Adherence to insulin therapy and associated factors among patients with diabetes mellitus in public hospitals of Central Zone of Tigray, Ethiopia, 2018: a cross-sectional study

Introduction Adherence to insulin therapy is a critical factor for adequate control of diabetes mellitus. Despite the multiple well-known benefits of adherence to insulin therapy, poor adherence remains to be a common cause of diabetes mellitus-related complications. A better management of diabetes mellitus requires determining the level of patient adherence and identifying why non-adherence to insulin therapy occurs. Therefore, this study was designed to assess the level of adherence to insulin therapy and associated factors among diabetes mellitus patients. Methods The study was conducted from May 1 to July 1, 2018, using a cross-sectional study design. Interviewer-administered questionnaire was employed for data collection and systematic random sampling technique was used to select study participants. The collected data were entered using Epi data version 3.1.1 and exported to SPSS version 22 for analysis. Logistic analysis was carried out to check the level of association between adherence to insulin therapy and the independent variables with significance level of 0.05 at 95% confidence interval. Results 273 respondents were selected with a 100% response rate. Near to one-fourth (24.2%) of the respondents were adherent to their insulin therapy. The study revealed that good knowledge of diabetes mellitus [AOR=6.51; 95% CI [1.58, 26.71], age [>30 years] [AOR=2.63; 95% CI [1.27, 5.42], knowledge regarding insulin self-injection [AOR=4.21; 95%CI [1.06,16.65], favorable attitude towards insulin injection [AOR=2.14; 95% CI [1.04,4.41], free-of-cost insulin therapy [AOR= 4.62, 95% CI [1.06,16.65], having of glucometer at home [AOR= 2.82, 95% CI [1.12,7.09], and being a member of Ethiopian diabetic association [AOR= 5.41, 95% CI [2.31,12.64] were found to significantly affect adherence to insulin therapy. Conclusion Nearly one-fourth of the study participants were adherent to their insulin therapy. Good knowledge and favorable attitude towards insulin injection, good knowledge regarding diabetes mellitus, being a member of the Ethiopian Diabetes Association, age greater than thirty years old, free-of-cost insulin therapy and having glucometer at home were found to be significant predictors of adherence to insulin therapy.


Introduction
Diabetes mellitus (DM) is a group of metabolic disorder characterized by elevated blood glucose levels (hyperglycemia) [1]. It is one of the major public health problems affecting over 451 million people globally and with a projected increase to 693 million by 2045 [2]. Inadequate control of blood glucose level can lead to several disorders, including cardiovascular diseases, which are the leading causes of morbidity and mortality for people with diabetes mellitus. Moreover, it accounts for the higher direct and indirect medical expense of people affected by DM [3]. To control the burden of DM, patients need to use insulin therapy as ordered by the health care providers. Type 1 DM (T1DM) patients are treated by multiple-dose insulin injection or continuous subcutaneous insulin infusion, but type 2 DM (T2DM) is a progressive disease with a treatment protocol adjusted in line with the disease's progression. Many patients with T2DM eventually require insulin to maintain adequate glycemic control [4]. With the ever-increasing number of DM patients and its negative effect on economy, health system and society, there has been a need for rapid progress in the care and management of patients with insulin treatment which, by improving glycemic control, slows down the progress of microvascular and macrovascular complications [5,6]. Anti-diabetic medications are effective only when they are taken according to the recommendations of health care providers. However, proper orders without good adherence would not result in the intended treatment outcome [7,8].
Good adherence to insulin treatment, compared to a low-level of adherence, has been associated with improved levels of glycemic control [9] which, in turn, reduces the risk of developing acute and chronic diabetic complication [10,11] and mitigating its economic burden [12,13].
Despite the widely known fact of multiple benefits of adherence to insulin therapy, poor adherence among patients remains a common problem in substantial number of patients with diabetes mellitus [14][15][16][17]. Specifically, while insulin is the most efficient glucose lowering agent [18], previous researches have indicated that a large proportion of T2DM patients discontinue their basal insulin therapy [16,19,20]. In general, 41-57% of insulin-treated persons were non-adherent with a 20% non-adherence rate [14].
Consequently, a significant number of patients die prematurely from cardiovascular diseases [21]. In addition, some studies in Ethiopia shows that there is poor adherence to insulin therapy among DM patients [22,23]. To improve patient's adherence to insulin treatment, it is important to determine the extent of patient adherence and why non-adherence to insulin therapy occurs. Identifying predictors of poor adherence are necessary to enable stakeholders to take all the initiatives to tackle these factors. However, there is a lack of information about the extent of adherence and associated factors to poor adherence for insulin therapy in developing countries. In Ethiopia, the level of adherence and associated factors of insulin therapy remain to be insufficiently studied. Therefore, this study was designed to assess the level adherence to insulin therapy and associated factors among patients with DM in public hospitals of central zone of Tigray regional state, Ethiopia.

Methods
The study was conducted in public hospitals of the Central Zone, Participants were selected using systematic random sampling technique. The participants were selected from each hospital in accordance with the magnitude of the population they entertain.
Lottery method was used to select participants from each hospital.
Data were collected using interviewer-administered structured questionnaire. The questionnaire contained six parts: the first part (part I) was used to collect socio-demographic data, part II was used to collect adherence to insulin therapy, part III was used to collect health profile data of the study subjects, part IV was regarding Diabetes Knowledge Questionnaire (DKQ), part V was knowledge regarding insulin therapy and part VI was used to measure the patients' attitude towards insulin therapy. The data were collected using a structured interviewer-administered questionnaire. A modified morisky adherence scale of 8 items (MMAS-8) was used to measure patients' adherence to their insulin therapy. MMAS-8 is a validated tool which consists of 8 item self report measure of adherence. Items 1 through 7 had response choices of "yes" or "no" whereas item 8 had 5-point Likert scales. Each "no" response was rated as "1" and each "yes" was rated as "0" except for item 5 (reversed score), in which the response "yes" was rated as "1" and "no" was rated as "0". The 5-point Likert scales of item 8, which is concerned on the difficulty to remember taking their insulin therapy, were never/rarely = 0, once in a while = 1, sometimes = 2, usually = 3 and all the time = 4. In this scale, if patients choose response "0", the score is "1" and if they choose response "1,2,3, or 4", the score is "0". Higher scores indicate higher adherence levels. The total scores of all the items range from 0 to 8 and was grouped into two levels: adherent (score of 6 to 8), and non-adherent (score < 6).
Knowledge regarding insulin therapy and knowledge regarding DM Multicollinearity was assessed by variance inflation factor and all assumptions of binary logistic regression were checked. A statistical analysis was made at a confidence interval of 95% with P value less than 0.05.

Operational definitions
Adherence: is the extent to which a person's behavior; taking medication, following a prescribed diet, and/or executing lifestyle changes corresponds with agreed recommendations from the health care provider. This was measured with Morisky's 8-item scale which, when the MMAS-8 score is ≥ 6, the study participant is considered to be adherent.

Results
Demographic profile of study participants: a total of 273 respondents were selected in this study with a response rate of 100%.
Of these study subjects, 37.4% were female participants and more than half of them (59%) were living in rural areas. Close to one-third of the subjects (36.6%) were single and with respect to their educational level, few (16.5%) couldn't read and write. Regarding ethnicity and religion, the majority were Tigrayans (98.9%) and Orthodox Christian (92.7%) ( Table 1).
Health profile of the respondents: in most of the participants (79.5%), the duration since they were diagnosed with DM was less than 10 years. 10% of the participants were getting insulin therapy free of cost. With respect to family history and having glucometer, 16.8% had family history of DM, and 18% had glucometer at home.
Nearly two-third (65.8%) of the respondents attended DM-related education, and more than a quarter (29.3%) of the subjects were member of the Ethiopian Diabetic Association (EDA). Almost all (98.5%) of the respondents had visited healthcare provider regularly (Table 2).
Knowledge and attitude of the respondents: the majority of the respondents had good knowledge regarding insulin self-injection (94.1%) and diabetes mellitus (82.8%). With respect to attitude towards insulin injection, more than a quarter (27.8%) of the respondents had a favorable attitude (Table 3).
Factors associated with adherence to insulin therapy: analysis of the data with multivariable logistic regression showed seven variables to be significantly associated with adherence to insulin therapy at 5% level of significance. Knowledge of insulin self-injection was significantly associated with adherence to insulin therapy. The odds of adherence to insulin therapy among those who had good knowledge of insulin self-injection was 4

Discussion
This study provided an insight into the level of adherence to insulin therapy and associated factors among patients with DM in public hospitals of central zone, Tigray regional state, Ethiopia in the year 2018. The level of adherence to insulin therapy was found to be 24.2%. This was considerably lower than findings from other studies

Conclusion
Nearly one-fourth of the study participants were found to be adherent to their insulin therapy. Good knowledge and favorable attitude towards insulin therapy, good knowledge regarding diabetes mellitus, being member of a diabetes association, older age (>30 years), getting insulin therapy for free and having glucometer at home were found to be the predictors of adherence to insulin therapy.

What is known about this topic
• Poor adherence to insulin therapy is the principal cause for the development of diabetes mellitus complications; • There is a lack of information about the extent of adherence and associated factors to insulin therapy in the developing countries; • While insulin is the most efficient glucose-lowering agent, previous research has indicated that a large proportion of diabetes mellitus patients discontinue their insulin therapy due to various factors.

What this study adds
• Nearly one-fourth of the study participants were adherent to their insulin therapy; • Free-of-cost insulin therapy, having glucometer at home, a good knowledge of DM, favorable attitude towards insulin therapy, being members of diabetes association and older age were the significant predictors of adherence to insulin therapy; • Therefore, health care professionals must encourage patient to be members of DM association and take actions to increase patient's knowledge, attitude and practice towards insulin therapy by strengthening the delivery of information, education and communication program.

Competing interests
The authors declare no competing interests.

Acknowledgments
We thank all the subjects of the study and the data collectors for their genuine contribution to the success of our work.