CC BY-NC-ND 4.0 · Journal of Social Health and Diabetes 2018; 06(02): 090-095
DOI: 10.1055/s-0038-1675687
Original Article
NovoNordisk Education Foundation

Noncompliance to Diet and Medication among Patients with Type 2 Diabetes Mellitus in Selected Hospitals of Kathmandu, Nepal

Nisha Kusum Kafle
1  Department of Public Health, Tribhuvan University Institute of Medicine, Kathmandu, Nepal
,
Resham Raj Poudel
2  Department of Internal Medicine, Northside Medical Center, Youngstown, Ohio, United States
,
Sushan Man Shrestha
1  Department of Public Health, Tribhuvan University Institute of Medicine, Kathmandu, Nepal
› Author Affiliations
Funding None.
Further Information

Address for correspondence

Nisha Kusum Kafle, BPH
Department of Public Health, Tribhuvan University Institute of Medicine
Kathmandu
Nepal   

Publication History

Publication Date:
12 November 2018 (online)

 

Abstract

Background Diabetes is a major public health problem affecting people of all ages globally. Noncompliance compromises the effectiveness of treatment and adversely affects patients' health. The main purpose of this study was to assess and compare the proportion of noncompliance to diet and medication between patients with type 2 diabetes mellitus (T2DM) visiting public and private hospitals in Kathmandu, Nepal.

Methods Descriptive cross-sectional study was conducted in T2DM patients visiting public and private hospitals. Eight item Morisky Medication Adherence Questionnaire (MMAQ) for medication adherence and Perceived Dietary Adherence Questionnaire (PDAQ) for dietary adherence were used. Epidata was used for data entry and SPSS for data analysis. Chi-square test was used as a test of significance. Odds ratio (OR) and the corresponding 95% confidence intervals (CI) were calculated.

Results The study involved 182 T2DM patients. Participants' age was ≥ 17 years and they were under treatment for ≥ 6 months. Mean age of the participants was 54.67 years with standard deviation (SD) ± 11.69. Prevalence of medication noncompliance was seen in 126 (69.2%) patients, whereas prevalence of dietary noncompliance was seen in 166 (91.2%) patients. Illiterate participants were more likely to be noncompliant than literate to medication (OR 4.32, p = 0.001). Self-employed were more likely to be noncompliant to medication than job holders (OR 2.93, p = 0.008). People visiting public hospital were more likely to be noncompliant to diet than those visiting private hospital (OR 4.89, p = 0.009). Illiterate participants were more likely to be noncompliant to diet than literate (OR 10.94, p = 0.005).

Conclusion The T2DM patients visiting public hospitals were more noncompliant to diet. Illiterate patients were more noncompliant to both medication and diet. Self-employed compared with job holders were more noncompliant to medication. Patient education and counseling should be aggressively addressed mainly in public hospitals. There was no significant difference in medication noncompliance between public and private hospitals (p = 0.108).


#

Introduction

Diabetes is a group of metabolic disorders associated with long-term damage, dysfunction, and failure of different organs, especially the eyes, kidneys, nerves, heart, and blood vessels.[1] Majority (90–95%) of patients have type 2 diabetes mellitus (T2DM).[2] Compliance to diet and medication is defined as an extent to which a person's behavior in terms of taking medications and following diet coincides with the health care provider's recommendations.[3]

According to the World Health Organization (WHO), diabetes is the sixth leading cause of death accounting for 1.59 million deaths in 2015.[4] According to International Diabetes Federation (IDF), diabetes affects approximately 415 million people worldwide, and the number is expected to reach 642 million by 2040 with two-thirds of all diabetes cases and > 75% of diabetes deaths occurring in lowto middle-income countries.[5] According to the IDF data for Nepal, prevalence of T2DM in 20 to 79 years age group was 4% in 2017, and the predicted number of undiagnosed cases was 532,100. IDF estimates the prevalence to reach 6.1% and 1,264,200 undiagnosed cases in 2045.[6] Diabetes is the third most common noncommunicable disease in Nepal, which causes 12% of all hospitalizations.[7] T2DM is emerging as a major health care problem in Nepal, with rising prevalence and its complications, especially in urban population complicated by noncompliance of diet and medication[8] Centralized health care, poor referrals and consultation system, and increasing trends of urban lifestyle in Nepal further complicate diabetes management[9] In a cross-sectional study in Nepal, dietary noncompliance was 87.5% and 12.5% were poorly compliant.[10] In another study, only one-fifth of the patients believed that being compliant to dietary advice helps reduce blood glucose.[11] Diabetes is a chronic disease that requires lifelong treatment. It greatly increases the risk of serious, long-term complications and affects health care costs and overall quality of life. Noncompliance to long-term therapy severely compromises the effectiveness of treatment and adversely affects the patient's condition.[12] Compliance to medication and dietary recommendations lessens the disease burden by reducing morbidity, mortality, and complications associated with T2DM.[13]


#

Methods

Between August 2017 and October 2017, a descriptive, analytical study with cross-sectional design was performed at four hospitals in Kathmandu, Nepal: two tertiary care public hospitals (TU Teaching Hospital and Bir Hospital) and two private hospitals (Metro Hospital and Diabetes Thyroid & Endocrinology Care Center). Hospitals were selected keeping in mind to cover the most representative population. The study population included all registered T2DM patients attending outpatient departments of selected hospitals during research period. All participants were of age ≥ 17 years and were under treatment for at least 6 months. Patients with gestational diabetes, severe comorbidity, and severe mental illness were excluded.

The minimum required sample size was calculated as 182 by using the formula n = z 2 pq/d 2, where n = required sample size, p = prevalence of noncompliance to diet, which was 87.5% [10], q = 1-p and d = deviation of ± 5% from true prevalence, and z = level of confidence measured; for 95% confidence interval (CI) (α = 0.05), z = 1.96. The study tool was pretested in 15 patients (8% of sample size) at Sahid Gangalal National Heart Center, Kathmandu. Necessary corrections and adjustments were made, and tools were finalized. Responses from pretest were not included in final analysis. Equal number of sample from each hospital was taken, that is, 46 samples from each four hospital. Because of limited time, resources purposive sampling was used for selecting samples. Data were collected by face-to-face interview with the patient by the researcher, after taking informed written consent.

Morisky Medication Adherence Questionnaire (MMAQ) that is a validated questionnaire was used to assess medication noncompliance.[14] [15] MMAQ consists of eight questions in which questions 1 to 7 have response choice Yes or No and question 8 has 5-point Likert response choice. Based on score obtained, 8 was considered as high compliance, 6 and 7 as medium compliance, and < 6 was considered as low compliance.

Dietary noncompliance was assessed by using validated questionnaire Perceived Dietary Adherence Questionnaire (PDAQ).[16] PDAQ is a 7-point Likert scale-based tool to measure dietary compliance. It has a total of nine questions, with scores ranging from lowest 0 to highest 7. Total score of PDAQ is 63. Based on the score obtained, > 75% was considered as high compliance, 50 to 75% as medium compliance, and < 50% as low compliance. Sociodemographic, behavioral, and other related variables used were based on previous studies and WHO NCD STEPS instrument.[17] For statistical analysis of both medication and diet noncompliance, only high compliance was considered as true compliance, and middle and low compliance were considered as noncompliance.

After coding, editing, and cross-checking, data were entered in EpiData ver. 3.1 (The EpiData Association Odense, Denmark 2004) and then exported to SPSS ver. 21 (IBM Corp: Armonk, NY, US 2012) for further analysis. Descriptive analysis was done in terms of number and percent for qualitative data, and mean and standard deviation (SD) for quantitative data. Bivariate analysis was performed to see the crude association of independent variable with the outcome variable by using chi-square test. p-Value (< 0.05) and 95% CI were used to see the significance of association.


#

Results

Total 182 T2DM patients were interviewed. The mean age of participants was 54.67 ± 11.69 years, 53.3% were male, and 46.7% were female. Of total participants, 89% were married, 25.8% were illiterate, approximately 40% were household worker, and 32.4% were self-employed. Almost 85% of the participants were from city/municipality. Nearly 35% of participants had family history of diabetes. Mean duration of diabetes was 6.88 ± 6.14 years, 56% of participants had diabetes for < 5 years, and remaining had diabetes for 5 to 28 years. More than one-half (62.5%) of the participants had hypertension, and 38.46% had no additional disease besides T2DM. Of total, 14.1% of the participants were current smoker whereas 40.1% were ever smoker and 59.9% were never smoker. In case of drinking habit, 19.8% were current drinkers, 41.2% were ever drinkers, and 58.8% were never drinkers. Responses of the participants on MMAQ, responses on PDAQ, and the participants' noncompliance status are presented in [Tables 1] to [3], respectively.

Table 1

Response on MMAQ–8

Questions

Number (%), n = 182

Yes

No

Abbreviations: MMAQ–8, Morisky Medication Adherence Questionnaire 8; SD, standard deviation.

Q1. Forget to take medicine sometimes

80 (44)

102 (56)

Q2. Any days forgot to take medicine over past 2 week

39 (21.4)

143 (78.6)

Q3. Stop taking medicine without telling physician when felt worse

20 (11)

162 (89)

Q4. Sometimes forget to bring along medicine when traveling or leaving home

42 (23.1)

140 (76.9)

Q5. Took all medicines yesterday

174 (95.6)

8 (4.4)

Q6. Sometimes stop taking medicines when symptoms are under control

16 (8.8)

166 (91.2)

Q7. Ever feel hassled while sticking to treatment plan

57 (31.3)

125 (68.9)

Q8. Difficulty in remembering to take all your medicines

Mean ± SD

0.89 ± 0.23

Table 2

Response on PDAQ

Questions

Mean ± SD, n = 182

Abbreviation: PDAQ, Perceived Dietary Adherence Questionnaire; SD, standard deviation.

Q1.

No. of days followed healthful eating plan in past 7 days

5.12 ± 1.42

Q2.

No. of days ate adequate fruits and vegetables in past 7 days

5.27 ± 1.34

Q3.

No. of days ate carbohydrate–containing food with low glycemic index in past 7 days

4.28 ± 1.95

Q4.

No. of days remove food high in sugar in past 7 days

6.11 ± 1.38

Q5.

No. of days ate high–fiber food in past 7 days

4.36 ± 2.26

Q6.

No. of days carbohydrates were spaced evenly throughout the day in past 7 days

5.96 ± 1.62

Q7.

No. of days ate fish or food high in omega–3 fats in past 7 days

0.66 ± 1.22

Q8.

No. of days ate food that contained or was prepared with canola, walnut, in past 7 days

1.79 ± 1.99

Q9.

No. of days remove foods high in fat in past 7 days

5.711± 0.61

Table 3

Participants' noncompliance status

Variables

Number (%)

Number (%)

Number (%) n = 182

High compliance

Medium compliance

Low compliance

Medication compliance: Score 8 high compliance, 6–7 medium compliance, and < 6 low compliance. Dietary compliance: > 75% high compliance score, 75–50% medium compliance, and < 50% low compliance.

Medication advice

56 (30.8)

75 (41.2)

51 (28.0)

Dietary advice

16 (8.8)

144 (79.1)

22 (12.1)

After considering only high compliance as compliance, and middle and low compliance as noncompliance: Prevalence of medication noncompliance, score < 8 on MMAQ was 69.2%. Prevalence of dietary noncompliance, score < 75% on PQDA was 91.2%.

Chi-square test was used as a test of significance to see the association of independent variables (age, sex, marital status, health facility type, family history of diabetes, duration of diabetes, occupation, education, place of residence, smoking habit, and drinking habit) with outcome variable—noncompliance. Odds ratio (OR) and the corresponding 95% CI were calculated, and two-sided p-value < 0.05 was considered significant.

Factors found to be significantly associated with medication noncompliance on bivariate analysis were level of education and occupation of the participants. Illiterate (no formal education) participants were 4.32 times more likely to be noncompliant than literate (formal education) (CI: 2.00–9.30, p = 0.001). Self-employed participants were 2.93 times more likely to be noncompliant than job holder (CI: 1.30–6.59, p = 0.008).

Factors found to be significantly associated with dietary noncompliance on bivariate analysis were type of health facility and level of education. Participants who visited public hospital were 4.89 times more likely to be noncompliant than those who visited private hospital (CI: 1.34–17.79, p = 0.009). Illiterate participants were 10.94 times more likely to be noncompliant than literate participants (CI: 1.41–84.75, p = 0.005). Characteristics of study participants are shown in [Table 4]. Comparison of medication and diet noncompliance is shown in [Fig. 1].

Table 4

Study participants' characteristics

Variables

Number (%) (n = 182)

Medication

Diet

Compliant (n = 56)

Noncompliant (n = 126)

Odds ratio

p–Value

Compliant (n = 16)

Noncompliant (n = 166)

Odds ratio

p–Value

Health facility type

Public

91 (50.0)

23 (25.3)

68 (74.7)

1.68

0.108

3 (3.3)

88 (96.7)

4.89

0.009

Private

91 (50.0)

33 (36.3)

58 (63.7)

13 (14.3)

78 (85.7)

Age

17–60

128 (70.3)

42 (32.8)

86 (67.2)

0.72

0.358

10 (7.8)

118 (92.2)

1.48

0.473

> 60

54 (29.7)

14 (25.9)

40 (74.1)

6 (11.1)

48 (88.9)

Sex

Male

97 (53.3)

34 (35.1)

63 (64.9)

0.65

0.181

12 (12.4)

85 (87.6)

0.35

0.68

Female

85 (46.7)

22 (25.9)

63 (74.1)

4 (4.7)

81 (95.3)

No. of year with diabetes mellitus (DM)

0.5–5

104 (57.1)

31 (30.4)

71 (69.6)

1.04

0.901

6 (5.9)

96 (94.1)

2.29

0.118

> 5

78 (42.9)

25 (31.3)

55 (68.7)

10 (12.5)

70 (87.5)

Occupational status

Self–employed

153 (84.1)

41 (26.8)

112 (73.2)

2.93

0.008

15 (9.8)

138 (90.2)

0.33

0.268

Job holders

29 (15.9)

15 (51.7)

14 (48.3)

1 (3.4)

28 (96.6)

Level of education

Illiterate (no formal education)

71 (39.0)

10 (14.1)

61 (85.9)

4.32

0.001

1 (1.4)

70 (98.6)

10.94

0.005

Literate (formal education)

111 (61.0)

46 (41.4)

65 (58.6)

15 (13.5)

96 (86.5)

Place of residence

Village municipality

27 (14.8)

5 (18.5)

22 (81.5)

2.16

0.135

1 (3.7)

26 (96.3)

2.79

0.312

City/Municipality

155 (85.2)

51 (32.9)

104 (67.1)

15 (9.7)

140 (90.3)

Marital status

Married

162 (89.0)

48 (29.6)

114 (70.4)

1.58

0.441

15 (9.3)

147 (90.7)

0.52

0.526

Others

20 (11.0)

8 (40)

12 (60)

1 (5.0)

19 (95.0)

Family history of DM

No

117 (64.3)

31 (26.5)

86 (73.5)

1.73

0.094

7 (10.8)

58 (89.2)

1.45

0.482

Yes

65 (35.7)

25 (38.5)

40 (61.5)

9 (7.7)

108 (92.3)

Additional problem

No

70 (38.5)

27 (38.6)

43 (61.4)

0.56

0.071

9 (8.0)

103 (92.0)

0.79

0.649

Yes

112 (61.5)

29 (25.9)

83 (74.1)

7 (10.0)

63 (90.0)

Tobacco use

Current nonsmoker

155 (85.16)

44 (28.4)

111 (71.6)

2.02

0.095

15 (9.7)

140 (90.3)

0.36

0.312

Current smoker

27 (14.8)

12 (44.4)

15 (55.6)

1 (3.7)

26 (96.3)

Never smoker

109 (59.89)

31 (28.4)

78 (71.6)

1.31

0.406

102 (93.6)

7 (6.4)

0.25

0.168

Ever smoker

73 (40.11)

25 (34.2)

48 (65.8)

64 (87.7)

9 (12.3)

Alcohol consumption

Current nondrinkers

146 (80.22)

44 (30.1)

102 (69.9)

1.16

0.710

13 (8.9)

133 (91.1)

0.93

0.914

Current drinkers

36 (19.8)

12 (33.3)

24 (66.7)

3 (8.3)

33 (91.7)

Never drinkers

107 (58.79)

34 (31.8)

73 (68.2)

0.89

0.725

6 (5.6)

101 (94.4)

2.59

0.070

Ever drinkers

75 (41.2)

22 (29.3)

53 (70.7)

10 (13.3)

65 (86.7)

Zoom Image
Fig. 1 Comparison of medication and diet noncompliance.

#

Discussion

This is the first study conducted to see medication noncompliance in Nepal, whereas few studies were conducted for dietary noncompliance showing noncompliance rate from 87.5 to 58.9%. This study shows that noncompliance to medication among T2DM patients was 69%. Study in India performed by using same tool showed noncompliance varying from 55[18] to 60%.[19] Medication noncompliance in a study done in eastern Uganda was 16.7% in Ethiopia[20] and 28 to 31.2% in Kolkata,[21] [22] with other studies showing 42.3% in India,[23] 50% in Spain,[24] 54.5% in Kenya,[25] and 67.9% in Saudi Arabia.[8]

This difference in medication noncompliance between our and other studies is due to variation in categorization of the “degree of noncompliance.” In our study only the score of 100% in MMAQ was considered as compliance to medication. In other studies, Adama (Gelaw et al)[21] and south India (Divya and Nadig, Manobharathi et al)[18] [19] ≥ 75% were considered as compliance. Study of Uganda[20] and Kolkata[23] considered score ≥80% as compliance. Most of these studies, that is, studies of Ethiopia,[22] Kenya,[25] Adama,[21] Kolkata,[23] and South India[18] covered data of single health care center. In this study there was a significant illiterate versus literate difference in noncompliance rate. Illiterate people were 4.32 times more likely to be noncompliant than literate (p = 0.001). Similar study in India showed that noncompliance to medication was significantly associated with educational status (p = 0.022) [23], (p = 0.04527).[26] This study showed that self-employed participants were 2.93 times more likely to be noncompliant (p = 0.008) than job holder participants, which can be correlated with similar study in tertiary care hospital in India, which showed that noncompliance to medication was significantly associated with employment status (p = 0.0001).[26]

Prevalence of dietary noncompliance in this study was 91.2%. Other studies in Nepal showed noncompliance to diet from 41 [11] to 100% (medium + poor).[10] International studies show dietary noncompliance variations from 97.8% in Egypt,[27] 62% in Mexico,[28] 48% in eastern Washington,[29] to 37% in Botswana, South Africa.[30] This study found that illiterate participants were 10.94 times more likely to be noncompliant to diet than literate. Participants who visited public hospital were 4.89 times more likely to be noncompliant to diet than those who visited private hospital/diabetes clinic. No similar study was conducted previously for comparison.


#

Limitation of the Study

In 7 days recall method, sometimes there could be recall bias. However, it is the most suitable method comparatively.


#

Conclusion

High rate of noncompliance to medication advice and dietary advice was found among T2DM patients in Kathmandu, Nepal. Dietary noncompliance was higher than medication noncompliance. It was found that place of treatment had significant effect on patients' dietary compliance. Medication noncompliance was affected by the participants' education and occupation status. Dietary noncompliance was influenced by level of education and place of treatment. Health care providers should be aware of such high prevalence of noncompliance in patients and put more efforts in educating patients regarding the necessity of compliance and poor outcomes that come from being noncompliant (with more focus in public hospitals). Further studies should be performed to find out more specifics on the determinants of noncompliance, which would help in intervention strategy.

Ethics Approval and Consent to Participate

Ethical approval was taken from National Health Research Council (NHRC). Permissions were taken from Teaching Hospital, Bir Hospital, Metro Hospital, Diabetes Thyroid & Endocrinology Care Center, and Sahid Gangalal National Heart Center, before data collection.


#
#
#

Conflict of Interest

None.

Acknowledgments

We would like to offer especial thanks to all the participants who are actively involved in this research and to all the hospitals that allowed and provided friendly environment for data collection. We are grateful to the Department of Community Medicine and Public Health (IOM) for providing opportunity to carry out the research.


Address for correspondence

Nisha Kusum Kafle, BPH
Department of Public Health, Tribhuvan University Institute of Medicine
Kathmandu
Nepal   


  
Zoom Image
Fig. 1 Comparison of medication and diet noncompliance.