Factors determining compliance with screening mammography

Contents

• Abstract
• Résumé
• Introduction
• Methods
• Results
• Discussion
• References

[Table of Contents]

Résumé

[Table of Contents]

Many population surveys on the factors associated with either the intent to undergo mammography or past mammography experience have been published. Some constants have been observed: age over 60 years and poverty are associated with low screening rates,[5,12-14] and knowledge of risk factors for breast cancer and of recommendations to undergo mammography does not appear to increase compliance consistently.[5,6] Investigators who have included Health Belief Model (HBM) variables in their analyses suggest that perceived susceptibility to breast cancer and perceived barriers to mammography are the principal operative HBM variables.[15-19] Some have suggested that women who perceive themselves as being unhealthy are less likely to undergo screening mammography.[8,14]

Of all the factors associated with exposure to screening mammography, a recommendation by a physician appears to be the strongest stimulus to action.[5,6,12,15-17,20] Hence, family physicians can play an important role in improving women's participation in screening programs. However, studies of the effectiveness of interventions by family physicians have shown that up to 40% of the targeted women did not follow their physician's advice.[13,21-26] Unfortunately, none of these studies controlled for the factors known to affect mammography rates, giving no insight into the characteristics of the women who declined their physician's recommendation. Only one of these studies was conducted in Canada.[25]

Between October 1991 and June 1992 we conducted a nonrandomized controlled trial at two family medicine clinics in Montreal to evaluate the effectiveness of an opportunistic approach to screening with mammography.[27] All eligible women visiting the "experimental" clinic were systematically given a written prescription for a mammography by the family physician, regardless of the reason for the visit. A computer-generated reminder to prescribe mammography was placed on the woman's chart before the visit. There was no intervention to modify the regular process of care at the "control" clinic, where physicians reported prescribing mammography mostly during visits related to health main- tenance. At the end of the study period the odds ratio (OR) for having undergone mammography in the en- tire study population was 14.11 (95% confidence interval [CI] 7.85 to 28.58) among the women at the experimental clinic.[27] The effect of the intervention was so strong that none of the other explanatory factors considered was associated with the probability of undergoing mammography during the study period.

However, 33.6% of the women at the experimental clinic did not comply with the physician's prescription, a figure comparable to that reported in other studies.[13,21-26] Since we had already collected the information in the first study, we decided to perform a secondary analysis involving only the women at the experimental clinic to determine which factors were associated with the likelihood of undergoing screening mammography. We hypothesized that socioeconomic variables, health status, previous use of health services, risk factors for breast cancer and HBM variables could be independent factors affecting compliance with the prescribed mammography.

[Table of Contents]

Methods

The experimental clinic was a university-affiliated family medicine clinic in Montreal that served mainly a low socioeconomic, white, French-speaking population. The clinic had 13 general practitioners paid on a fee-for-service basis and 11 family medicine residents. Only 20% of the patients were followed by the residents, who acted as the attending physician under supervision. The family medicine clinic was reported as the only source of primary care during the 6 months preceding the study period by 75% of the subjects.

Patient population

Intervention

Questionnaire

• Sociodemographic characteristics: age, level of education, marital status, and economic status (poor or not poor) according to income level and number of dependants, as it pertains to the city of Montreal.[29]
• Health status and psychological well-being: the subjects' perceived health status was measured with the use of a validated French translation of the self-rated perceived health status scale.[30] Subjects were asked to qualify their perception of their state of health and to compare it with that of people their age on a 5-point Likert scale. Its reliability has been well established,[31] and it has been used in Quebec health surveys[4,32] and by Fink and Shapiro[14] in their study of the factors associated with responding to an invitation to be screened. We used the Affect-Balance Scale[33] to measure psychological well-being. It consists of 10 questions and has been properly validated.[34] Other health-related variables measured were previous breast disease, hysterectomy, hormonal replacement therapy and smoking status.
• Risk factors for breast cancer: age at menarche and menopause, age at first pregnancy, parity and family history of breast cancer were measured using questions from the Quebec health survey.[4]
• Previous use of preventive and primary care services: use of other primary care resources and gynecological services, appropriateness of Papanicolaou testing schedule, number of visits at the clinic in the previous 6 months and previous use of mammography.
• Beliefs and attitudes: subjects' beliefs and attitudes toward breast cancer in general and screening in particular were measured using the HBM and questions to evaluate their knowledge of screening recommendations and their perception of what women their age actually do. At the planning stage of the study no properly validated HBM questionnaire in relation to screening mammography was available. On the basis of previously published questionnaires[35-39] we constructed scales measuring the following HBM variables: perceived personal risk of breast cancer, perceived severity of breast cancer, effectiveness of screening and treatments, perceived barriers to screening and treatments, perceived efficiency in performing breast self-examination, knowledge of recommendations for screening mammography and perception of the social norm concerning breast cancer screening. Since the telephone interview was held after the intervention, questions about potential barriers to screening asked for reasons why the woman would refuse to undergo another mammogram if she had already had one.

Statistical analysis

[Table of Contents]

Results

Of the 171 women 113 (66.1%) underwent the prescribed mammography within 2 months after the visit to the clinic. The difference in the distribution of reasons for the visit between the women who did and those who did not undergo mammography was not statistically significant, although the proportion of women who visited the clinic for a medical problem not related to a gynecological reason or a health check-up was greater in the noncompliant group (69.5% v. 86.4%; p = 0.13).

The overall response rate to the telephone questionnaire was 87.1% (149/171). Nine women, five of whom were in the noncompliant group, refused to answer the questions. Thirteen women, nine of whom were in the noncompliant group, could not be reached. Thus, the response rate was 92.9% in the compliant group and 75.9% in the noncompliant group (p = 0.02). The respondents and nonrespondents who did or did not comply with the prescribed mammography were comparable in terms of age and type of visit (with or without an appointment); however, slightly more nonrespondents than respondents in the compliant group visited the clinic for reasons other than health check-up or gynecological problems (Table 1). The distribution of the principal sociodemographic and medical characteristics of the respondents according to mammography status are shown in Table 2. There was no significant difference in any of the variables between the compliant and noncompliant groups.

The results of the final logistic regression model appear in Table 3. The model predicts noncompliance. It accounted for 87.5% of the outcome (c2 for goodness of fit = 164.4; p = 0.0001). None of the sociodemographic variables affected the compliance rate. Smoking was the only health status variable associated with the outcome, in all steps of the analysis. Neither the overall score for breast cancer risk factors nor a family history of breast cancer was related to compliance.

Appropriateness of the Papanicolaou testing schedule was negatively associated with noncompliance (OR 0.49, 95% CI 0.29 to 0.84; p = 0.009) and the number of visits to the clinic within the previous 6 months was positively associated with noncompliance (OR 1.24, 95% CI 0.99 to 1.58; p = 0.06) when utilization variables were entered in the fifth step of the analysis. However, the effect of these two variables was diminished when the number of previous mammograms was entered into the equation (Table 3). This suggests that these two factors were associated with compliance in a more general pattern of health services utilization in which preventive services are less valued and that the number of previous mammograms was the strongest predictor of further participation in mammography screening.

Having undergone mammography previously was the strongest predictor of compliance. Such women were less likely to be noncompliers than those who had not undergone a previous mammogram (OR 0.11, 95% CI 0.02 to 0.51; p = 0.005). Women who said that a recommendation by a physician would convince them to undergo mammography were also less likely than the others to be noncompliers (OR 0.21, 95% CI 0.07 to 0.60; p = 0.004). Perceived barriers to mammography were the only HBM variables found to affect compliance. The women who expressed fear of mammography ("I fear the x-rays," "I fear the results") and time constraints ("I do not have the time," "I cannot miss work") were more likely than the others to be noncompliant (OR 2.09, 95% CI 1.08 to 4.02, p = 0.03; and OR 3.07, 95% CI 1.21 to 7.80, p = 0.02, respectively). Finally, the women who did not smoke were less likely than those who did smoke to be noncompliant (OR 0.43, 95% CI 0.22 to 0.86; p = 0.02).

[Table of Contents]

Discussion

However, one third of the women chose not to undergo the prescribed mammography; this rate of noncompliance is similar to that reported in other primary care settings.[13,21-26,41] Our analysis of the women's characteristics suggests that they were particularly reluctant toward screening mammography and health maintenance behaviours in general, since they were more likely than the compliant women to have never obtained a mammography, to have infrequent Papanicolaou tests and to be smokers. As observed with community-based screening interventions,[8,15,17,19] the noncompliant women were more likely to express fear of mammography and to report being under time constraints. However, perceived susceptibility to breast cancer and perceived health status did not seem to influence behaviour as it did in some of those studies.[17,19,22] Finally, the noncompliant women expressed more frequently than the compliant women the view that a recommendation by a family physician was unlikely to make them undergo screening mammography, a finding we did not expect.

Obviously, our findings do not apply to all clinical settings. First, they represent a set of constraints that are operative in the absence of any direct financial barriers to screening; in Quebec the cost of physician visits and mammography are covered by the provincial health insurance program. Second, the study population was a culturally homogenous one. Third, a substantial proportion of the women belonged to a lower socioeconomic group; however, we deliberately targeted this population because data from the 1986 Quebec health survey revealed that up to 72% of such women had never undergone mammography.[4]

The fact that our study was conducted in an academic setting may affect the generalizability of the results. It explains in part the high rate of adherence to the study protocol by the physicians and the strong practice commitment that was obtained through group discussion before the study began. Also, computer-generated reminder systems are not widely available in family practices in Canada.[42] However, the mammography prescription rate of 40% before the study began was comparable to the one observed in the control clinic, which was a high-volume, nonacademic, fee-for-service practice located in the same neighbourhood.[27] This suggests that the physicians in the experimental clinic were similar to others practising in the same neighbourhood. Of course, the objective of this study was not to determine the physicians' behaviours but, rather, to determine the women's reactions to the physicians' intervention. The issue of representativeness was more crucial for the women than for the physicians. Women in our study were representative of women from a francophone background living in a less privileged neighbourhood of a large city such as Montreal.

Would we have obtained different answers to the HBM questions if the questionnaire had been administered before the women had undergone the mammography? We will never know. We chose to administer the questionnaire 2 months after the visit in order not to interfere with the effect of the intervention. This may have led to an overestimation of the apparent detrimental effect of believing that mammography is uncomfortable. On the other hand, it may have meant that fear of discomfort was not a strong barrier for all women who held that belief. In an attempt to overcome this problem, we formulated the questions about the barriers to mammography in relation to potential reasons not to have another mammogram if women answered that they had undergone one previously.

Would immediate availability of mammography have produced a different compliance rate, particularly since time constraints were expressed as barriers to screening? We do not think so, since comparable compliance rates were achieved in settings in which women had to make appointments,[22-25] as in our study, and in settings in which the service was available immediately.[13] The only exception was the study by Wolosin,[41] who found that offering to make the appointment for the patient right after the visit during which mammography was recommended and mailing a reminder 4 days before the appointment increased the likelihood of compliance. However, the effectiveness of this approach varied from 58% to 84% depending on the setting in which it took place. Unfortunately, Wolosin did not provide any explanation for this range.

The different rates of response to the questionnaire in the compliant and the noncompliant groups should be addressed. Although the absolute numbers were small, the difference was statistically significant. It was mostly due to the larger number of women who could not be reached rather than to those who refused to answer the questions. Analysis confirmed that there was no relation between the response rate, the mean age of the respondents and the type of visit in each study group. However, in the compliant group nonrespondents were more likely than respondents to have visited the clinic because of a specific problem not related to health maintenance or gynecological problems. This subgroup of women may not have been representative of our regular clientele at the clinic. As a result, the impact of negative attitudes toward health care professionals or of the effect of the presence of priorities more important than prevention at the time of the visit may have been underestimated.

Notwithstanding these constraints, we believe that our study has many strengths. First, the outcome measure was based on direct evidence of mammograms done rather than on self-reported measures, which have been shown to overestimate mammography rates.[43] Second, we dealt with screening mammography only and not a mix of diagnostic and screening examinations, as in most studies. Finally, unlike most studies conducted in primary care settings, our explanatory model took into account a vast array of factors known to be associated with the use of mammography screening.

Our results show that some women who are not likely to be reached through a population screening strategy,[6,8,10,11] namely those who are poor, over 60 years of age and less educated, respond favourably to an intervention by their family physician. These results are important in the present Canadian context, since all provinces are planning to create dedicated screening centres. The creation of such centres must not occur in isolation from the actual primary care network.

Unfortunately, our results show that some women will not respond positively to their physician's recommendation to undergo screening mammography for reasons that do not appear to be directly related to social or medical characteristics. These women seem to be less oriented toward prevention and to hold negative views of advice from health care professionals. Their priorities may be elsewhere. It is also possible that screening mammography provokes too much fear. Research of a more qualitative nature would help to explore the reasons why some women are so reluctant to undergo screening mammography and how their worries or beliefs may be addressed more appropriately. New innovative approaches using the power of peer groups may be more effective in reaching these individuals.

This study was supported by grant 911357-104 from the Fonds de la recherche en santé du Québec.

[Table of Contents]

References

1. The Workshop Group: Reducing deaths from breast cancer in Canada. CMAJ 1989; 141: 199-201
2. Battista RN: Adult cancer prevention in primary care: patterns of practice in Quebec. Am J Public Health 1983; 73: 1036-1039
3. Battista RN, Palmer CS, Marchand BM et al: Patterns of preventive practice in New Brunswick. CMAJ 1985; 132: 1013-1015
4. Santé Québec: Et la santé, ça va?, Publications du Québec, Québec, 1988: 337
5. Mah Z, Bryant H: Age as a factor in breast cancer knowledge, attitudes and screening behaviour. CMAJ 1992; 146: 2167-2174
6. Vernon SW, La Ville EA, Jackson GL: Participation in breast screening programs: a review. Soc Sci Med 1990; 30: 1107-1118
7. Zapka JG, Costanza ME, Harris DR et al: Impact of a breast cancer screening community intervention. Prev Med 1993; 22: 34-53
8. Fink R, Shapiro S, Lewison J: The reluctant participant in a breast cancer screening program. Public Health Rep 1968; 83: 479-490
9. Hurley SF, Jolley DJ, Livingston PM et al: Effectiveness, costs, and cost-effectiveness of recruitment strategies for a mammographic screening program to detect breast cancer. J Natl Cancer Inst 1992; 84: 855-863
10. Miller AB, Baines CJ, To T et al: Canadian National Breast Screening Study: 2. Breast cancer detection and death rates among women aged 50 to 59 years. CMAJ 1992; 147: 1477-1488
11. Kee F, Telford AM, Donaghy P et al: Attitude or access: reasons for not attending mammography in Northern Ireland. Eur J Cancer Prev 1992; 1: 311-315
12. Rimer BK, Trock B, Engstrom PF et al: Why do some women get regular mammograms? Am J Prev Med 1991; 7: 69-74
13. Mandelblatt J, Traxler M, Lakin P et al: Breast and cervical cancer screening of poor, elderly, black women: clinical results and implications. Am J Prev Med 1993; 9: 133-138
14. Fink R, Shapiro S: Significance of increased efforts to gain participation in screening for breast cancer. Am J Prev Med 1990; 6: 34-41
15. Calnan M: The Health Belief Model and participation in programmes for the early detection of breast cancer: a comparative analysis. Soc Sci Med 1984; 19: 823-830
16. Champion VL: Compliance with guidelines for mammography screening. Cancer Detect Prev 1992; 16: 253-258
17. Stein JA, Fox SA, Murata PJ et al: Mammography usage and the Health Belief Model. Health Educ Q 1992; 19: 447-462
18. Fulton JP, Buechner JS, Scott HD et al: A study guided by the Health Belief Model of the predictors of breast cancer screening. Public Health Rep 1991; 106: 410-420
19. Hyman RB, Baker S, Ephraim R et al: Health Belief Model variables as predictors of screening mammography utilization. J Behav Med 1994; 17: 391-406
20. Zapka JG, Costanza ME, Greene HL Jr: Breast cancer screening and primary care providers. Prog Clin Biol Res 1989; 293: 57-63
21. Nattinger AB, Panzer RJ, Janus J: Improving the utilization of screening mammography in primary care practices. Arch Intern Med 1989; 149: 2087-2092
22. Burack RC, Liang J: The early detection of cancer in the primary-care setting: factors associated with the acceptance and completion of recommended procedures. Prev Med 1987; 16: 739-751
23. Taplin SH, Anderman C, Grothaus L et al: Using physician correspondence and postcard reminders to promote mammography use. Am J Public Health 1994; 84: 571-574
24. Swinker M, Arbogast JG, Murray S: Why do patients decline screening mammography? Fam Pract Res J 1993; 13: 165-170
25. Bass B, Pross D, Bell P: Recruitment for breast screening in a rural practice. Trial of a physician's letter of invitation. Can Fam Physician 1994; 40: 1730-1739
26. Kee F: Do general practitioners facilitate the breast screening programme? Eur J Cancer Prev 1992; 1: 231-238
27. Beaulieu MD, Roy D, Béland F et al: Dépistage du cancer du sein. Efficacité d'une intervention par le médecin de famille. Union Med Can 1994; 123: 154-162
28. Lamberts H, Wood M: ICPC. International Classification of Primary Care, Oxford University Press, Oxford, England, 1987: 201
29. Conseil national du bien-être social: Profil de la pauvreté, 1980-1990, Gouvernement du Canada, Ottawa, 1992
30. Fylkesnes K, Forde OH: Determinants and dimensions involved in self-evaluation of health. Soc Sci Med 1992; 35: 271-279
31. Pampalon R, Saucier A, Berthiaume N et al: Des indicateurs de besoin pour l'allocation interrégionale des ressources, Ministère de la Santé et des Services sociaux du Québec, Québec, 1994
32. Santé Québec: Et la santé, ça va en 1992-1993?, Publications du Québec, Québec, 1995: 412
33. Bradburn NM (ed): The Structure of Psychological Well-being, Aldine, Chicago, 1969
34. McDowell I, Newell C: Measuring Health. A Guide to Rating Scales and Questionnaires, Oxford University Press, New York, 1987
35. Jette AM, Cummings KM, Brock BM et al: The structure and reliability of health belief indices. Health Serv Res 1981; 16: 81-98
36. Becker MH: The Health Belief Model and personal health behavior. Health Educ Monogr 1974; 2: 324-473
37. Janz NK, Becker MH: The Health Belief Model: a decade later. Health Educ Q 1984; 11: 1-47
38. Fletcher SW, Morgan TM, O'Malley MS et al: Is breast self-examination predicted by knowledge, attitudes, beliefs, or sociodemographic characteristics? Am J Prev Med 1989; 5: 207-215
39. Baines CJ, To T, Wall C: Women's attitudes to screening after participation in the National Breast Screening Study. A questionnaire survey. Cancer 1990; 65: 1663-1669
40. Hosmer DW, Lemeshow S: Applied Logistic Regression, John Wiley & Sons, New York, 1989
41. Wolosin RJ: Effect of appointment scheduling and reminder postcards on adherence to mammography recommendations. J Fam Pract 1990; 30: 542-547
42. Bass MJ, Elford RW: Preventive practice patterns of Canadian primary care physicians. Am J Prev Med 1988; 4 (4, suppl): 17-23
43. Fulton-Kehoe D, Burg MA, Lane DS: Are self-reported dates of mammograms accurate? Public Health Rev 1992-93; 20: 233-240