- Email: [email protected]
The Belief in Conventional Medicine and Adherence to Treatment in Non–Insulin- Dependent Diabetes Mellitus Patients
ORIGINAL ARTICLES
The Belief in Conventional Medicine and Adherence to Treatment in Non–InsulinDependent Diabetes Mellitus Patients Maria Eugenia Garay-Sevilla Juan Manuel Malacara H. Felipe Gonza´lez-Parada Lourdes Jorda´n-Gine´s
ABSTRACT We investigated the role of belief in conventional medicine, the type of medical care, and familiar and socioeconomic factors on the adherence to treatment in non–insulin-dependent diabetes mellitus (NIDDM) patients. In a cross-sectional design, we selected 156 patients from two institutions, who agreed to fill out a questionnaire, which included general data, socioeconomic level, somatometric data, type of medical care, complications, if they had friends and relatives with diabetes, the family function, and a score on the belief in conventional medicine. Factors associated with adherence to diet and medication were analyzed. Patients had a mean age of 55.6 years and 8.9 years since diagnosis. A total of 51.3% of them were not covered by social security, and 62.8% received attention by a general physician.
Patients under the care of a specialist had better adherence to diet and medication, and better belief in conventional medicine. The principal factor associated with adherence to medication and diet was the belief in conventional medicine (p ⬍ 0.001 in both). Adherence to diet was also associated with the socioeconomic level (p ⫽ 0.001) and years since diagnosis (p ⫽ 0.004). Adherence to medication was also associated with schooling (p ⫽ 0.001). We concluded that belief in conventional medicine is strongly associated with adherence to treatment and other factors such as schooling, socioeconomic level, and medical care under a specialist; adherence to diet was better in patients with more years since diagnosis. ( Journal of Diabetes and Its Complications 12; 5: 239–245, 1998.) 1998 Elsevier Science Inc.
INTRODUCTION
the therapeutic plan. Nonadherence is the most serious problem facing medical practice, as only a third of the patients properly follow medical indications. Non–insulin-dependent diabetes mellitus (NIDDM) has a high prevalence and its control requires a careful compliance to medical indications, which demands a permanent behavioral change to comply with diet and medication and to assure a continued monitoring of metabolic control. The reported frequency of nonadherence to diet varies from 35% to 75%.2 Interestingly, patients attribute their difficulties to follow medical indications to environmental and economical factors,
A
successful ambulatory treatment for patients with chronic diseases depends on various factors. Becker1 identified four of them: Recognition of il health, the diagnosis of illness, planning of treatment, and the patient’s adherence to
Instituto de Investigaciones Me´dicas, Universidad de Guanajuato, Leon, Mexico Reprint requests to be sent to: Dr. Juan Manuel Malacara H., Instituto de Investigaciones Me´dicas, Universidad de Guanajuato, 20 de Enero 929, Apdo, Postal 874, 37320 Leo´n Gto, Mexico. Journal of Diabetes and Its Complications 1998; 12:239–245 1998 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
1056-8727/98/$19.00 PII S1056-8727(97)00075-5
240
GARAY-SEVILLA ET AL.
as well as to personal problems with the spouse, family, and health professionals.3 In contrast, physicians ascribe nonadherence to personal factors.4 The interactions within the family may have an effect on adherence to treatment, a better adherence is associated with non aggressive, genuine and frank discussions about diabetes.5,6 We have found that adherence is related to some aspects of the family function, such as control of behavior,7 these family factors may interact with the social support they receive. An appropriate compliance to treatment for a lifelong disease requires that the model for health service is fully accepted and is not in conflict with the cultural background of the patient. The patients’ attitude toward disease has been studied by means of the health belief model,8–11 in order to understand and predict adherence to treatment. It has been proposed that adherence to treatment of diabetes increases with perceived susceptibility to disease, perceived severity, perception of benefits, barriers or costs, and cues to motivate changes of behavior. Compliance to diet is difficult to archive because food is an important cultural feature, and traditional food in most cultural groups does not comply with the American Diabetes Association (ADA) recommendations.12 Foods are also a common subject for healthrelated myths. Conventional medicine may be considered by diverse indigenous cultures as an alien institution. Non-scientific types of medicine are still prevalent in most cultures. Modern medicine, developed in industrialized countries, includes facilities for medical attention and diagnostic and treatment procedures that may be in conflict with traditional values of native societies. Based on these considerations, we propose that belief in modern medicine, may play a role in compliance to treatment for diabetes. In this study, we examined the interaction of the belief in medicine with compliance to treatment, the family support, and other factors in patients with NIDDM. METHODS Study Design. We studied 156 patients (43 men and 113 women) with NIDDM, according to the diagnostic criteria of the American Diabetes Association,13 in a cross-sectional design. Two groups of patients were examined: With and without social security coverage. The latter group included 80 patients who attended the Instituto de Investigaciones Me´dicas requiring free medical attention and laboratory test for control of diabetes (no free medication). The group with social security included 76 patients recruited from a Social Security Hospital in Leo´n, Mexico, where they received free medical attention. The patients from this group were invited during the weekly diabetes club meeting, to have the purpose of the study explained to them. The
J Diab Comp 1998; 12:239–245
acceptance rate was close to 30%. The socioeconomic status was middle and low in both groups. Inclusion Criteria. For inclusion in the study, we considered non-hospitalized NIDDM patients without pregnancy, intercurrent diseases, severe metabolic alterations, or complications that required additional treatment. The patients showed adequate physical and mental capacity to answer questions about their disease. Data Collection. Patients answered a questionnaire by direct questioning by an investigator including the following items: General data: Age, years since diagnosis of diabetes, schooling (in years), occupation, monthly family income (scores from 0 to 4 with 0 equal to less than Pesos$100.00 and 4 more than Pesos$5,000.00/month), and age of the spouse. The socioeconomic level was obtained with a score combining schooling and family income. The quartiles for each value were added, and the final score of socioeconomic level ranged from 1 to 7. Somatometric variables: Weight (kg), height (cm) and body-mass index (BMI) were obtained. Type of medical care: The type of medical care considered two aspects: coverage by social security, and the physician, general practitioner or specialist (internist or endocrinologist). Complications: We collected antecedents of myocardial infarction, stroke, angor pectoris, cataract, amaurosis, diabetic feet, and amputations. Family function: Family function was assessed by means of a questionnaire used in previous studies7,14 inquiring about communication within the family, and adding the answers to four questions scored 1–4. The score range was 4 to 16. Belief in conventional medicine: This was assessed with an instrument designed by us. The content validity of the test was approached including 15 topics on the patient’s most frequently mentioned objections to accepting medicine (or reasons to seek non-medical treatments), for health problems related to chronic diseases, maternal–child health, vaccination, and infectious diseases. The construct validity15 was tested in a pilot study on twenty NIDDM patients, analyzing intrinsic consistency with a correlation matrix. For the final instrument were considered the six questions with a higher intrinsic correlation. The questions included were (1) the frequency in which they sought medical care when they felt ill; (2) the extent of confidence on their physician’s proficiency; (3) their perception of the time required by medical care; (4) their perception on the cost of medical care; (5) their use of medical attention for prevention; (6) their perception of the effectiveness of herbs and vegetables to cure diabetes. Answers to each
J Diab Comp 1998; 12:239–245
question were scored with a Likert scale ranging from 1 to 4. Question six had a negative polarity (higher scores for answer of “never”). The scores for each question were added for a final score ranging from 6 to 24. Relatives and friends with diabetes: The questionnaire collected patient awareness of close relatives or friends with uncomplicated or complicated diabetes, or recently deceased as a result of complications of diabetes (answers yes/no). Adherence to treatment: Adherence to medical treatment was studied in two aspects: adherence to diet and to medication. These aspects were evaluated with the questionnaire used in our previous work.7 Adherence to diet was assessed with seven questions scored 1 to 4, answers were added to obtain a score range of 7–28. The questionnaire on adherence to medication had three questions, and the score was added, with a result ranging from 3 to 12. Higher scores indicated better adherence to treatment. Metabolic control: Blood glucose was measured with a glucose oxidase method (GOD-PAD, Lakeside) and glycated hemoglobin with an ion exchange chromatography method (Sigma Chemical, St. Louis, MO, USA) (normal values, 5.5%–8.5%). Food intake: Food intake was assessed in a semi-quantitative manner asking the type and amount of food recently consumed. The category of was selected from a list of the 29 foods most frequent consumed in the region. We calculated daily intake of protein, fat (in grams) and calories using the tables of the Instituto Nacional de la Nutricio´n (Mexico City).16 We registered the type of dietary prescription: oral or written, quantitative or qualitative. Statistical Analysis. The characteristics of the total group of patients were analyzed with descriptive parametric statistics. Differences between groups under the care of a general physician and a specialist, and between patients with or without social security were analyzed with an unpaired Student’s t test. The determinants of metabolic control were tested with a multiple-regression analysis, taking blood glucose and glycated hemoglobin as dependent variables, and age, years since diagnosis, adherence to diet, adherence to medication, and intake of calories, fat, and proteins as candidate regressors. The determinants of adherence to diet and medication were tested with multiple-regression analysis taking as candidate regressors: age, years since diagnosis, schooling, age of the spouse, socioeconomic level, medical service, type of physician, friends and relatives with diabetes or deceased as a result of diabetes, communication in the family, BMI, and belief in conventional medicine.
ADHERENCE TO TREATMENT IN NIDDM
241
RESULTS Characteristics of the Patients. The characteristics of the patients are shown in Table 1. The means for age and years since diagnosis were 55.6, and 8.9, respectively. Schooling was low (mean of 3.7 years). Most patients had middle or low socioeconomic level (1–4 on the scale of 1–7). They were mainly obese (mean BMI was 28). In relation to medical care, most patients were under the attention of a general physician. Among complications studied, amaurosis was found in 12.2% and cataract in 8.3% of the cases. Family Function. In regard to family function, communication was mainly clear and direct, and no difference was found among the groups. This factor was not collected in 25 widowers, 19 who were not married, three who did not have children, and four who did not respond. A total of 75% of the patients had at least a relative with diabetes, and 61% had a diabetic friend. These factors did not have an influence on adherence to treatment. Medical Care. The analysis of the characteristics of patients according to their physician’s type is shown in Table 2. Patients under a general physician’s care had fewer years since diagnosis, had lower scores of adherence to diet and medication, and lower scores of belief of conventional medicine. No differences between groups were found for metabolic control, food intake, or rates of complications. The characteristics for groups with and without social security coverage are shown in Table 3. Patients with social security were older and with more years since diagnosis, and had lower glucose levels than those without social security. Adherence to treatment was similar in both groups, and belief in conventional medicine had marginally higher scores in the social security group (p ⫽ 0.04). Adherence to Treatment. Adherence to diet had a mean of 20.1 (95% CI, 20.8–19.4) from a range of 7–28, and adherence to medication a mean of 10.7 (95% CI, 11.0–10.4) from a range of 3–12. A total of 34.6% of the patients did not follow their prescribed diet, and 6.4% said they followed it strictly. The type of diet prescription was quantitative and written for 32%, qualitative and written for 5.8%, quantitative but not written for 21%, and qualitative and not written for 3.8% of the patients. Metabolic Control. Metabolic control was not satisfactory in most patients: mean glucose level was 192.1 mg/dL and mean HbA1c was 12.9%. In the regression analysis for determinants of metabolic control, adherence to medications was determinant for both glucose levels (r ⫽ ⫺0.182, p ⫽ 0.023) and glycated hemoglobin (r ⫽ ⫺0.174, p ⫽ 0.029).
242
GARAY-SEVILLA ET AL.
J Diab Comp 1998; 12:239–245
TABLE 1. CHARACTERISTICS OF THE PATIENTS Variable Socioeconomic Level 1–2 (low) 3–4 (middle) 5–7 (high) Medical care With social security Without social security General physician Specialist Complications Myocardial infarction Cataract Amaurosis Diabetic foot Amputations Family function: Communication within the family Clear and direct Clear and indirect Masked and direct Unmasked and indirect Not applicable Relatives and friends with diabetes Relatives with diabetes Relatives with complicated diabetes Relatives deceased as a result of diabetes Friends with diabetes Friends with complicated diabetes Friends deceased as a result of diabetes Metabolic control Glucose (mg/dL) HbAlc % Other characteristics Age (years) Age of the spouse (years) Time since diagnosis (years) Schooling (years) BMI Protein intake Fat intake Calorie intake
n
%
55 66 30
35.3 42.3 19.1
76 80 98 55
48.7 51.3 62.8 35.2
2 13 19 8 3
1.3 8.3 12.2 5.3 1.9
81 17 6 1 51
51.9 10.9 3.8 0.6 32.7
117 50 65 95 37 41
75.0 32.1 41.7 60.9 23.7 26.3
Mean
95% CI
192.1 12.9
181.2–202.9 12.4–13.3
55.6 52.2 8.9 3.7 28.0 50.1 40.7 1230
53.8–57.5 53.5–56.9 7.8–10.1 3.2–4.2 28.9–27.2 47.1–53.5 37.8–43.7 1157–1304
CI, confidence interval; HbAic, glycated hemoglobin; BMI, bodymass index.
Factors Associated with Adherence to Treatment. The factors associated with adherence to treatment as analyzed by means of the multiple regression procedure are shown in Table 4. Adherence to diet and medication were highly associated with belief in conventional medicine (Figure 1); additionally, adherence to diet was associated negatively with socioeconomic level and positively with years since diagnosis; and adherence to medication correlated negatively with schooling. DISCUSSION Conventional medicine has exhibited an explosive evolution, resulting in the scientific and technical ad-
vances, developed in industrialized countries. However, modern medical concepts and techniques may be in conflict with cultural values from traditional societies. This may hamper the acceptance of medical treatments in marginated communities who preserve traditional values. Societies with ancient traditions still have mythical concepts about health and disease. They seek magic cures for their ailments, and are reluctant to accept long-lasting medical treatments, mainly if a prompt benefit is not experienced. This is also true when medication have components and labels revealing an alien origin. Diet is an important element in the treatment
J Diab Comp 1998; 12:239–245
ADHERENCE TO TREATMENT IN NIDDM
243
TABLE 2. CHARACTERISTICS OF PATIENTS ACCORDING TO TYPE PHYSICIAN’S CARE
Age (years) Years since diagnosis Weight (kg) Height (cm) BMI Socioeconomic level Adherence to diet Adherence to medication Belief in conventional medicine Glucose (mg/dL) Glycated hemoglobin (%) Intake of calories Intake of fat (g) Intake of protein (g)
General Mean (SD)
Specialist Mean (SD)
t
p
54.3 (11.6) 7.9 (6.5) 67.3 (12.6) 155 (0.09) 28.0 (5.0) 3.1 (1.5) 19.4 (4.4) 10.5 (1.7) 17.9 (3.8) 194 (71) 13.2 (3.1) 1229 (422) 39.3 (16) 48.9 (17)
58.1 (11.5) 11.2 (8.3) 69.9 (13.0) 158 (0.09) 28.0 (5.4) 3.3 (1.4) 21.5 (4.1) 11.1 (1.8) 19.8 (3.4) 190 (65) 12.5 (2.7) 1211 (521) 42.2 (23) 51.0 (20)
⫺1.92 ⫺2.79 ⫺1.16 ⫺1.96 0.02 ⫺0.82 ⫺3.00 ⫺2.38 ⫺3.05 0.40 1.50 0.22 ⫺0.96 ⫺0.68
0.056 0.005 0.247 0.052 0.981 0.411 0.003 0.019 0.003 0.689 0.136 0.82 0.340 0.50
n (%) 4 (7.3) 3 (5.4) 4 (7.3)
2 0.16 3.82 0.724
p 0.684 0.051 0.395
n (%) 9 (9.2) 16 (17.4) 4 (4.1)
Cataract Amaurosis Diabetic foot BMI, body-mass index.
of NIDDM, and accepted diet guidelines are in conflict with alimentary traditions of several societies. Food is a factor of cultural identification for every ethnic group. In Mexico, traditional food has a high content of carbohydrate and saturated fat, and a low fiber content. These considerations explain difficulties for the acceptance of modern medicine. In this study, we developed a simple questionnaire intended to evaluate belief
in conventional medicine. The association of the scores obtained with compliance to medication and diet are in agreement with a predictive value of this instrument15 for use in subsequent studies. The findings from this study should be considered to design adequate strategies for a more culture-based approach for behavior modification. An important strategy should be a careful education on the meaning and practical appli-
TABLE 3. CHARACTERISTICS OF PATIENTS ACCORDING TO SOCIAL SECURITY
Age (years) Years since diagnosis Weight (kg) Height (cm) BMI Socioeconomic level Adherence to diet Adherence to medication Belief in conventional medicine Glucose (mg/dL) Glycated hemoglobin (%) Intake of calories Intake of fat (g) Intake of protein (g) Cataract Amaurosis Diabetic foot BMI, body-mass index.
Social Security Mean (SD)
No Social Security Mean (SD)
t
p
59.3 (9.8) 10.2 (7.7) 68.5 (13.1) 157 (0.1) 27.7 (4.8) 3.3 (1.5) 20.3 (4.2) 10.8 (2.0) 19.2 (3.4) 178 (60) 12.5 (2.7) 1261 (490) 42.8 (19.8) 52.3 (18.8)
52.1 (12.1) 7.8 (6.8) 68.5 (13.2) 156 (0.1) 28.3 (5.4) 3.0(1.4) 20.0 (4.5) 10.6 (1.7) 18.0 (4.0) 206 (73) 13.3 (3.2) 1200 (439) 38.8 (17.5) 48.0 (18.5)
4.06 2.00 ⫺0.13 1.00 ⫺0.73 1.17 0.41 0.59 2.05 ⫺2.61 ⫺1.60 0.82 1.36 1.42
⬍0.001 0.046 0.989 0.320 0.466 0.244 0.682 0.555 0.042 0.010 0.102 0.414 0.175 0.158
n (%) 6 (7.9) 9 (11.8) 2 (2.6)
n (%) 7 (8.7) 10 (12.5) 6 (7.5)
2 0.04 0.01 1.90
p 0.847 0.900 0.160
244
GARAY-SEVILLA ET AL.
J Diab Comp 1998; 12:239–245
TABLE 4. REGRESSION ANALYSIS TAKING AS DEPENDENT VARIABLE ADHERENCE TO TREATMENT 
r
p
0.355 ⫺0.234 0.206
0.375 ⫺0.258 0.228
⬍0.001 0.001 0.004
0.458 ⫺0.223
0.473 ⫺0.253
⬍0.001 0.001
Regressor Adherence to diet Belief in conventional medicine Socioeconomic level Years since diagnosis Adherence to medication Belief in conventional medicine Schooling
cations of scientific medicine, a factor largely overlooked in the education programs for diabetic patients. In our group of study, a large proportion of the patients had partial or complete noncompliance with medical indications, and adherence was strongly associated with years since diagnosis of diabetes. This agrees with our previous report,7 but other investigators, found only a weak association between both factors.17 We interpreted this finding to mean that in the
FIGURE 1 Association of adherence to diet (upper panel) and adherence to medication (lower panel) with belief in medicine, in patients with non–insulin-dependent diabetes mellitus. The p values for both cases were ⬍ 0.001.
progress of time, diminishes the patient’s denial of disease, frequent at early stages.18 After the onset of complications, the patient accepts the disease, and may use it as an adaptive contrivance to obtain more care from the family. The disease may frequently become a central subject modifying the dynamics of family function. It is possible that the association between adherence to diet and years since diagnosis, also results from the survival bias that appears in cross-sectional studies. Better compliants should have better survival. It was surprising to find that patients with lower schooling and socioeconomic level comply better with medical treatment. Before interpretation, one must consider that the group of study had a very low schooling (mean of 3.7 years). Therefore, conclusions should be limited to a narrow socioeconomic range. Knowledge of diabetes, an important factor for adherence to treatment,7 is not acquired in elementary school. This finding emphasizes the importance of information on diabetes given to groups of patients. Another source for information may be the mass communication media, that may facilitate compliance to treatment, by means of a complete information of the devastating results from complications. Patients under the specialist’s care had higher scores of belief in conventional medicine and adherence both to diet and to medication as compared with those under the care of a general physician. These results are in agreement with the findings from a study in hypertensive patients17 reporting better adherence in patients treated by a cardiologist than by a general physician. The explanation of these results is that patients who better accept conventional medicine look for the advantages of a specialist’s care, and are more prone to better compliance to treatment. An important aspect of social support for diabetic patients is a proper interaction with the health-care providers and with the health-care system.19,20 We found only a marginally higher belief in conventional medicine in the group with social security. Support received within the family is a most important aspect of social support. Communication is relevant to enhance family support. The homogeneity of
J Diab Comp 1998; 12:239–245
this characteristic in the group of study did not permit us to examine the influence of communication as part of the social support for better adherence to treatment, as in our previous study.7 Several investigations have shown the impact of morbidity or mortality of diabetic relatives and friends on a patient’s compliance to treatment, permitting a better perception of the dreadful consequences of diabetes complications.8,21 Yet, our results did not show a significant association of these factors. Differences can be explained by different methods and criteria, because we did not use the health belief model.21 This model proposed that the acceptance of medical advice depend on the perception of vulnerability to illness, including acceptance of the diagnosis, and perceived benefits. In NIDDM patients, both may be low, in particular, perception of the benefits of appropriate metabolic control. We concluded that the adherence to treatment in NIDDM patients is strongly associated with belief in conventional medicine and negatively associated to schooling and socioeconomic level. Adherence to diet was associated with years since diagnosis. The belief in medicine was higher in patients under the care of a specialist, and was not clearly associated with social security coverage. ACKNOWLEDGMENTS This work was supported in part by grant 3417-M from CONACYT.
REFERENCES 1.
Becker MH: Patient adherence to prescribed therapies. Med Care 23:539–555, 1985.
2.
Kurtz SM: Adherence to diabetes regimens: Empirical status and clinical applications. Diabetes Educ 16:50–56, 1990.
3.
Ary DV, Toobert D, Wilson W, Glasgow RE: Patient perspectives on factors contributing to nonadherence to diabetes regimens. Diabetes Care 9:168–172, 1986.
4.
House WC, Pendelton L, Parker L: Patients versus physicians attributions of reasons for diabetic patients’ non compliance with diet. Diabetes Care 9:434, 1986.
5.
6.
Anderson BJ, Miller IP, Auslander WF, Santiago JV: Family characteristics of diabetic adolescents: Relationship to metabolic control. Diabetes Care 4:586–594, 1981. Carney RM, Schechfer K: Familial eating patterns and metabolic control in juvenile onset diabetes. Diabetes 3(suppl 2):13A, 1982.
ADHERENCE TO TREATMENT IN NIDDM
245
7.
Garay-Sevilla ME, Nava LE, Malacara JM, Huerta MR, Diaz de Leo´n J, Mena A, Fajardo ME: Adherence to treatment and social support in patients with non-insulin dependent diabetes mellitus. J Diabet Complications 9:81–86, 1995.
8.
Alogna M: Perception of severity of disease and health locus of control in compliant and noncompliant diabetics patients. Diabetes Care 3:533–534, 1980.
9.
Boom CAB, Hart LK: The relationship between the health belief model and compliance of persons with diabetes mellitus. Diabetes Care 3:594–598, 1980.
10. Brownlee-Duffeck M, Peterson L, Simmonds JF, Goldstein D, Kilo C: The role of health beliefs in the regimen adherence and metabolic control of adolescents an adults with diabetes mellitus. Diabetes Care 55:139– 144, 1987. 11. Harris R, Linn MW, Skyler JS, Sandifer R: Development of the diabetes health belief scale. Diabetes Educ 13:292– 297, 1987. 12. American Diabetes Association: Nutritional recommendations and principles for individuals with diabetes mellitus. Diabetes Care 15:21–28, 1992. 13. American Diabetes Association: Office guide to diagnosis and classification of diabetes mellitus and other categories of glucose intolerance. Diabetes Care 15(suppl 2):4–6, 1992. 14. Huerta MR, Mena A, Malacara JM, Diaz de Leo´n J: Symptoms at perimenopausal period: Its association with attitudes toward sexuality, life-style, family function and FSH levels. Psychoneuroendocrinology 20:135– 148, 1995. 15. Nunnally JC, Bernstein IH: Psychometric Theory, 3rd ed. New York: McGraw-Hill, 1994. 16. Herna´ndez M, Chavez A, Bourges H: Valor Nutritivo de los Alimentos Mexicanos, 10th ed. Me´xico: Publicaciones de la divisio´n de nutricio´n, Instituto Nacional de la Nutricio´n, Me´xico, 1987, pp. 1–34. 17. Krasvitz RL, Havs RD, Sherbourne CD, MiMatteo R, Rogers WH, Orway L, Greenfield S: Recall of recommendations and adherence to advice among patients with chronic medical conditions. Arch Intern Med 153:1869–1878, 1993. 18. Rollins M: Denial? (not me!). Diabetes Forecast 7:67–68, 1992. 19. Doyle BJ, Warw JE: Physician conduct and other factors that effect consumer satisfaction with medical care. J Med Educ 52:793–801, 1977. 20. Glasgow RE, Toobert DJ: Social environment and regimen adherence among type II diabetic patients. Diabetes Care 11:377–386, 1988. 21. Janz NK, Becker MH: The health belief model: A decade later. Health Educ Q 11:1–47, 1984.
The Belief in Conventional Medicine and Adherence to Treatment in Non–InsulinDependent Diabetes Mellitus Patients Maria Eugenia Garay-Sevilla Juan Manuel Malacara H. Felipe Gonza´lez-Parada Lourdes Jorda´n-Gine´s
ABSTRACT We investigated the role of belief in conventional medicine, the type of medical care, and familiar and socioeconomic factors on the adherence to treatment in non–insulin-dependent diabetes mellitus (NIDDM) patients. In a cross-sectional design, we selected 156 patients from two institutions, who agreed to fill out a questionnaire, which included general data, socioeconomic level, somatometric data, type of medical care, complications, if they had friends and relatives with diabetes, the family function, and a score on the belief in conventional medicine. Factors associated with adherence to diet and medication were analyzed. Patients had a mean age of 55.6 years and 8.9 years since diagnosis. A total of 51.3% of them were not covered by social security, and 62.8% received attention by a general physician.
Patients under the care of a specialist had better adherence to diet and medication, and better belief in conventional medicine. The principal factor associated with adherence to medication and diet was the belief in conventional medicine (p ⬍ 0.001 in both). Adherence to diet was also associated with the socioeconomic level (p ⫽ 0.001) and years since diagnosis (p ⫽ 0.004). Adherence to medication was also associated with schooling (p ⫽ 0.001). We concluded that belief in conventional medicine is strongly associated with adherence to treatment and other factors such as schooling, socioeconomic level, and medical care under a specialist; adherence to diet was better in patients with more years since diagnosis. ( Journal of Diabetes and Its Complications 12; 5: 239–245, 1998.) 1998 Elsevier Science Inc.
INTRODUCTION
the therapeutic plan. Nonadherence is the most serious problem facing medical practice, as only a third of the patients properly follow medical indications. Non–insulin-dependent diabetes mellitus (NIDDM) has a high prevalence and its control requires a careful compliance to medical indications, which demands a permanent behavioral change to comply with diet and medication and to assure a continued monitoring of metabolic control. The reported frequency of nonadherence to diet varies from 35% to 75%.2 Interestingly, patients attribute their difficulties to follow medical indications to environmental and economical factors,
A
successful ambulatory treatment for patients with chronic diseases depends on various factors. Becker1 identified four of them: Recognition of il health, the diagnosis of illness, planning of treatment, and the patient’s adherence to
Instituto de Investigaciones Me´dicas, Universidad de Guanajuato, Leon, Mexico Reprint requests to be sent to: Dr. Juan Manuel Malacara H., Instituto de Investigaciones Me´dicas, Universidad de Guanajuato, 20 de Enero 929, Apdo, Postal 874, 37320 Leo´n Gto, Mexico. Journal of Diabetes and Its Complications 1998; 12:239–245 1998 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
1056-8727/98/$19.00 PII S1056-8727(97)00075-5
240
GARAY-SEVILLA ET AL.
as well as to personal problems with the spouse, family, and health professionals.3 In contrast, physicians ascribe nonadherence to personal factors.4 The interactions within the family may have an effect on adherence to treatment, a better adherence is associated with non aggressive, genuine and frank discussions about diabetes.5,6 We have found that adherence is related to some aspects of the family function, such as control of behavior,7 these family factors may interact with the social support they receive. An appropriate compliance to treatment for a lifelong disease requires that the model for health service is fully accepted and is not in conflict with the cultural background of the patient. The patients’ attitude toward disease has been studied by means of the health belief model,8–11 in order to understand and predict adherence to treatment. It has been proposed that adherence to treatment of diabetes increases with perceived susceptibility to disease, perceived severity, perception of benefits, barriers or costs, and cues to motivate changes of behavior. Compliance to diet is difficult to archive because food is an important cultural feature, and traditional food in most cultural groups does not comply with the American Diabetes Association (ADA) recommendations.12 Foods are also a common subject for healthrelated myths. Conventional medicine may be considered by diverse indigenous cultures as an alien institution. Non-scientific types of medicine are still prevalent in most cultures. Modern medicine, developed in industrialized countries, includes facilities for medical attention and diagnostic and treatment procedures that may be in conflict with traditional values of native societies. Based on these considerations, we propose that belief in modern medicine, may play a role in compliance to treatment for diabetes. In this study, we examined the interaction of the belief in medicine with compliance to treatment, the family support, and other factors in patients with NIDDM. METHODS Study Design. We studied 156 patients (43 men and 113 women) with NIDDM, according to the diagnostic criteria of the American Diabetes Association,13 in a cross-sectional design. Two groups of patients were examined: With and without social security coverage. The latter group included 80 patients who attended the Instituto de Investigaciones Me´dicas requiring free medical attention and laboratory test for control of diabetes (no free medication). The group with social security included 76 patients recruited from a Social Security Hospital in Leo´n, Mexico, where they received free medical attention. The patients from this group were invited during the weekly diabetes club meeting, to have the purpose of the study explained to them. The
J Diab Comp 1998; 12:239–245
acceptance rate was close to 30%. The socioeconomic status was middle and low in both groups. Inclusion Criteria. For inclusion in the study, we considered non-hospitalized NIDDM patients without pregnancy, intercurrent diseases, severe metabolic alterations, or complications that required additional treatment. The patients showed adequate physical and mental capacity to answer questions about their disease. Data Collection. Patients answered a questionnaire by direct questioning by an investigator including the following items: General data: Age, years since diagnosis of diabetes, schooling (in years), occupation, monthly family income (scores from 0 to 4 with 0 equal to less than Pesos$100.00 and 4 more than Pesos$5,000.00/month), and age of the spouse. The socioeconomic level was obtained with a score combining schooling and family income. The quartiles for each value were added, and the final score of socioeconomic level ranged from 1 to 7. Somatometric variables: Weight (kg), height (cm) and body-mass index (BMI) were obtained. Type of medical care: The type of medical care considered two aspects: coverage by social security, and the physician, general practitioner or specialist (internist or endocrinologist). Complications: We collected antecedents of myocardial infarction, stroke, angor pectoris, cataract, amaurosis, diabetic feet, and amputations. Family function: Family function was assessed by means of a questionnaire used in previous studies7,14 inquiring about communication within the family, and adding the answers to four questions scored 1–4. The score range was 4 to 16. Belief in conventional medicine: This was assessed with an instrument designed by us. The content validity of the test was approached including 15 topics on the patient’s most frequently mentioned objections to accepting medicine (or reasons to seek non-medical treatments), for health problems related to chronic diseases, maternal–child health, vaccination, and infectious diseases. The construct validity15 was tested in a pilot study on twenty NIDDM patients, analyzing intrinsic consistency with a correlation matrix. For the final instrument were considered the six questions with a higher intrinsic correlation. The questions included were (1) the frequency in which they sought medical care when they felt ill; (2) the extent of confidence on their physician’s proficiency; (3) their perception of the time required by medical care; (4) their perception on the cost of medical care; (5) their use of medical attention for prevention; (6) their perception of the effectiveness of herbs and vegetables to cure diabetes. Answers to each
J Diab Comp 1998; 12:239–245
question were scored with a Likert scale ranging from 1 to 4. Question six had a negative polarity (higher scores for answer of “never”). The scores for each question were added for a final score ranging from 6 to 24. Relatives and friends with diabetes: The questionnaire collected patient awareness of close relatives or friends with uncomplicated or complicated diabetes, or recently deceased as a result of complications of diabetes (answers yes/no). Adherence to treatment: Adherence to medical treatment was studied in two aspects: adherence to diet and to medication. These aspects were evaluated with the questionnaire used in our previous work.7 Adherence to diet was assessed with seven questions scored 1 to 4, answers were added to obtain a score range of 7–28. The questionnaire on adherence to medication had three questions, and the score was added, with a result ranging from 3 to 12. Higher scores indicated better adherence to treatment. Metabolic control: Blood glucose was measured with a glucose oxidase method (GOD-PAD, Lakeside) and glycated hemoglobin with an ion exchange chromatography method (Sigma Chemical, St. Louis, MO, USA) (normal values, 5.5%–8.5%). Food intake: Food intake was assessed in a semi-quantitative manner asking the type and amount of food recently consumed. The category of was selected from a list of the 29 foods most frequent consumed in the region. We calculated daily intake of protein, fat (in grams) and calories using the tables of the Instituto Nacional de la Nutricio´n (Mexico City).16 We registered the type of dietary prescription: oral or written, quantitative or qualitative. Statistical Analysis. The characteristics of the total group of patients were analyzed with descriptive parametric statistics. Differences between groups under the care of a general physician and a specialist, and between patients with or without social security were analyzed with an unpaired Student’s t test. The determinants of metabolic control were tested with a multiple-regression analysis, taking blood glucose and glycated hemoglobin as dependent variables, and age, years since diagnosis, adherence to diet, adherence to medication, and intake of calories, fat, and proteins as candidate regressors. The determinants of adherence to diet and medication were tested with multiple-regression analysis taking as candidate regressors: age, years since diagnosis, schooling, age of the spouse, socioeconomic level, medical service, type of physician, friends and relatives with diabetes or deceased as a result of diabetes, communication in the family, BMI, and belief in conventional medicine.
ADHERENCE TO TREATMENT IN NIDDM
241
RESULTS Characteristics of the Patients. The characteristics of the patients are shown in Table 1. The means for age and years since diagnosis were 55.6, and 8.9, respectively. Schooling was low (mean of 3.7 years). Most patients had middle or low socioeconomic level (1–4 on the scale of 1–7). They were mainly obese (mean BMI was 28). In relation to medical care, most patients were under the attention of a general physician. Among complications studied, amaurosis was found in 12.2% and cataract in 8.3% of the cases. Family Function. In regard to family function, communication was mainly clear and direct, and no difference was found among the groups. This factor was not collected in 25 widowers, 19 who were not married, three who did not have children, and four who did not respond. A total of 75% of the patients had at least a relative with diabetes, and 61% had a diabetic friend. These factors did not have an influence on adherence to treatment. Medical Care. The analysis of the characteristics of patients according to their physician’s type is shown in Table 2. Patients under a general physician’s care had fewer years since diagnosis, had lower scores of adherence to diet and medication, and lower scores of belief of conventional medicine. No differences between groups were found for metabolic control, food intake, or rates of complications. The characteristics for groups with and without social security coverage are shown in Table 3. Patients with social security were older and with more years since diagnosis, and had lower glucose levels than those without social security. Adherence to treatment was similar in both groups, and belief in conventional medicine had marginally higher scores in the social security group (p ⫽ 0.04). Adherence to Treatment. Adherence to diet had a mean of 20.1 (95% CI, 20.8–19.4) from a range of 7–28, and adherence to medication a mean of 10.7 (95% CI, 11.0–10.4) from a range of 3–12. A total of 34.6% of the patients did not follow their prescribed diet, and 6.4% said they followed it strictly. The type of diet prescription was quantitative and written for 32%, qualitative and written for 5.8%, quantitative but not written for 21%, and qualitative and not written for 3.8% of the patients. Metabolic Control. Metabolic control was not satisfactory in most patients: mean glucose level was 192.1 mg/dL and mean HbA1c was 12.9%. In the regression analysis for determinants of metabolic control, adherence to medications was determinant for both glucose levels (r ⫽ ⫺0.182, p ⫽ 0.023) and glycated hemoglobin (r ⫽ ⫺0.174, p ⫽ 0.029).
242
GARAY-SEVILLA ET AL.
J Diab Comp 1998; 12:239–245
TABLE 1. CHARACTERISTICS OF THE PATIENTS Variable Socioeconomic Level 1–2 (low) 3–4 (middle) 5–7 (high) Medical care With social security Without social security General physician Specialist Complications Myocardial infarction Cataract Amaurosis Diabetic foot Amputations Family function: Communication within the family Clear and direct Clear and indirect Masked and direct Unmasked and indirect Not applicable Relatives and friends with diabetes Relatives with diabetes Relatives with complicated diabetes Relatives deceased as a result of diabetes Friends with diabetes Friends with complicated diabetes Friends deceased as a result of diabetes Metabolic control Glucose (mg/dL) HbAlc % Other characteristics Age (years) Age of the spouse (years) Time since diagnosis (years) Schooling (years) BMI Protein intake Fat intake Calorie intake
n
%
55 66 30
35.3 42.3 19.1
76 80 98 55
48.7 51.3 62.8 35.2
2 13 19 8 3
1.3 8.3 12.2 5.3 1.9
81 17 6 1 51
51.9 10.9 3.8 0.6 32.7
117 50 65 95 37 41
75.0 32.1 41.7 60.9 23.7 26.3
Mean
95% CI
192.1 12.9
181.2–202.9 12.4–13.3
55.6 52.2 8.9 3.7 28.0 50.1 40.7 1230
53.8–57.5 53.5–56.9 7.8–10.1 3.2–4.2 28.9–27.2 47.1–53.5 37.8–43.7 1157–1304
CI, confidence interval; HbAic, glycated hemoglobin; BMI, bodymass index.
Factors Associated with Adherence to Treatment. The factors associated with adherence to treatment as analyzed by means of the multiple regression procedure are shown in Table 4. Adherence to diet and medication were highly associated with belief in conventional medicine (Figure 1); additionally, adherence to diet was associated negatively with socioeconomic level and positively with years since diagnosis; and adherence to medication correlated negatively with schooling. DISCUSSION Conventional medicine has exhibited an explosive evolution, resulting in the scientific and technical ad-
vances, developed in industrialized countries. However, modern medical concepts and techniques may be in conflict with cultural values from traditional societies. This may hamper the acceptance of medical treatments in marginated communities who preserve traditional values. Societies with ancient traditions still have mythical concepts about health and disease. They seek magic cures for their ailments, and are reluctant to accept long-lasting medical treatments, mainly if a prompt benefit is not experienced. This is also true when medication have components and labels revealing an alien origin. Diet is an important element in the treatment
J Diab Comp 1998; 12:239–245
ADHERENCE TO TREATMENT IN NIDDM
243
TABLE 2. CHARACTERISTICS OF PATIENTS ACCORDING TO TYPE PHYSICIAN’S CARE
Age (years) Years since diagnosis Weight (kg) Height (cm) BMI Socioeconomic level Adherence to diet Adherence to medication Belief in conventional medicine Glucose (mg/dL) Glycated hemoglobin (%) Intake of calories Intake of fat (g) Intake of protein (g)
General Mean (SD)
Specialist Mean (SD)
t
p
54.3 (11.6) 7.9 (6.5) 67.3 (12.6) 155 (0.09) 28.0 (5.0) 3.1 (1.5) 19.4 (4.4) 10.5 (1.7) 17.9 (3.8) 194 (71) 13.2 (3.1) 1229 (422) 39.3 (16) 48.9 (17)
58.1 (11.5) 11.2 (8.3) 69.9 (13.0) 158 (0.09) 28.0 (5.4) 3.3 (1.4) 21.5 (4.1) 11.1 (1.8) 19.8 (3.4) 190 (65) 12.5 (2.7) 1211 (521) 42.2 (23) 51.0 (20)
⫺1.92 ⫺2.79 ⫺1.16 ⫺1.96 0.02 ⫺0.82 ⫺3.00 ⫺2.38 ⫺3.05 0.40 1.50 0.22 ⫺0.96 ⫺0.68
0.056 0.005 0.247 0.052 0.981 0.411 0.003 0.019 0.003 0.689 0.136 0.82 0.340 0.50
n (%) 4 (7.3) 3 (5.4) 4 (7.3)
2 0.16 3.82 0.724
p 0.684 0.051 0.395
n (%) 9 (9.2) 16 (17.4) 4 (4.1)
Cataract Amaurosis Diabetic foot BMI, body-mass index.
of NIDDM, and accepted diet guidelines are in conflict with alimentary traditions of several societies. Food is a factor of cultural identification for every ethnic group. In Mexico, traditional food has a high content of carbohydrate and saturated fat, and a low fiber content. These considerations explain difficulties for the acceptance of modern medicine. In this study, we developed a simple questionnaire intended to evaluate belief
in conventional medicine. The association of the scores obtained with compliance to medication and diet are in agreement with a predictive value of this instrument15 for use in subsequent studies. The findings from this study should be considered to design adequate strategies for a more culture-based approach for behavior modification. An important strategy should be a careful education on the meaning and practical appli-
TABLE 3. CHARACTERISTICS OF PATIENTS ACCORDING TO SOCIAL SECURITY
Age (years) Years since diagnosis Weight (kg) Height (cm) BMI Socioeconomic level Adherence to diet Adherence to medication Belief in conventional medicine Glucose (mg/dL) Glycated hemoglobin (%) Intake of calories Intake of fat (g) Intake of protein (g) Cataract Amaurosis Diabetic foot BMI, body-mass index.
Social Security Mean (SD)
No Social Security Mean (SD)
t
p
59.3 (9.8) 10.2 (7.7) 68.5 (13.1) 157 (0.1) 27.7 (4.8) 3.3 (1.5) 20.3 (4.2) 10.8 (2.0) 19.2 (3.4) 178 (60) 12.5 (2.7) 1261 (490) 42.8 (19.8) 52.3 (18.8)
52.1 (12.1) 7.8 (6.8) 68.5 (13.2) 156 (0.1) 28.3 (5.4) 3.0(1.4) 20.0 (4.5) 10.6 (1.7) 18.0 (4.0) 206 (73) 13.3 (3.2) 1200 (439) 38.8 (17.5) 48.0 (18.5)
4.06 2.00 ⫺0.13 1.00 ⫺0.73 1.17 0.41 0.59 2.05 ⫺2.61 ⫺1.60 0.82 1.36 1.42
⬍0.001 0.046 0.989 0.320 0.466 0.244 0.682 0.555 0.042 0.010 0.102 0.414 0.175 0.158
n (%) 6 (7.9) 9 (11.8) 2 (2.6)
n (%) 7 (8.7) 10 (12.5) 6 (7.5)
2 0.04 0.01 1.90
p 0.847 0.900 0.160
244
GARAY-SEVILLA ET AL.
J Diab Comp 1998; 12:239–245
TABLE 4. REGRESSION ANALYSIS TAKING AS DEPENDENT VARIABLE ADHERENCE TO TREATMENT 
r
p
0.355 ⫺0.234 0.206
0.375 ⫺0.258 0.228
⬍0.001 0.001 0.004
0.458 ⫺0.223
0.473 ⫺0.253
⬍0.001 0.001
Regressor Adherence to diet Belief in conventional medicine Socioeconomic level Years since diagnosis Adherence to medication Belief in conventional medicine Schooling
cations of scientific medicine, a factor largely overlooked in the education programs for diabetic patients. In our group of study, a large proportion of the patients had partial or complete noncompliance with medical indications, and adherence was strongly associated with years since diagnosis of diabetes. This agrees with our previous report,7 but other investigators, found only a weak association between both factors.17 We interpreted this finding to mean that in the
FIGURE 1 Association of adherence to diet (upper panel) and adherence to medication (lower panel) with belief in medicine, in patients with non–insulin-dependent diabetes mellitus. The p values for both cases were ⬍ 0.001.
progress of time, diminishes the patient’s denial of disease, frequent at early stages.18 After the onset of complications, the patient accepts the disease, and may use it as an adaptive contrivance to obtain more care from the family. The disease may frequently become a central subject modifying the dynamics of family function. It is possible that the association between adherence to diet and years since diagnosis, also results from the survival bias that appears in cross-sectional studies. Better compliants should have better survival. It was surprising to find that patients with lower schooling and socioeconomic level comply better with medical treatment. Before interpretation, one must consider that the group of study had a very low schooling (mean of 3.7 years). Therefore, conclusions should be limited to a narrow socioeconomic range. Knowledge of diabetes, an important factor for adherence to treatment,7 is not acquired in elementary school. This finding emphasizes the importance of information on diabetes given to groups of patients. Another source for information may be the mass communication media, that may facilitate compliance to treatment, by means of a complete information of the devastating results from complications. Patients under the specialist’s care had higher scores of belief in conventional medicine and adherence both to diet and to medication as compared with those under the care of a general physician. These results are in agreement with the findings from a study in hypertensive patients17 reporting better adherence in patients treated by a cardiologist than by a general physician. The explanation of these results is that patients who better accept conventional medicine look for the advantages of a specialist’s care, and are more prone to better compliance to treatment. An important aspect of social support for diabetic patients is a proper interaction with the health-care providers and with the health-care system.19,20 We found only a marginally higher belief in conventional medicine in the group with social security. Support received within the family is a most important aspect of social support. Communication is relevant to enhance family support. The homogeneity of
J Diab Comp 1998; 12:239–245
this characteristic in the group of study did not permit us to examine the influence of communication as part of the social support for better adherence to treatment, as in our previous study.7 Several investigations have shown the impact of morbidity or mortality of diabetic relatives and friends on a patient’s compliance to treatment, permitting a better perception of the dreadful consequences of diabetes complications.8,21 Yet, our results did not show a significant association of these factors. Differences can be explained by different methods and criteria, because we did not use the health belief model.21 This model proposed that the acceptance of medical advice depend on the perception of vulnerability to illness, including acceptance of the diagnosis, and perceived benefits. In NIDDM patients, both may be low, in particular, perception of the benefits of appropriate metabolic control. We concluded that the adherence to treatment in NIDDM patients is strongly associated with belief in conventional medicine and negatively associated to schooling and socioeconomic level. Adherence to diet was associated with years since diagnosis. The belief in medicine was higher in patients under the care of a specialist, and was not clearly associated with social security coverage. ACKNOWLEDGMENTS This work was supported in part by grant 3417-M from CONACYT.
REFERENCES 1.
Becker MH: Patient adherence to prescribed therapies. Med Care 23:539–555, 1985.
2.
Kurtz SM: Adherence to diabetes regimens: Empirical status and clinical applications. Diabetes Educ 16:50–56, 1990.
3.
Ary DV, Toobert D, Wilson W, Glasgow RE: Patient perspectives on factors contributing to nonadherence to diabetes regimens. Diabetes Care 9:168–172, 1986.
4.
House WC, Pendelton L, Parker L: Patients versus physicians attributions of reasons for diabetic patients’ non compliance with diet. Diabetes Care 9:434, 1986.
5.
6.
Anderson BJ, Miller IP, Auslander WF, Santiago JV: Family characteristics of diabetic adolescents: Relationship to metabolic control. Diabetes Care 4:586–594, 1981. Carney RM, Schechfer K: Familial eating patterns and metabolic control in juvenile onset diabetes. Diabetes 3(suppl 2):13A, 1982.
ADHERENCE TO TREATMENT IN NIDDM
245
7.
Garay-Sevilla ME, Nava LE, Malacara JM, Huerta MR, Diaz de Leo´n J, Mena A, Fajardo ME: Adherence to treatment and social support in patients with non-insulin dependent diabetes mellitus. J Diabet Complications 9:81–86, 1995.
8.
Alogna M: Perception of severity of disease and health locus of control in compliant and noncompliant diabetics patients. Diabetes Care 3:533–534, 1980.
9.
Boom CAB, Hart LK: The relationship between the health belief model and compliance of persons with diabetes mellitus. Diabetes Care 3:594–598, 1980.
10. Brownlee-Duffeck M, Peterson L, Simmonds JF, Goldstein D, Kilo C: The role of health beliefs in the regimen adherence and metabolic control of adolescents an adults with diabetes mellitus. Diabetes Care 55:139– 144, 1987. 11. Harris R, Linn MW, Skyler JS, Sandifer R: Development of the diabetes health belief scale. Diabetes Educ 13:292– 297, 1987. 12. American Diabetes Association: Nutritional recommendations and principles for individuals with diabetes mellitus. Diabetes Care 15:21–28, 1992. 13. American Diabetes Association: Office guide to diagnosis and classification of diabetes mellitus and other categories of glucose intolerance. Diabetes Care 15(suppl 2):4–6, 1992. 14. Huerta MR, Mena A, Malacara JM, Diaz de Leo´n J: Symptoms at perimenopausal period: Its association with attitudes toward sexuality, life-style, family function and FSH levels. Psychoneuroendocrinology 20:135– 148, 1995. 15. Nunnally JC, Bernstein IH: Psychometric Theory, 3rd ed. New York: McGraw-Hill, 1994. 16. Herna´ndez M, Chavez A, Bourges H: Valor Nutritivo de los Alimentos Mexicanos, 10th ed. Me´xico: Publicaciones de la divisio´n de nutricio´n, Instituto Nacional de la Nutricio´n, Me´xico, 1987, pp. 1–34. 17. Krasvitz RL, Havs RD, Sherbourne CD, MiMatteo R, Rogers WH, Orway L, Greenfield S: Recall of recommendations and adherence to advice among patients with chronic medical conditions. Arch Intern Med 153:1869–1878, 1993. 18. Rollins M: Denial? (not me!). Diabetes Forecast 7:67–68, 1992. 19. Doyle BJ, Warw JE: Physician conduct and other factors that effect consumer satisfaction with medical care. J Med Educ 52:793–801, 1977. 20. Glasgow RE, Toobert DJ: Social environment and regimen adherence among type II diabetic patients. Diabetes Care 11:377–386, 1988. 21. Janz NK, Becker MH: The health belief model: A decade later. Health Educ Q 11:1–47, 1984.