Personal characteristics of the studied pharmacists

The scale measuring TPB constructs was designed based on the methodology described by Ajzen [12]. The constructs were measured in two ways: direct and indirect measurements. Responses to each item were scored on a five-point Likert scale ranging from (1) strongly disagree to (5) strongly agree.

Direct measures: Pharmacists’ intention to provide diabetic care was measured by five statements (example: “I intend to provide diabetic patients with information concerning the suitable diet” and “I intend to provide diabetic patients with information on the importance of performing regular exercise”) and was classified into low (5-11), moderate (12-18) and high (19-25). The statements were tested for reliability using Cronbach’s alpha, which was 0.868.

Pharmacists’ attitude towards diabetic care was assessed through six statements (example: “In general, I believe that pharmacists should be involved in giving instructions on use of glucose meters” and “In general, I believe that pharmacists should be involved in advising patients to perform regular check for diabetes complications.” Attitude was classified into; negative (6-13), neutral (14-22) and positive (23-30). The statements were tested for reliability using Cronbach’s alpha, which was 0.901.

Subjective norm was assessed using two statements: “Most people I deal with as a pharmacist encourage my involvement in diabetes care” and “It is expected of me to be involved in diabetes care”. It was classified into; negative (2-4), neutral (5-7) and positive (8-10). The statements were tested for reliability using the Cronbach’s alpha, which was 0.595.

Perceived behavioral control was assessed using two statements: “I am confident that I could identify clients at risk of diabetes mellitus type 2” and “Counseling patients on the drug intake and compliance is easy to me”. It was classified into; low (2-4), moderate (5-7) and high (8-10). The statements were tested for reliability using the Cronbach’s alpha which was 0.671.

Indirect measures-Behavioral beliefs: Respondents were asked to indicate their agreement on 6 outcomes to practicing diabetes care (example: “counseling patients on drug intake and compliance can decrease drug interactions”), and their evaluation of these outcomes (example: In my opinion decreasing drug interactions is (very unimportant, 1 to very important, 5)).

Responses to each pair of items were multiplied together and an indirect measure of attitude (i.e., a belief-based measure of attitude) was subsequently constructed by summing these products across all beliefs. The range of possible scores was 6 to 150.

Normative beliefs: Respondents were asked to indicate the extent to which 5 important referents would approve or disapprove of pharmacists practicing diabetes care (example: Most pharmacists I work with promote the profession of pharmacy through involvement in patient counseling) and the extent to which they were motivated to comply with these referents (example: Generally, I would like to do what other pharmacists do). Each of the two variables was measured using 6 statements rated on a 5-point ranging from strongly disapprove (1) to strongly approve (5). Responses to each pair of items were multiplied together and an indirect measure of subjective norm (i.e., a belief-based measure of subjective norm) was subsequently constructed by summing these products across all referents. The range of possible scores was 6 to 150.

Control beliefs: Respondents indicated the presence or absence of 5 control factors that would hinder practicing diabetes care (example: I feel I do not have enough time to counsel diabetic patients on drug intake and compliance) and the impact of each factor on hindering the behavior (example: Lack of time makes it more difficult for me to counsel patients on drug intake and compliance). Each of the two variables was scored using a 5-point Likert scale ranging from strongly disapprove (5) to strongly approve (1). Responses to each pair of items were multiplied together and an indirect measure of perceived behavioral control (i.e., a belief-based measure of perceived behavioral control) was subsequently constructed by summing these products across all beliefs. The range of possible scores was 5 to 125.