Volume 23
Number 4
2002

[Table of Contents]

Perceptions of disease severity and barriers to self-care predict glycemic control in aboriginal persons with type 2 diabetes mellitus

Mark Daniel and Lynne C Messer

Abstract

The Health Belief Model (HBM) was evaluated for secondary prevention of type 2 diabetes mellitus in an Aboriginal population in British Columbia. Glycemic markers (glycated hemoglobin [HbA_1c]), insulin and post-load glucose), diabetes health beliefs (susceptibility, severity, benefits and barriers), knowledge and behaviour were measured for 16 men and 18 women with diabetes (age [SD] = 57.7 [11.6]). Eighteen months later, HbA_1c and behaviour were measured for all participants, and health beliefs obtained for 17 of them. Perceived severity and perceived barriers were related to glycemic status at baseline and follow up, and predicted reduction in HbA_1c (ß[SE] ³ |0.40| [0.18], p < 0.05). The results support a therapeutic emphasis on belief in the severity of diabetes complications, and the complementary belief that barriers to therapeutic behaviour can be overcome in efforts to support Aboriginal persons with diabetes to manage their disease. The empirical utility of the HBM in glycemic control was partially upheld.

Key words: Aboriginal, North American; diabetes mellitus, type 2; glycemic control; health beliefs

Introduction

Diabetes mellitus, mostly type 2 diabetes mellitus, is a serious public health problem in the Canadian Aboriginal population.¹ Among First Nations and Inuit peoples, diabetes prevalence is 8% for men and 13% for women, corresponding to age-adjusted rates 3.6 and 5.3 times greater for Aboriginal men and women, respectively, than for their counterparts in the general Canadian population.²

Community-level strategies against diabetes attempt to screen for and prevent development of the disease amongst susceptible persons and the community in general.³ Prevention initiatives also try to help persons with diabetes to control hyperglycemia and limit the risk of macrovascular and microvascular complications. Behavioural therapy is informed by an understanding of relations between effective treatment and health beliefs associated with having diabetes.⁴ For persons with type 1 diabetes, treatment compliance is positively associated with perceived benefit, emotional stability and supportive structure, and negatively related to perceived barriers and negative social environment.⁵

Effective control of type 2 diabetes requires behavioural compliance with diabetic regimens, which is difficult to predict and challenging to influence. The Health Belief Model (HBM)⁶ was initially developed to understand why people did not participate in disease detection programs^7,8 but has since been used to explicate factors underlying patient compliance with treatment regimens for diseases, including diabetes.⁹ In its application to diabetes, the theoretical utility of the HBM lies in its potential to inform more effective interventions for treatment adherence and improved glycemic control.¹⁰ Founded on the value-expectancy theories of social psychology, the HBM posits that individuals will be more likely to take healthful behavioural action if they desire to stay healthy and believe such action will effectively protect their health.^9,11 The HBM posits further that positive behavioural action depends on convictions about the severity of the disease threat and the belief that barriers to the execution of therapeutic behaviour can be surmounted.¹²

The utility of the HBM in accounting for health behaviour has been examined in studies ranging from acceptance of preventive health recommendations^7,8 to adherence to treatments for acute and chronic illness.^13,14 While positive health beliefs are linked to patient compliance for chronic diseases, including diabetes, barrier perceptions are most strongly, and severity perceptions are most weakly, related to positive health behaviour.^14-19 A meta-analysis of the HBM's consistency of validation found the evidence to be lacking, however, in its overall utility for predicting behaviour change.²⁰ Conceding that some behavioural research upheld a simplistic cognitive emphasis on changing attitudes or beliefs in isolation from the contexts in which they occur,²¹ it can also be challenging to assess behaviours with sufficient sensitivity to link them to specific health beliefs,²² especially in ethnically distinct and disadvantaged populations.

Studies applying the HBM to diabetes have relied largely on self-reported data in focusing on relationships between health beliefs and behavioural compliance with prescribed treatment regimens. Early research on diabetes tended to target beliefs in relation to behaviour alone, but a growing pool of studies has targeted relationships between health beliefs and behaviour as well as glycemic control in the same individuals.^10,23 Such studies are important in assessing the utility of the HBM in relating health beliefs to indicators of glycemic control, yet the evidence would suggest the model explains behavioural compliance better than glycemic control.^10,23,24 Few studies, however, have examined the capacity of the HBM in longitudinal prediction of glycemic control while allowing for the potential influence of behaviour on such relations.²⁵ We undertook such an analysis in a sample of Aboriginal Canadians from a population at high risk for diabetes and its complications, and this is the report of the results. Research on health beliefs and glycemic control in Aboriginal people has not yet been published. The hypotheses tested were that in a cohort of individuals with diabetes: (a) health beliefs would be related to glycemic outcomes at baseline screening and a followup survey 18 months later; and (b) health beliefs would predict improved glycemic control at follow up.

Methods

Population and setting

Data analyzed in this report are from a diabetes screening program implemented among on-reserve “registered” Indians in British Columbia's rural Okanagan region. Persons indigenous to this area are of the Interior Salishan linguistic group and the Plateau area culture. Registered Indians in this area are of limited educational attainment and low socio-economic status relative to the general population.²⁶ Of adults age 15–49, 50% have completed secondary school, and of adults age 50-64, 40% have completed grades 1–8. Of the total adult population 44% are employed, 16% are unemployed, and 40% are not involved in the labour force.

Participant selection and measurement protocol

Participants were volunteers for a community-based diabetes diagnostic and risk factor screening initiative. Pregnant women and minors under 18 years of age were excluded from screening. The study received ethical approval from the University of British Columbia Behavioural Sciences Screening Committee, and all participants provided their informed consent. Of the eligible adult population, the study participation rate was 57% (n = 202). The main reason for non-participation was lack of interest. Male gender and age less than 30 years were the primary correlates of non-participation. Tests were done in meeting halls between 7:30 a.m. and 12:00 noon. World Health Organisation criteria²⁷ were applied to classify individuals as normoglycemic (n = 155), having impaired glucose tolerance (IGT) (n = 10), or having diabetes (n = 37).²⁸

This paper reports on relations among persons with diabetes only, as established at the study baseline. Of 37 individuals identified as having diabetes, 34 agreed to provide data on health beliefs. Seventeen of these persons were verified as having existing diabetes (self-reported, median duration of 3.2 years), with the rest being newly diagnosed cases of diabetes. All persons with diabetes were referred for education and management to local medical practitioners and to a diabetes day program at a regional hospital an hour's drive away. Eighteen months after the baseline survey, HbA_1c was obtained for all 34 individuals and health beliefs obtained for 17 of them. Persons for whom follow-up data were available all complied in keeping referrals to medical practitioners and biannual visits to the diabetes day program.

Blood samples and analytic methodology

For screening purposes, participants gave a venous sample after fasting overnight. Concentrations were determined from these samples for glycated hemoglobin A_1c (HbA_1c) and insulin. Post-load glucose levels were determined from blood drawn two hours after consumption of a 75-g glucose load. HbA_1c, insulin and two-hour glucose concentrations were determined at baseline, but HbA_1c alone was re-measured 18 months later.

Whole blood specimens collected in ethylene-diamine-tetra-acetate (EDTA) anticoagulant were used to determine HbA_1c. Analyses for insulin and two-hour glucose were performed on serum specimens. Specimens were transported at 4° C for analysis at a regional laboratory on the day of collection. Insulin concentration was measured using microparticle enzyme immunoassay kits. Percent HbA_1c was determined using ion capture assay kits. Two-hour blood glucose concentration was assessed using enzymatically linked assay kits. Intra- and inter-assay coefficients of variation were, respectively: insulin, 3.8% and 4.2%; HbA_1c, 4.4% and 4.6%; and two-hr glucose, 1.2% and 1.8%.

Anthropometric measurements

Participants wore light clothing, with footwear removed. Weight and height were assessed using a beam balance and stadiometer. Body mass index (BMI) (weight [kg] / height [m²]) was calculated.

Diabetes knowledge and behaviours

Knowledge and behaviours were measured at both baseline and follow up. Knowledge of diabetes was assessed as the number of correct responses to 13 questions on diabetes risk factors, symptoms and complications. The two-week test-retest stability for responses was high (r = 0.89). Physical activity was assessed using the Pima Indian physical activity scale²⁹ to obtain metabolic equivalent [MET] hours of occupational and leisure-time physical activity for the month preceding measurement. For Pima Indians, correlations for test-retest stability (over two weeks) are 0.62–0.96. Dietary behaviour was determined using self-completed records for food and drink consumed over three consecutive days (including one weekend day). The reliability of three-day food records is well established.³⁰ Follow up interviews resolved imprecise entries. Nutrient intakes were computed using a nutrient database/software system for Canadian nutrition surveys.³¹ Energy intake (in kilocalories) and the relative quantity of protein, lipid, and carbohydrate consumed (g/100 kcal) were determined.

Diabetes health beliefs

Beliefs about diabetes were assessed at six weeks after baseline testing and at follow up using the 16-item Health Belief Model Diabetes Scale^32-35 (Appendix). This instrument surveys four domains of perceived beliefs relevant to having diabetes: (a) susceptibility to diabetes complications; (b) severity of diabetes complications; (c) benefits of diabetes control; and (d) barriers to executing therapeutic behaviours. Responses were scored on a five-point Likert scale. The construct validity of the scale has been established, and concurrent and predictive validity estimates uphold its criterion validity.^33,34 The instrument was pre-tested with 10 persons with IGT, for whom test-retest stability, r, for subscales after three weeks ranged from 0.67–0.73. Cronbach's a for subscales, based on the 34 persons with diabetes, ranged from 0.69–0.82. The questionnaire required 15 minutes to complete.

Statistical analysis

The hypotheses that diabetes beliefs about susceptibility, severity, benefits and barriers are related to HbA_1c, insulin and two-hour glucose concentrations were tested using linear regression models with adjustments for age and gender. Separate models were analyzed for each of the four diabetes beliefs. The linearity of dependence relationships was established by evaluating plots of residuals. Standardized beta coefficients are reported. P-values less than 0.05 are considered to be statistically significant. STATA (6.0) software was used for analyses.

Diabetes knowledge, behaviour and BMI, at baseline and follow up, were included separately and together in preliminary analyses, to assess their effect together with diabetes beliefs on glycemic outcomes. Behaviours were not related to beliefs, and in no case was any association between beliefs and glycemic status influenced by behaviour, nor was any behaviour related to any glycemic measure in models including beliefs. For example, in assessing relations between perceived severity at baseline and 18-month change in HbA_1c, the inclusion in models of behaviours measured at follow up, with or without BMI, yielded results in keeping with models omitting behavioural variables altogether. In all such preliminary analyses the relationship between perceived severity and change in HbA_1c stayed significant, with the strength of the association (ß[SE]) ranging from 0.39 (0.18) (when specified with knowledge) to 0.50 (0.22) (when specified with carbohydrate intake). In such models, ß(SE) for BMI ranged from -0.25 (0.21) to -0.18 (0.18) (i.e., BMI was negatively related, but not significantly so, to healthful change in HbA_1c). Furthermore, in no case was any behaviour related to change in HbA_1c. For behaviours, ß(SE) ranged from 0.03 (0.21) for knowledge to 0.31 (0.24) for protein intake. Given a consistent lack of any measurable effect of behaviour, to conserve statistical power, final models did not include knowledge, BMI or any behaviour, and focused solely on beliefs and glycemic outcomes.

Results

Descriptive characteristics at baseline for the persons surveyed are given in Table 1. There were no significant differences between men and women for age, BMI, HbA_1c, insulin, or years of schooling. Diabetes beliefs did not differ between the genders (Table 2). Correlations between diabetes beliefs are given in Table 3. Knowledge of diabetes and behaviours did not differ by gender at baseline or follow up, and there were no significant changes in behaviours, nor HbA_1c or BMI, between baseline and follow up (Table 4). Knowledge and behaviours were not related to beliefs or glycemic measures.

After 18 months of following up the 34 individuals classified at baseline as having diabetes, 16 were managing the disease by diet alone, 16 were managing by diet and medication with oral hypoglycemic agents, and two were managing by diet and insulin therapy. Management strategy and existing versus newly diagnosed diabetes were not associated with HbA_1c at baseline or follow up. Diabetes beliefs at follow up for 17 persons who provided such data, were not related to treatment strategy or existing versus newly diagnosed diabetes. The distribution of gender and age did not differ between persons participating in both surveys and those participating at baseline only.

Baseline health beliefs and baseline glycemic measures

Diabetes health beliefs were not associated with HbA_1c concentration at baseline (Table 5). Insulin concentration was inversely associated with perceived severity (p = 0.042) and positively associated with perceived barriers (p = 0.05). Hence low insulin concentration was related to high perceived severity of complications and low perception of barriers to therapeutic behaviours. High perceived benefit to complying with a diabetic regimen was related to low two-hour glucose level (p = 0.012). Perceived susceptibility to complications was unrelated to any glycemic measure.

Baseline health beliefs and change in HbA_1c

Perceived severity of consequences of uncontrolled diabetes at baseline was positively related (Table 5) to positive change in HbA_1c concentration (p = 0.036) over the 18 months of follow up, and thus predicted reduction of HbA_1c. Perceived susceptibility, perceived benefits and perceived barriers at baseline were not significantly associated with change in HbA_1c.

Health beliefs at follow up and HbA_1c at follow up

HbA_1c concentration was inversely associated at follow up with perceived severity of complications (p=0.037), and positively related to perceived barriers (p=0.0004) to therapeutic behaviours (Table 5). Greater perceived severity and low perceived barriers were thus associated with healthful HbA_1c. Perceived susceptibility and perceived benefits were not related to HbA_1c concentration at follow up.

Follow up health beliefs and change in HbA_1c

Perceived severity of diabetes complications at follow up was positively associated with positive, or healthful, change in HbA_1c concentration between baseline and follow up (Table 5). Similarly, low perceived barriers to upholding therapeutic behaviours was associated with healthful reduction in HbA_1c between baseline and follow up. Perceived susceptibility and perceived benefits at follow up were not related to change in HbA_1c concentration.

Discussion

In this sample of Aboriginal people with type 2 diabetes, neither diabetes health beliefs nor HbA_1c concentration changed at the group level between baseline and a follow up survey conducted 18 months later. All participants were referred to and followed through in consulting local medical practitioners for treatment of diabetes, in addition to participating in a diabetes education program at a regional hospital. This report aimed not to evaluate treatment strategies or patient compliance with them, but to assess diabetes health beliefs at both points in time and to determine their relations to glycemic outcomes. Despite no relation at baseline between diabetes beliefs and HbA_1c, baseline perceptions of the severity of diabetes predicted reduced HbA_1c at the follow up survey. At follow up, high perceived severity of diabetes and low perceived barriers to therapeutic behaviours were related to healthful HbA_1c concentration as well as reductions in HbA_1c. Such results indicate that individual beliefs about barriers related to control and severity of complications are important factors influencing the ability of Aboriginal people with diabetes to achieve control of blood glucose.

Of further importance is the evident utility of extending the HBM from its original purpose to predict health-related behaviour from health beliefs, to directly predicting physiological outcomes. Restricting use of the HBM to planning and assessing behaviour change strategies in diabetes could short change health professionals and clients alike, as the model is of modest utility in this task.^10,23-25 It is noteworthy that controlling for behaviour at baseline and follow up in estimating relationships between diabetes beliefs and glycemic measures at each time, and controlling for behaviour at follow up in estimating relations between beliefs and change in glycemic control, was associated with either no change or minor attenuation only of model coefficients. This does not necessarily mean, however, that behaviour did not influence the positive results observed. Despite the lack of relationship found between any behaviour and belief or glycemic outcome, it is possible that behaviours, unlike HbA_1c and perceived beliefs, were not measured with sensitivity sufficient to implicate their mediation of the link between diabetes health beliefs and glycemic outcomes.

The consistent relation found between perceived severity and glycemic control in this study is an important finding, and is supported by earlier research indicating that perceived severity is the component of the HBM most strongly related to behavioural compliance.³⁶ Nevertheless, while a link between high perceived severity and healthful HbA_1c has been reported before,¹⁹ most studies indicate that high perceived severity is related to poor glycemic control.^10,37 Some of the discrepancy in these findings could reflect those studies linking poor glycemic control to high perceived severity being of a cross-sectional rather than longitudinal design, as well as their focus on persons with long established diabetes (five years or more). The duration of diabetes cases may influence such findings. In the present study, newly diagnosed cases were surveyed on diabetes beliefs six weeks after diagnosis, and the median duration of diabetes among established cases was 3.2 years at the time of the survey.

A relation between perceived severity and HbA_1c could be framed in terms of the “perceived threat” (of diabetes complications) reflecting the consequence of initial perceptions about severity, followed by perceptions of susceptibility.⁶ Thus, a heightened state of perceived severity could be required before perceived susceptibility might predict glycemic control. That susceptibility was not related to glycemic outcomes in this study suggests many persons surveyed were in the process of adjusting to diabetes. One might expect susceptibility to increase, and severity to decrease, given further experience with diabetes. Longitudinal research is necessary to clarify whether the inverse relation between perceived severity and HbA_1c reverses over time, and whether susceptibility mediates such a change. At this time, the role of perceived susceptibility in diabetes management is inconsistent.³⁶

The second component of the HBM explaining positive glycemic control in this analysis, consistent with other reports,^6,10,19,37 was perceived barriers at follow up. Baseline perceived barriers were positively related to insulin concentration. Underlying the HBM is an assumption that the value one places on health influences his or her health-seeking behaviour.^10,38 This premise presupposes freedom to pursue health-oriented behaviour. Aboriginal peoples face obstacles to improving health, including educational, economic and power disparities.² Conceding various challenges of daily living for Aboriginal populations, both subjective (perceived) and objective (structural) barriers to health-related activity must be addressed by clinical and community level initiatives to predispose, enable and reinforce health-related behaviour and reduce risk of developing diabetes and its complications.³

A combination of the perceived benefits and perceived barriers components of the HBM is believed to provide “direction” for motivated health actions.⁹ Thus, understanding the nature of the barriers faced by Aboriginal people may shed light on why the perceived benefits component of the scale did not predict glycemic outcomes. It could be difficult, perhaps, for Aboriginal people to be impressed by the potential benefits of glycemic control, given the more immediate hardships they face on a daily basis. That two-hour glucose concentration was associated with perceived benefits at baseline is supported by other research documenting an association between perceived benefits and behavioural compliance in self-management of diabetes.³⁷ The physiological response to a diagnostic glucose load given by two-hour glucose concentration is in large part a function of risk behaviours linked to the development of diabetes, embedded as they are in lifestyle.³

The hypotheses evaluated by this study were only partly upheld. The four HBM component beliefs were not consistently related to glycemic outcomes. Perceived severity and perceived barriers were the best predictors of glycemic status. Perceived benefits was related to insulin concentration but not glucose control per se (i.e., HbA_1c concentration), and perceived susceptibility was not related to any glycemic outcome. This modest empirical confirmation of the model is consistent with the balance of the literature on the capacity of the HBM in diabetes control and management.³⁶ While this in itself may be a reasonable conclusion, there are additional aspects of this application of the HBM that merit appraisal, including its applicability to the Aboriginal population surveyed. It may be that the predictive value of the HBM is limited for populations facing extensive socio-structural barriers to health action, where non-attitudinal factors inhibit those beliefs, or the effects of such beliefs, that in more advantaged populations might relate better to health outcomes. One might also question how much the Health Belief Model Diabetes Scale, in terms of perceptions of susceptibility to diabetes complications and benefits of effective control, captured salient beliefs held by the Aboriginal population surveyed.³⁹

Limitations

The primary limitation of this study is its small sample size and the consequence of this on statistical power. It seems unlikely, though, that low statistical power was a factor influencing the limited effect of perceived benefit and a lack of effect of perceived susceptibility on glycemic outcomes. Standard errors for the standardized beta coefficients for these two components of the Health Beliefs Model Diabetes scale are stable relative to those for other components of the model, and beta coefficients are close to zero. Low statistical power also precluded performing multivariate analyses including all four diabetes belief constructs in the same model at the same time (each construct was tested alone).

The small sample size also limits the generalizability of the study results to populations. As noted, however, empirical research on health beliefs and glycemic control in Aboriginal people with diabetes has not been published, and a longitudinal analysis, however small, is arguably a worthwhile contribution to the literature on diabetes in this disadvantaged and under-researched population.

A third matter is the cultural appropriateness of the Health Belief Model Diabetes Scale. The scale appears to have some utility for the population reported on here, as evidenced by the reliability of the instrument sub-scales. Wider use of the scale in other Aboriginal populations would benefit from further psychometric evaluation to see how accurately it captures the constructs it purports to tap. Scales assessing diabetes health beliefs may be population specific in their applicability.^38,40

Practice Implications

Recommended strategies for secondary prevention of type 2 diabetes include individualized, patient-centred strategies drawing on physicians, diabetes educators and other health personnel.⁴¹ High-risk populations may also benefit from community-based initiatives to support individual interventions to control diabetes.^3,42 Advocated aggressive diabetes treatment strategies (e.g., low saturated fat diet, regular activity and medication use) are complex, however, and Aboriginal people with diabetes can experience difficulty in interactions with non-Aboriginal health professionals with different values.⁴³

The results of this study support a therapeutic emphasis on belief in the severity of diabetes complications and a companion belief that barriers to therapeutic behaviour can be surmounted in efforts to support people with diabetes to manage their disease. The utility of beliefs in the benefits of diabetes control, and susceptibility to complications, was not upheld. But it may be reasonable not to entirely preclude attention to such perceptions. The manner by which an emphasis on perceptions is effected is likely to be key. This is because the value placed on respect for autonomy in Aboriginal culture contra-indicates attempts to influence the autonomy of others, and Aboriginal people may regard western, didactic education as interference, interruption of lifestyle, or invasion of privacy.⁴⁴

Aboriginal understandings about diabetes do not simply reflect health education rhetoric or the teachings of biomedical practitioners, which are largely ahistorical, framed at the individual level, and fail to explain the rapid emergence in the 20^th century of diabetes among Aboriginal populations with little earlier experience with the disease.^45,46 Cultural convictions about the nature of the illness experience dictate Aboriginal responses more than beliefs about the disease diabetes.⁴⁷ Negotiating a shared understanding of diabetes process and control between Aboriginal people and western health practitioners has proved important in diabetes education, treatment and community intervention.⁴⁸

Conclusion

Diabetes beliefs are amenable to change through health education,²⁵ and this analysis in Aboriginal people has illustrated empirical relations between glycemic outcomes and perceptions about having diabetes. Little research has reported on cognitive factors influencing diabetes control in Aboriginal persons.⁴⁹ The majority of studies on beliefs, behaviour and diabetes have been cross-sectional. The primary recommendation arising from this longitudinal study is for culturally sensitive education of Aboriginal people emphasizing the severity of diabetes complications and beliefs about overcoming perceived barriers to diabetes control. Further research on diabetes beliefs, behaviour and glycemic control is required to better control the growing epidemic of diabetes in Aboriginal populations.

Acknowledgements

This research was supported by Health Canada's National Health Research and Development Program through a project grant (#6610-2022-ND) and fellowship (#6610-2086-47) to Mark Daniel. The authors extend their appreciation to the Salishan people of the Okanagan region of British Columbia for their support of this research. We acknowledge with gratitude the contributions of collaborators at Okanagan University College, Diane Gamble, RN, MN, and Joyce Henderson, RN, MPH, and the assistance of Sandy Burgess, RN, Medical Services Branch, Health Canada. Lowell Laidlaw, RT, and Betty Carlson, RT, Laboratory, Vernon Jubilee Hospital, provided expert technical assistance in the conduct of biological assays.

References

1. Young TK, Reading J, Elias B, O'Neil JD. Type 2 diabetes mellitus in Canada's First Nations: status of an epidemic in progress. CMAJ 2000;163:561–6.

2. Final Report. First Nations and Inuit Regional Health Survey. St. Regis, Quebec: National Steering Committee; 1999.

3. Daniel M, Green LW. Application of the Precede-Proceed planning model in diabetes prevention and control: A case illustration from a Canadian aboriginal community. Diabetes Spectrum 1995;8(2):74-84.

4. Daniel M, Rowley KG, Herbert CP, O'Dea K, Green LW. Lipids and psychosocial status in aboriginal persons with and at risk for Type 2 diabetes: Implications for tertiary prevention. Patient Education and Counseling 2001;43(1):85–95.

5. Nagasawa M, Smith MC, Barnes JH, Fincham JE. Meta-analysis of correlates of diabetes patients' compliance with prescribed medications. The Diabetes Educator 1990;16(3):192–200.

6. Strecher VJ, Rosenstock IM. The Health Belief Model. In: Glanz K, Lewis FM, Rimer BK, editors. Health Behaviour and Health Education. San Francisco: Jossey-Bass, 1997:41–59.

7. Hochbaum GM. Public participation in medical screening programs: A sociopsychological study. Washington, D.C: U.S. Government Printing Office, 1958.

8. Rosenstock IM. Historical origins of the health belief model. Health Education Monographs 1974;2:328–35.

9. Becker MH, Janz NK. The health belief model applied to understanding diabetes regimen compliance. The Diabetes Educator 1985;11(1):41–7.

10. Polly RK. Diabetes health beliefs, self-care behaviours, and glycemic control among older adults with non-insulin-dependent diabetes mellitus. The Diabetes Educator 1992;18(4):321–7.

11. Becker MH, Maiman LA, Kirscht JP, Haefner DP, Drachman RH. The health belief model and prediction of dietary compliance: A field experiment. J Health Soc Behav 1977; 18(4):348–66.

12. Given GW, Given BA, Gallin RS, Condon JW. Development of scales to measure beliefs of diabetic patients. Research in Nursing & Health 1983;6(3):127–41.

13. Becker MH, Drachman RH, Kirscht JP. A new approach to explaining sick-role behavior in low-income populations. Am J Public Health 1974;64:205–16.

14. Janz NK Becker MH. The health belief model: A decade later. Health Education Quarterly 1984;11:1–47.

15. Algona M. Perception of severity of diabetes and health locus of control in compliant and noncompliant diabetic patients. Diabetes Care 1980;3:533–44.

16. Kirscht, JP. The health belief model and illness behaviour. Health Education Monographs 1974;2:387–408.

17. Kirscht JP. Research related to the modifications of health beliefs. In: Becker MH, editor. The Health Belief Model and Personal Health Behaviour. New Jersey: Charles B. Slack Co., 1976.

18. Maiman L, Becker M. Scales for measuring health belief model dimensions: A test of predictive value, internal consistency, and relationships among beliefs. Health Education Monographs 1977;5:215–31.

19. Harris R, Linn MW, Skyler JS, Sandifer R. Development of the diabetes health belief scale. The Diabetes Educator 1987;13(3): 292–7.

20. Harrison JA, Mullen PD, Green LW. A meta-analysis of studies of the Health Belief Model. Health Education Research 1992;7(1):107–16.

21. Green LW. Should health education abandon attitude change strategies: Perspectives from recent research. Health Education Monographs 1970;1:24–48.

22. Brown SA, Hedges LV. Predicting metabolic control in diabetes: A pilot study using a meta-analysis to estimate a linear model. Nurs Res 1994;43(6):362–8.

23. Wilson W, Ary DV, Biglan A, Glasgow RE, Toobert DJ, Campbell DR. Psychosocial predictors of self-care behaviors (compliance) and glycemic control in non-insulin-dependent diabetes mellitus. Diabetes Care 1986;9(6):614–22.

24. Coates VE, Boore JR. The influence of psychological factors on the self-management of insulin-dependent diabetes mellitus. J Adv Nurs 1998;27(3):528–37.

25. Wooldridge KL, Wallston KA, Graber AL, Brown AW, Davidson P. The relationship between health beliefs, adherence, and metabolic control of diabetes. The Diabetes Educator 1992;18(6):445–50.

26. Statistics Canada. Schooling, Work and Related Activities, Income, Expenses and Mobility. 1991 Aboriginal Data. 1991 Census of Canada (Catalogue no. 89–534). Ottawa: Industry, Science and Technology Canada, 1993.

27. World Health Organisation. WHO Study Group on Diabetes Mellitus. Geneva: World Health Organisation (Technical Report Series No. 727), 1985.

28. Daniel M, Gamble D, Henderson J, Burgess S. Diabetes prevalence, behavioural and anthropometric risk factors, and psychosocial constructs in an aboriginal population in central British Columbia. Chronic Dis Can 1995;16(4):165–74.

29. Kriska AM, Knowler WC, LaPorte RE, Drash AL, Wing RR, Blair SN, Bennett PH, Kuller LH. Development of a questionnaire to examine the relationship of physical activity and diabetes in Pima Indians. Diabetes Care 1990;13:401–11.

30. Tremblay A, Sévigny J, Leblanc C, Bouchard C. The reproducibility of a three-day dietary record. Nutrition Research 1983;3:819–30.

31. Thompson JN, Brulé D. Canadian Dietary Information System, 4.0. Ottawa: Health Canada, 1994.

32. Bradley C, Brewin CR, Gamsu DS, Moses JL. Development of scales to measure perceived control of diabetes mellitus and diabetes-related health beliefs. Diabetic Medicine 1984;1(3):213–8.

33. Becker MH, Janz NK. The health belief model applied to understanding diabetes regimen compliance. The Diabetes Educator 1985; 11(1): 41–7.

34. Hurley AC. The health belief model: Evaluation of a diabetes scale. The Diabetes Educator 1990;16(1):44–8.

35. Lewis KS, Jennings AM, Ward JD, Bradley C. Health belief scales developed specifically for people with tablet-treated type 2 diabetes. Diabet Med 1990;7(2):148–55.

36. Kurtz SM. Adherence to diabetes regimens: Empirical status and clinical applications. The Diabetes Educator 1990;16(1):50–9.

37. Bond GG, Aiken LS, Somerville SC. The health belief model and adolescents with insulin-dependent diabetes mellitus. Health Psychol 1992;11(3):190–8.

38. Schwab T, Meyer J, Merrell R. Measuring attitudes and health beliefs among Mexican Americans with diabetes. The Diabetes Educator 1994;20(3):221–7.

39. McAllister G, Farquhar M. Health beliefs: A cultural division? J Adv Nurs 1992;17: 1447–54.

40. Skyler JS. Psychological issues in diabetes. Diabetes Care 1981;4:656–7.

41. Henry RR, Genuth S. Forum One – current recommendations about intensification of metabolic control in non-insulin-dependent diabetes mellitus. Ann Intern Med 1996;124:175–7.

42. Daniel M, Green LW. Community-based prevention and chronic disease self-management programs: Problems, praises and pitfalls. In: Wilde MI, editor. Patient-Centered Healthcare. Philadelphia: Adis International, 2000, 23–30.

43. Farkas C. Ethnospecific communication in diabetes education. In: Young TK, editor. Diabetes in the Canadian Native Population. Toronto: Canadian Diabetes Association, 1987, 123–32.

44. Brant C. Native ethics and rules of behaviour. Can J Psychiatry 1990;35:534– 9.

45. Garro LC. Individual or societal responsibility? Explanations of diabetes in an Anishinaabe (Ojibway) community. Soc Sci Med 1995;40:37–46.

46. Travers KD. Using qualitative research to understand the sociocultural origins of diabetes among Cape Breton Mi'kmaq. Chronic Dis Can 1995;16(4):140–3.

47. Grams GD, Herbert C, Brown D, Heffernan C, Calam B, Wilson MA, Grzybowski S, Brown D. Haida perspectives on living with non-insulin-dependent diabetes. CMAJ 1996;55(11):1563–9.

48. Boston P, Jordan S, MacNamara E, Kozolanka K, Bobbish-Rondeau E, Iserhoff H, Mianscum S, Mianscum-Trapper R, Mistacheesick I, Petawabano B, Sheshamush-Masty M. Using participatory action research to understand the meanings Aboriginal Canadians attribute to the rising incidence of diabetes. Chronic Dis Can 1997;18(1): 5–12.

49. Delamater AM, Jacobson AM, Anderson B, Cox D, Fisher L, Lustman P, Rubin R, Wysocki T. Psychosocial therapies in diabetes: Report of the Psychosocial Therapies Working Group. Diabetes Care 2001; 24(7):1286–92.

Author References

Mark Daniel, Department of Epidemiology and Department of Health Behaviour and Health Education, School of Public Health, The University of North Carolina at Chapel Hill, North Carolina, USA

Lynne C Messer, Department of Health Behaviour and Health Education, School of Public Health, The University of North Carolina at Chapel Hill, North Carolina, USA.

Correspondence: Mark Daniel, School of Public Health, The University of North Carolina at Chapel Hill, CB #7440, Rosenau Hall, Room 302, Chapel Hill, North Carolina 27599-7440, USA; Fax: (919) 966-2921; E-mail: danielm@email.unc.edu

[Previous] [Table of Contents] [Next]