (ASA) for any reason among physicians, by sex, age and cardiovascular disease (CVD) status

Age, yr; CVD status* < 50

. 50

Sex; % of physicians (and 95% confidence limits [CLs]) Male Female (n= 177) (n = 33) 3(1, 8) 4(0,18) (n = 110) (n = 28) 30 (19, 42) 0 (0,60) (n 4) (n = 64) 24 0 (n = 4) (n =55) 67 0 (n = 0) (n = 9) 13 3 =

NoCVD

CVD present All

*No respondent under 50 years of age had overt CVD. The data for four physicians (three men and one woman) whose age was unknown, one of whom used ASA routinely, were excluded.

reported taking 325 mg, 3 reported taking less than 325 mg and 3 reported taking more than 325 mg. Among the male physicians aged 50 years or more 6 (67%) of the 9 who had CVD took ASA routinely, and 9 (16%) of the 55 without CVD did so expressly

CVD. The recommendations on long-term ASA therapy for the seven hypothetical patients by the GPs, the ICNNs and the expert panel are given in Table 3. We did not tabulate the responses by the physicians in the "Other" group because they were unlikely to manage such problems. For the patient with a transient ischemic episode (patient 3) we report only the recommendations of the expert neurology panel. The five members of the expert cardiology panel agreed that long-term ASA therapy was indicated in the patients with unstable angina, recent myocardial infarction or a transient ischemic episode but not in the healthy middle-aged man or woman. Two and three panellists respectively would treat the asymptomatic man and woman at high risk because they were convinced that ASA prevents clinically important cardiovascular events and that these patients were at sufficient risk of such an event to justify the risks of ASA therapy. The panellists who would not treat these patients felt that the existing evidence on to prevent

Table 3: Recommendations of the sampled physicians and an expert panel on long-term ASA therapy for seven hypothetical patients*

Group General practitioners % who would recommend long-term ASA therapy No. who specified a daily dose Daily dose specified, % of physicians < 325mg 325mg > 325 mg Internists, cardiologists, neurosurgeons and neurologists % who would recommend long-term ASA therapy No. who specified a daily dose Daily dose specified, % of physicians < 325 mg 325mg > 325 mg Expert panelt

Patient no. 3 4

1

2

67

83

94

39

49

15 74 10

5

6

7

37

37

22

15

56

21

24

12

9

14 80 6

7 70 23

14 76 10

17 71 13

8 83 8

44 44 11

70

77

91

17

13

2

2

30

32

41

6

5

1

1

10 80 10 Yes

9 84 6 Yes

10 60 29 Yes

0 100 0 NC

0 100 0 NC

0 100 0 No

0 100 0 No

*Patient 1 = woman with unstable angina, patient 2 = man with recent myocardial infarction, patient 3 = man with a transient ischemic episode, patient 4 = man at high risk, patient 5 = woman at high risk, patient 6 = man at low risk, patient 7 = woman at low risk. See the Methods section for complete details. tThe recommendations are those of the expert cardiology panel except for patient 3 (expert cardiology and neurology panels). Yes ASA therapy was recommended by all the panellists, NC = no consensus was reached, No = no panellist recommended ASA therapy. =

NOVEMBER 1, 1991

CAN MED ASSOCJ 1991; 145(9)

1111

the efficacy of ASA could not be extrapolated to these cases, and they were concerned that the risks of therapy might outweigh the benefits. When the justifications for the responses to these two cases were circulated to the panellists none changed their opinion. The three neurologists agreed that longterm ASA therapy was indicated for the patient with a transient ischemic episode. Most (two-thirds or more) of the GPs and ICNNs advised long-term ASA therapy in the cases involving secondary prevention of CVD, whereas a minority (2% to 37%) recommended such therapy for primary prevention (Table 3). The GPs recommended ASA therapy for primary prevention more frequently than the ICNNs. The dose suggested by all the members of the expert cardiology panel was 325 mg/d except for the patient with a transient ischemic episode, for whom two advised 650 mg twice a day; after the justifications for the responses were circulated one of the two changed his response to 325 mg/d. Similarly, two of the neurologists recommended 650 mg twice a day and one 325 mg/d for the patient with a transient ischemic episode. Among the GPs and ICNNs the dose most frequently suggested was 325 mg/d, with a tendency to suggest higher doses for the patient with a transient ischemic episode.

Community survey Of the 993 listings 107 were excluded because they were not in service (in 74 cases) or were not household listings (in 10) or because the respondent was ineligible owing to age, language or place of residence (in 23). For the remaining 886 listings complete or near-complete interviews were obtained in 666 cases (75%), 172 people (19%) refused to participate and 48 households (5%) could not be reached. The features of the respondents are given in Table 4. The sample contained a greater than expected number of women, particularly those less than 24 years old or more than 75. Rural residents were overrepresented among females, but place of residence was not found to be related to ASA use. Because of the age groups used in Statistics Canada tabulations the level of education of the sample could be compared with Middlesex County data only for people aged 25 or more (this would exclude 230 respondents). Men of this age were more likely than the general population to have secondary school education or less (X2 = 7.2, 2 degrees of freedom [df], p < 0.05); there was no significant difference among the women. Since less education was found to be associated with long-term ASA use (X2 = 10.52, 2 df, p = 0.03) estimates of the proportions of men who used ASA routinely were adjusted for dif1112


ferences in education status. This adjustment had a negligible effect on the estimates (0.3% or less), and therefore the unadjusted estimates and CLs are presented. A total of 50 (8%) of the 666 respondents took ASA routinely. Prevention of heart attack or stroke was the most common reason for doing so (in 38% of cases). Other reasons for long-term ASA use were arthritis (in 32%), headaches (in 16%) and other musculoskeletal problems such as chronic back pain (in 6%). Four respondents took ASA routinely for the purpose of "blood thinning"; however, the context for this (e.g., deep venous thrombosis v. coronary heart disease) was not determined, and these respondents were not included with those who used ASA expressly for preventing ischemic CVD. Some respondents gave more than one reason for taking ASA routinely. Long-term ASA use for any reason was significantly associated with age of 50 years or more (X2 = 34.9, p < 0.00001) and presence of CVD (X2 = 97.5, p < 0.00001) but not sex (Tables 5 and 6). Of the 53 respondents with apparent CVD 43% (95% CLs 30 and 57) used ASA routinely. Of the 19 respondents who used ASA routinely for the prevention of CVD 8 were deemed not to have CVD and were assumed to be using ASA for primary prevention. Seven of the eight were men, of whom six were 50 years old or more. A total of 8% of the men in this age group used ASA for primary prevention, as compared with 1% of the women (p 0.006, Fisher's exact test). All but one person using ASA routinely for apparent primary or secondary treatment of CVD did so on the advice of a

physician. Sixteen of the 19 respondents who used ASA routinely to prevent CVD were able to provide the usual dose taken: none took the drug on alternate days, 9 took 325 mg/d, and 7 took 650 mg/d or more. Of the last seven subjects four were taking ASA to prevent a heart attack and three to prevent a stroke. A total of 43% of the respondents were unable to name any side effects of ASA, and 3% named side effects that are not associated with the drug (e.g., tumours). Among the better-known effects were upset stomach (named by 33% of the respondents), bleeding (17%) and ulcers (9%). People who used ASA routinely were no more knowledgeable about side effects than the other respondents: 42% could name no side effects, 34% named upset stomach, 26% bleeding, 10% ulcers and 8% tinnitus.

Discussion Our main findings are that (a) there is general LE ler NOVEMBRE 1991

acceptance by physicians in London and the immediate area of long-term ASA therapy for secondary prevention of CVD but little acceptance of a role for ASA therapy in primary prevention, (b) 16% of male physicians aged 50 years or more take ASA routinely for primary prevention of CVD, (c) about half the people with CVD in the community do not take ASA routinely, (d) 19% of men and 14% of women 50

years or older in the community take ASA routinely, and 8% and 1% respectively do so apparently for primary prevention of CVD, and (e) medically unsupervised long-term ASA use for primary or secondary prevention of CVD is uncommon. Several factors should be considered in interpreting the results of the community survey. First, households with unlisted or no telephone service

Table 4: Ch.araceristics of a community sample of 666 subjects compared with those of the population of Middles.ex County12

.Characteristic Sex, no. (and %)* Sample :County. Mean age (and SD) of sample, yr Age, no. (and %). 20-34 yr Sample ... County 35-44 yr... ... Sample........

Male 260 (40.4) 113 665 (47.3)

384 (59.6) 126 850 (52.7) X2 =11.9, 1 df,t p < 0.001

..

County...

.45-54 yr ........ .Sa.mple....... .CnY e. ..... County 55-64 yr Sam.ple.. .County..

.. 65

43.1 (16.7)

47.9 (19.0)

96 (36.9) 44 535 (39.2)

115 (29.9) 47 335 (37.3)

53 (20.4) 22 745 (20.0)

81 (21.1) 23 905 (18.8)

32 (12.3) 16 295 (14.3)

44 (11.5) 16 890 (13.3)

44 (16.9) 14845(13.1)

50 (13.0) 16475(13.0)

yr..

Sa.mple.. County Place of residence, no. (and %)* :London.

Sample

C:ounty:

Highest level of education,. no. (and %)t Secondary schoo or less Some postsecondary and other

Community college University degree Employment in sample,

Clerical or service *Manual Unemployed Other

x

35 (13.5) 15 245 (13.4) 4.06, 4 df, p = 0.39

205 (78.8) 93 970 (82.7)

Co:unt.y.. Outside London Sample

no. (and )t Retired Student Homemaker Employed Professional

Female

.

55 (21.2) 19 695(17.3) x2-2.4, 1 df, p= 0.12

94 (24.5) 22 245 (17.5) x2=18.1, 4 df, p =-0.001 291 (75.8) 107 060 (84.4)

93 (24.2) 19 790 (15.6) x2 20.9, 1 df, p < 0.001

140 (52.2)

218 (55.5)

56 (20.9) 22 (8.2) 50 (18.7)

77 (19.6) 45 (11 .4)

53 (13.5)

53(19.8) 1 (0o4)

85 (21.8) 22 (5.6) 71 (18.2)

44 (16.4) 63 (23.5) 72 (26.9) 2 (0.7) 8 (3.0)

58 (14.9) 123 (31.5) 16 (4.1) 9 (2.3) 11 (2.8)

. 28 (10.4)

*Eleven were excluded because they were less than 20 years old and another 11 because of missing data.

tFive People In the education category and eight in the employment category were excluded because of missing data. tdf = degrees of freom.

NOVEMBER 1, 1991


1113

m

(about 10% of all households in the survey area) were not accessible. Second, in 24% of the households sampled the subject could not be contacted or refused to participate. The underrepresentation of men in the sample has been observed in at least one other telephone survey2' and is consistent with nonresponse bias. Third, the accuracy of the estimates depended on the recall of the respondents. Evidence on the reliability of self-reporting of drug use suggests that the probability of agreement beyond chance in respect to never versus ever using a drug varies from about 0.20 to 0.60.22 Higher levels of agreement have been observed for drugs that are commonly used (antihypertensives) or are usually taken indefinitely (levothyroxine),22 situations analogous to ASA treatment for CVD. Fourth, the survey was restricted to London and the immediate area because of financial constraints. Generalization of the estimates to populations outside this region is limited by the nature of the population surveyed (which is largely white, English speaking and urban based); the fact that London is a research and treatment centre for cerebrovascular disease and CVD should also be taken into account. Similar nonresponse and recall biases may apply to the physician survey. Furthermore, we made no effort to ascertain the physicians' actual practices in prescribing long-term ASA therapy; rather, we studied their recommendations for hypothetical cases. With the exception of a survey done in 1983, in which 14% and 21% of physicians reported that they routinely prescribe ASA for acute myocardial infarction and after myocardial infarction respectively,23 we are unaware of any other recent survey of physicians' attitudes toward the prescription of ASA for CVD. Most of the GPs and ICNNs in our survey recommended the drug for patients with unstable angina, recent myocardial infarction or a transient

ischemic episode. All three indications were unanimously agreed on by the expert panellists and are supported by the results of well-designed clinical trials2'3 and an overview analysis.5 The higher proportion of GPs and ICNNs suggesting ASA therapy for patients with a transient ischemic episode than for those with the other two conditions may reflect the fact that this indication was the first to be established through an adequate clinical trial.3 The low level of long-term ASA use among people with overt CVD in the community (43%) is in contrast to the physicians' recommendations and is of concern. Several aspects of the survey may have led to underestimation of appropriate long-term ASA use among those with CVD. First, we did not establish whether respondents with CVD who were not using ASA were unable to tolerate the drug. Second, there was instability (i.e., a wide confidence interval) in the estimate because of a low number of Table 6: Prevalence of long-term ASA use for any reason in the community sample, by sex age and CVD status

Sex; % of subjects* Male Female

Age, yr; CVD status 50 No CVD

3

CVO present 50 No

CVD

CVD present

(n - 174) 20 (n = 5) 16 (n = 74) 36 (n- 14)

2

(n = 225) 50 (n 2), (n 129) 50 (n- 32)

`The data for the 11 respondents who did not give their age were excluded.

Table 5: Prevalence of ASA use for any

reason in the

rvommunity sample by sex, age and pattern of use Sex: % of subjects (and 95%

CLs)t Age. yr. ASA use` 50 None Occasional Long-term - 50 None

Occasional Long-term

Male

Female

73 24 3 (1. 7)

68 30 2 (1, 5)

52 28 19 (11 28)

66 20 14 (8, 19)

no ASA was taken in the 4 weeks before the interview, occasional some ASA was taken in the 4 weeks before the interview but not enough for the use to be classified as long-term according to our definition (see the

'None

Methods section). 'The data tor 'I respondents who did not give their age, ASA rotutineiy. were excluded.

1114

CAN MED ASSOC J

1991; 145 (9)

none

of whom used

LE 1 er NOVEMBRE 1991

respondents with apparent CVD. Finally, we did not distinguish between hemorrhagic and thrombotic stroke. However, validation of the prevalence of long-term ASA use for secondary treatment of CVD may be warranted in view of the particularly low level observed. Although the sampled physicians and the expert cardiology panel generally agreed that ASA therapy was not indicated in middle-aged men or women at low risk, there was less certainty about the role of the drug in asymptomatic middle-aged men and women with risk factors for atherosclerosis. These responses can be interpreted in light of the trials of ASA for primary prevention of coronary heart disease, which showed the efficacy of ASA in preventing myocardial infarction in middle-aged men at low risk.4'7"' The fact that most of the GPs and ICNNs did not advocate long-term ASA use in asymptomatic men and women at low risk may be related to a lack of familiarity with the primary prevention studies, a belief that the results of these studies are not generalizable or a belief that the risks of using ASA (i.e., side effects and assignment of asymptomatic people to the sick role) are not justified by the benefit (which, in absolute terms, was small, with no overall effect on the rate of death from cardiovascular disease). GPs were consistently more likely than ICNNs and experts to recommend ASA therapy for asymptomatic people, regardless of risk factors. One possible explanation for this difference is that GPs more frequently deal with CVD from the perspective of primary prevention. Differing perceptions of risk and benefit may also be important. In a survey of family practitioners and specialists on the use of oral anticoagulation therapy in patients with nonvalvular atrial fibrillation more family practitioners than cardiologists recommended such therapy.24 The difference was related to lower estimates of the risk of systemic embolism by cardiologists. The usual ASA dose recommended by physicians and taken by patients for CVD was 325 mg/d, a higher dose being suggested for patients with a transient ischemic episode. The use of a larger dose in this condition may reflect the findings of an earlier study of ASA therapy in threatened stroke.3 However, the use of higher doses is not supported by a direct comparison of low (300 mg/d) and high (1200 mg/d) doses of ASA in patients with a transient ischemic episode,25 by an overview analysis that indirectly showed equivalent efficacy of a lower ASA dose5 or by dose-dependent ASA toxicity.25 There have been a number of other populationbased surveys of ASA use.26-31 However, none collected data after the results of the trials of ASA for primary prevention of coronary heart disease had been published.4"l' The reference populations, definiNOVEMBER 1, 1991

tions of long-term ASA use, if provided, and means of data collection in these studies were also different from those in our study, which makes it difficult to directly assess the effect of the trials on ASA use in the community. Whether or not the results of the trials and the associated media interest have affected long-term ASA use in the community, it is reassuring that people in our survey who used ASA routinely usually did so on the advice of a physician. The potential implications of the findings of our community survey can be assessed in light of results of the Physicians' Health Study.4 From that trial it is estimated that about 315 men 50 years or older would have to take 325 mg of ASA every other day for 1 year for one fatal or nonfatal myocardial infarction to be prevented. In Ontario there are about 1.08 million men aged 50 years or more.32 Our data suggest that 84% of these men do not have overt atherosclerosis and that of the 84%, 16% use ASA routinely. This may result in the primary prevention of 460 myocardial infarctions per year in the province. This assumes that the benefit from taking ASA in terms of preventing myocardial infarction did not vary over time in the Physicians' Health Study, that the findings from that study and our community survey can be extrapolated to Ontario men and that the subjects in our survey who used ASA routinely did so for at least 1 year. A similar approach can be used to assess the possible harm from long-term ASA use, such as bleeding necessitating transfusion, in the community. In the Physicians' Health Study about 2760 men over 40 years old had to take ASA for 1 year for one such complication to occur- (data on this complication in subjects aged 50 years or more were not provided in the final report of the study). The population of Ontario men over 40 years old is about 1.64 million.32 Our data suggest that 88% of these men do not have CVD and that 13% of the 88% take ASA routinely. Thus, about 70 serious bleeding episodes per year could be expected among men in this age group as a result of ASA use under these circumstances. If the assessment of the risks and benefits of the long-term ASA use observed in our community survey is limited to these two outcomes, benefit easily exceeds risk. However, this assessment is oversimplified and does not take into consideration other adverse effects of ASA use, such as drug interactions and, possibly, intracranial hemorrhage. In conclusion, the general indication of longterm ASA therapy for symptomatic CVD appears to be well established among physicians after more than 10 years of clinical trials. However, the prevalence of long-term ASA use in people with overt CVD in the community may not be optimal. The role of ASA in the primary prevention of CVD is less CAN MED ASSOC J 1991; 145 (9)

1115

accepted, but such use is detectable in the community and may increase. The effect of ASA on the course of atherosclerosis, its side effects and the potential result of these features together with unrestricted access to the drug will continue to make it important to monitor ASA use in the community. We acknowledge the assistance of Volunteer Services, University Hospital, London, Ont., and the expert panellists: Henry J.M. Barnett, MD, FRCPC (University of Western Ontario, London), John A. Cairns, MD, FRCPC (McMaster University, Hamilton, Ont.), Robert B. C6te, MD, FRCPC (McGill University, Montreal), Jack Hirsh, MD, FRCPC (McMaster University), Martin G. Myers, MD, FRCPC (University of Toronto), John W. Norris, MD, FRCPC (University of Toronto), Eldon R. Smith, MD, FRCPC (University of Calgary), and David D. Waters, MD, FRCPC (University of Montreal). This work was supported by grant R89-34 from Physicians' Services Incorporated of Ontario. Dr. Mahon was an Ontario Ministry of Health Research Fellow during this work, Kathleen Steel holds a doctoral training grant from the Easter Seal Research Institute, Dr. Laupacis is a career scientist of the Ontario Ministry of Health, and Dr. Pederson is a National Health Research Scholar of the Department of National Health and Welfare.

References 1. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group: Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,189 cases of suspected acute myocardial infarction. Lancet 1988; 2: 34-39 2. Cairns JA, Gent M, Singer J et al: Aspirin, sulfinpyrazone or both in unstable angina. N Engl J Med 1985; 313: 1369-1375 3. Canadian Cooperative Study Group: A randomized trial of aspirin and sulfinpyrazone in threatened stroke. N Engl J

Med 1978; 299: 53-59 4. Steering Committee of the Physicians' Health Study Research Group: Final report on the aspirin component of the ongoing Physicians' Health Study. N Engl J Med 1989; 321: 129-135 5. Antiplatelet Trialists' Collaboration: Secondary prevention of vascular disease by prolonged antiplatelet treatment. BMJ 1988; 296: 320-331 6. Dalen JE, Hirsh J (eds): Second American College of Chest Physicians Conference on Antithrombotic Therapy. Chest 1989; 95 (suppl 2): 1S-169S 7. Hennekens CH, Peto R, Hutchison GB et al: An overview of the British and American aspirin studies [C]. N Engl J Med 1988; 318: 923-924

8. US Preventive Services Task Force: Aspirin prophylaxis for cardiovascular disease. Am Fam Physician 1989; 40: 117-120 9. Aspirin for prevention of myocardial infarction and stroke. Med Lett Drugs Ther 1989; 31: 77-79

1116


10. Lasagna L: Does the public need protection from itself? Clin

Ther 1987; 9: 240-242 11. Peto R, Gray R, Collins R et al: Randomised trial of prophylactic daily aspirin in British male doctors. BMJ 1988; 296: 313-316 12. 1986 Canada Census (cat 94-1 11), Statistics Canada, Ottawa, 1987: 247, 253 13. Ontario Medical Directory, College of Physicians and Surgeons of Ontario, Toronto, 1988 14. Collins R, Cairns JA, Hirsh J et al: Aspirin and other platelet active drugs. Chest 1989; 95 (suppl 2): 12S- 18S 15. Bombardier C, Tugwell P, Sinclair A et al: Preference for endpoint measures for clinical trials: results of structured workshops. J Rheumatol 1982; 9: 798-801 16. Household Facilities and Equipment (cat 64-202 [annual]), Statistics Canada, Ottawa, 1987: 45 17. O'Rourke D, Blair J: Improving random respondent selection in telephone surveys. J Mark Res 1983; 20: 428-432 18. Aaron TH, Muttitt ELC: Reactions to acetylsalicylic acid. Can Med Assoc J 1982; 126: 609-611 19. Snedecor GW, Cochran WG: Statistical Methods, 7th ed, Iowa St U Pr, Ames, Iowa, 1980: 121 20. Armitage P, Berry G: Statistical Methods in Medical Research, Blackwell Sci, Oxford, 1987: 118-119 21. Pederson LL, Bull SB, Ashley MJ et al: A population survey on legislative measures to restrict smoking in Ontario: 1. Design, methodology, and sample representativeness. Am J Prev Med 1986; 2: 307-315 22. Paganini-Hill A, Ross RK: Reliability of recall of drug usage and other health-related information. Am J Epidemiol 1982; 116: 114-127 23. Dalen JE, Goldberg RJ, Gore JM et al: Therapeutic interventions in acute myocardial infarction. Survey of the ACCP section on clinical cardiology. Chest 1984; 86: 257-262 24. Chang HJ, Bell JR, Deroo DB et al: Physician variation in anticoagulating patients with atrial fibrillation. Arch Intern Med 1990; 150: 83-86 25. United Kingdom Transient Ischaemic Attack (UK-TIA) Study Group: United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: interim results. BMJ 1988; 296: 316-

320 26. Folsom AR, Iso H, Sprafka M et al: Use of aspirin for prevention of cardiovascular disease - 1981-82 to 1985-86: the Minnesota Heart Survey. Am Heart J 1988; 116: 827-830 27. Paganini-Hill A, Choa A, Ross RK et al: Aspirin use and chronic diseases: a cohort study of the elderly. BMJ 1989; 299: 1247-1250 28. Rubin RJ, Brown DJ, Taylor JW: Public awareness of aspirin and sources of aspirin information in a rural Iowa community. J Community Health 1983; 8: 229-239 29. National Center for Health Statistics: Use Habits among Adults of Cigarettes, Coffee, Aspirin, and Sleeping Pills, 1976 (Vital and Health Statistics ser 10, no 131) (DHEW publ [PHS] 80-1559), US Govt Printing Office, Washington, 1979: 1-48 30. Stein CM, Gora NP, Macheka BM: Self-medication in urban and rural Zimbabwean communities. Br J Pharmacol 1989; 27: 741-747 31. Peach H: Trends in self-prescribing and attitudes to selfmedication. Practitioner 1983; 227: 1609-1615 32. 1986 Canada Census (cat 93-101), Statistics Canada, Ottawa, 1987: 1-7

For prescribing information see page 1 168