This 6-month, open-labelled, prospective, parallel group, randomized controlled trial was conducted from December 2017 to October 2018. The study was conducted at diabetes clinic of Capital Hospital, located in the capital city of Pakistan (Islamabad). The average diabetes patients’ turn over in this healthcare facility is about 80 per week. Ethics approval for the study was obtained from Monash University Human Research Ethics Committee (Ethics approval number 10817; approval date: 22nd September 2017). The protocol of this trial met Consolidated Standards of Reporting Trials (CONSORT) guidelines and American Medical Association (ADA) guidelines for medical care of people with diabetes (Association AD, 2002). This trial has been registered with Australian New Zealand Clinical Trials Registry (Registration No. ACTRN12617001327370; Registration date: 15th September 2017).

The details about demographic characteristics, self-care behaviors and disease knowledge of intervention and control group participants were collected at the start (baseline) and at the end (24th week) of the trial.

The study sample included both male and female patients older than 30 years of age with a confirmed diagnosis of type 2 diabetes and a history of HbA1c ≥ 7% within the preceding month. Patients were excluded if they were involved in any educational trial related to diabetes in past 3 months, suffering with type 1 diabetes, gestational diabetes, cognitive impairment, and terminal illness.

An informed written consent for participation in the study was taken from the participants, after explaining them the study objective and procedure. Participants were randomized into intervention and control groups by using a computer-generated randomization list and allocation was concealed by using envelop method. The participants of intervention group and control group were called on alternate days to receive pharmacist-led educational intervention and usual care, respectively, so as to minimize the interaction among them and possible contamination. The sample size of the study was calculated on the basis of its ability to detect an effect size of 1% with standard deviation of ±1.4% (Sarkadi and Rosenqvist, 2004; Hayward et al., 2005; Ali et al., 2012; Chen et al., 2012; Moreira et al., 2015), 80% sampling power, and .05 significance level, at 6-month in the intervention group compared to the control group. A sample size of sixty-two was calculated (n = 31 each in the intervention and control group), but seventy five patients were recruited in the study, in order to compensate 20% attrition rate (Chow et al., 2007; Hulley et al., 2013).

One hundred and thirty eight potentially eligible patients were identified from the hospital records. However, n = 11 patients had HbA1c levels less than 7%, n = 48 had no recent HbA1c levels, and, n = 4 refused to participate in the trial due to hectic nature of their job or they were residing far away from the city. A total of seventy-five T2DM patients were recruited from those seeking medical services for their diabetes at Capital Hospital Islamabad.

Of these seventy-five eligible participants, n = 38 were randomized to the pharmacist-led educational group and n = 37 were randomized to the usual care group. A total of sixty-six participants completed the 6-month study. The flow of study participants through the trial is presented in Figure 1 (compiled in accordance with the CONSORT guidelines). The baseline demographic characteristics of the intervention and control group are presented in Table 1. The demographic characteristics and clinical variables were similar in both groups at baseline (Table 1).

Participants in the intervention group received two face-to-face self-care educational sessions by the pharmacist. A data collection form was designed to collect the data from the participants of the study regarding their demographic characteristics. However lab data were collected by the researchers from the patients’ lab profile. In the first session after collecting the baseline information about participants’ demographic characteristics, self-care practices and disease knowledge, the pharmacist educated the intervention participants (approximately 30–40 min) about diabetes, its symptoms, normal blood glucose levels and its monitoring, food choices for diabetes, importance of regular exercise and medicine use, and, diabetes-associated completions and their monitoring. In the second visit (12th week), pharmacist reinforced the intervention group participants about importance of adhering to diabetes-related self-care practices (approximately 15–30 min). During the second session, the participants were also inquired about barriers to self-care faced by them in the past 3-month, and remedial strategy or visit to physician was suggested, if required. The intervention group participants were supplemented with printed educational material and informatory brochures (in the Urdu language) about diabetes, diabetes-associated complications, and self-care activities. Besides these two face-to-face educational sessions, intervention group participants were also followed-up telephonically after every 4-week (till the end of trial).

American Diabetes Association (ADA) guidelines have been followed to design the contents of the educational intervention and informatory leaflets for people with diabetes. In order to make the educational intervention culturally-sensitive and patient-tailored, the contents of pharmacist-led educational intervention have been validated after two rounds of the Delphi-technique, involving six endocrinologists practicing in Pakistan. The detailed methodology of this trial and educational intervention is online available in this study’s published protocol (Bukhsh et al., 2018d). The pharmacist involved in delivering the educational intervention is registered with Punjab Pharmacy Council and was not involved in any healthcare delivery process to the study participants previously.

The participants in the control group received usual medical care, but were provided with the educational session and informatory brochures after the completion of this study.

The primary outcome was change in glycated hemoglobin (HbA1c) level. Whereas, improvement in disease knowledge and diabetes-related self-care practices were the secondary outcomes of the study. Disease knowledge and self-care activities were examined by using Urdu versions of Diabetes Knowledge Questionnaire (DKQ) (Bukhsh et al., 2017a) and Diabetes Self-management Questionnaire (DSMQ) (Bukhsh et al., 2017b), respectively. Both of these study tools (DSMQ and DKQ) have been recently translated in to the Urdu language and psychometrically validated in type 2 diabetes patients in Pakistan. The scoring criteria of the DSMQ (Bukhsh et al., 2017b) and DKQ (Bukhsh et al., 2017a) have been described in detail in the published study protocol of this trial (Bukhsh et al., 2018d).

The data of the study were analyzed by using the Statistical Package for the Social Sciences (SPSS, Version 24 Inc., Chicago, IL, United States). Descriptive statistics, such numbers, standard deviation, and percentages were used to present and compare the baseline demographic characteristics of the control and experimental group participants. Chi-square test was used to test relationship between categorical variables of the participants in intervention and control. Independent-samples T-test (for continuous variables) and Fisher’s exact test (for categorical variables) was used to compare the group differences, whereas, Paired-Samples T-Test was applied to observe the difference between baseline and follow-up values. A p-value of less than .05 was considered significant for all analysis. The magnitude of effect of pharmacist-led intervention on primary outcome (HbA1c) was calculated by using Cohen’d (d = (M1−M2)/SD _P ). (Dunst et al., 2004).

Summary of changes in the outcomes of the study participants is presented in Table 2. Change in the levels of glycated hemoglobin (HbA1c) was the primary outcome of this trial. The baseline levels of glycated hemoglobin in the intervention (9.00 ± 1.43) and control (9.20 ± 1.22) group were similar at baseline (p > .05). Mean reducations for HbA1C level in the intervention group was significantly greater (.91%; p < .01) as compared to that in the control group (.28%; p = .06). Cohen’s d effect size of our intervention on glycated hemoglobin was .78. Figures 2, 3 shows the graphical representation of the changes in HbA1c levels of the control and the intervention group. After completion of the trial, the percentage of participants in this study who met the American Diabetes Association target of glycemic control (HbA1c < 7%) is presented in Figure 4.

A significant improvement (p < .05) in the self-care practices was observed in the intervention group in comparison to the usual care group. Scores for sum-scale and the four sub-scales of Diabetes Self-management Questionnaire (DSMQ) significantly (p < .05) improved in the participants of the intervention group. On the other hand, no significant improvement in DSMQ scores was noticed in the participants of the usual care group, except for Glucose Management (GM) sub-scale. Detailed result are presented in Table 2.

At the completion of trial, disease knowledge scores were significantly improved (p < .05) in the intervention group patients (p < .01) as compared to patients in the usual care group (Table 2).