A total of 468 patients participated in this study, while 67 refused, giving a response rate of 87.47%. Two-thirds of the participants were from the urban area. The average age of the participants was 58 years (SD ± 11.64) and more than half of them were females (58.8%). The majority of the participants were uneducated (64.7%, n = 303). Most were married (84.4%, n = 395). The median duration of diabetes was 10 years (range four years – thirty-two years). The sociodemographic characteristics of the participants are summarized in .
Demographic characteristics of the study sample
Regarding the regularity of follow-up in the clinic only 7.9% (n = 37) of the participants had not missed any appointment in the last one year, while almost half of them (49.4%)had missed an appointment once or twice, and 41% (n = 191) more than twice. More than 50% (n = 249) of the participants did not attend the clinic on the day of interview; 42.9% (n = 94) of them mentioned unavailability of transport as the excuse for non- attendance, and 15.5% (n = 34) said they had forgotten , while 28.7% (n = 63) of them considered it unnecessary as they were taking medicine from other sources. More than half of the participants (57.5%, n = 289) did not adhere to the anti-diabetic medication as advised by the GP. The same obtained in the advice on exercise, where 62.6% (n = 293) did not follow the instructions given by the GP. However, the instructions on diet were followed by 64.7% (n = 303) of the participants.
Regarding the information received from the general physician, most of the patients reported that they had got enough information on ‘how to take the medicine’ (91%), ‘how long it would take to act’, (68.6%), and ‘how long the medicine should be taken’ (69%), but this was not true with the information regarding the side effects. Sixty-one percent of the participants did not receive any information on the side effects of the medicine and 64% did not know what to do if there were any side effects from the medicine. Most of the participants (96%, n = 448) agreed that the attending physicians completely understood their health problem on the day of appointment, and 90% (n = 421) were comfortable with the multiple drug prescriptions.
The overall prevalence of therapeutic non-compliance, that is, Hb1Ac level of more than 7 with the optimum treatment among the participants was 67.9 % (n = 318, 95% CI 63.59 – 72.02%). The non-compliance of the males (69.34%) was higher than that of the females (65.45%, P = .003). The non-compliance in the urban participants was significantly higher than (71.04 vs. 60.15%, P = .023) in the rural participants. There was a statistically significant difference in the prevalence rate of non-compliance among the participants of different educational levels. It was highest among the illiterates (72.6%, P = .001), falling as the level of education rose. It was , 61.60% among those with primary school education , 47.61% among those with secondary school education , and 45.83% in those educated beyond high school. Patients who were regular on follow-up had a significantly higher compliance rate than those who were irregular (46.88% for those who never missed an appointment, 35.53% for those who had missed an appointment once or twice in a year, 26.40%for those who had missed the appointment more than twice in a year and 18.19% for those who never attended the clinic, P = .039). The non-compliance was higher among the patients who did not follow the exercise regime than those who followed it (66.66% vs. 54.67%, P = .012). However, this did not hold for instructions on diet where non-compliance was statistically insignificant .
Patients’ related factors associated with non compliance
The non-compliance was least (48%, P = .003) with the single drug regimen (Metformin) while it was highest (79.31%, P = .003) among patients who were on combined oral and insulin treatment. The non-compliance was also higher (79.03%, P = .003) in patients who were on insulin therapy only. With regard to a multiple oral drug regime, non-compliance was more among patients who were on Metformin and Glibenclamide (66.66% Vs 61.29%, P = .003) than those who were on Metformin and Gliclazide . Physician related factors are described in .
Factors found to be significantly associated with non-compliance on bivariate analysis were: Male gender (OR = 1.90, CI =1.32-4.57), education level (literacy) (OR = 5.27, CI = 4.63–7.19),urban population (OR = 5.22, CI = 3.65-8.22), irregularity of follow up (OR = 8.41, CI = 4.90-11.92), non- adherence to drug prescription (OR = 4.55 , CI = 3.54-5.56), non -adherence to instruction on exercise (OR = 5.55, CI = 4.2 66.86), insulin (OR = 1.29, CI = .71-1.87), insulin with oral antidiabetic (OR = 1.20, CI = .65-1.75).
Age, marital status, duration of diabetes, associated chronic disease, and attendance on the day of appointment were not significantly associated with non-compliance. However, non-compliance was significantly higher among the patients who did not attend the clinic on the day of appointment, because of unavailability of transport, than those who forgot the appointment day, and those who thought it unnecessary as they were taking medicine from other sources (78.3, 76.3, and 58.42%, respectively, P = .000). Patient–doctor interaction factors, such as, the patients’ perception of the physician's understanding of their health problems and the patients’ ease with the prescription of multiple drugs, were not significantly associated with the compliance rate.