Research in Social and Administrative Pharmacy j (2015) j–j

Original Research

An explanatory model of community pharmacists’ support in the secondary prevention of cardiovascular disease Hanni P. Puspitasari, B.Sc.(Pharm.), M.Phil., Ph.D.(c)a,b,*, Daniel S.J. Costa, B.Sc.(Hons.), Ph.D., M.P.H.c, Parisa Aslani, B.Pharm.(Hons.), M.Sc., Ph.D., Grad.Cert.Ed.Stud.(Higher Ed.)a, Ines Krass, B.Pharm., Dip.Hosp.Pharm., Grad.Dip.Educ.Studies(Health Ed), Ph.D.a a

Faculty of Pharmacy, The University of Sydney, Pharmacy and Bank Building A15, Sydney, NSW 2006, Australia b Fakultas Farmasi, Universitas Airlangga, Jalan Dharmawangsa Dalam, Surabaya, Jawa Timur 60286, Indonesia c School of Psychology, The University of Sydney, Lifehouse Building C39Z, Sydney, NSW 2006, Australia

Abstract Background: Community pharmacists have faced ongoing challenges in the delivery of clinical pharmacy services. Various attitudinal and environmental factors have been found to be associated with the provision of general clinical pharmacy services or services which focus on a specific condition, including cardiovascular disease (CVD). However, the interrelationship and relative influence of explanatory factors has not been investigated. Objective: To develop a model illustrating influences on CVD support provision by community pharmacists. Methods: Mail surveys were sent to a random sample of 1350 Australian community pharmacies to investigate determinants of CVD support provision. A theoretical model modified from the Theory of Planned Behavior (TPB) was used as a framework for the survey instrument. Structural equation modeling was used to determine how pharmacists’ attitudes and environmental factors influence CVD support. Results: A response rate of 15.8% (209/1320) was obtained. The model for CVD support provision by community pharmacists demonstrated good fit: c2/df ¼ 1.403, RMSEA ¼ 0.047 (90% CI ¼ 0.031–0.062), CFI ¼ 0.962, TLI ¼ 0.955 and WRMR ¼ 0.838. Factors found to predict CVD support included: two attitudinal latent factors (“subjective norms of pharmacists’ role in CVD support” and “pharmacists’ perceived responsibilities in CVD support”) and environmental factors i.e. pharmacy infrastructure (documentation and a private area), workload, location; government funded pharmacy practice programs; and pharmacists’ involvement with Continuing Professional Development and attendance at CVD courses. Conclusions: Pharmacists’ attitudes appeared to be the strongest predictor of CVD support provision. The TPB framework was useful in identifying “subjective norms” and “pharmacists’ beliefs” as key constructs * Corresponding author. Faculty of Pharmacy, The University of Sydney, Pharmacy and Bank Building A15, Sydney, NSW 2006, Australia. Tel.: þ61 2 9114 1159; fax: þ61 2 9351 4451. E-mail addresses: [email protected], [email protected] (H.P. Puspitasari). 1551-7411/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.sapharm.2015.04.008

2

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

of community pharmacists’ attitudes. Community pharmacies would be able to provide such an advanced clinical service if they strongly believed that this was an acknowledged part of their scope of practice, had adequate infrastructure and employed sufficient numbers of pharmacists with appropriate and relevant knowledge. Ó 2015 Elsevier Inc. All rights reserved. Keywords: Australia; Community pharmacists; Cardiovascular disease; Attitudes; Environmental factors

Introduction Community pharmacy is a unique industry wherein professional pharmacy services and patient-centered health care may be delivered within the retail environment.1 The literature, however, has reported ongoing challenges faced by community pharmacists in reorienting their practice to embrace delivery of professional pharmacy services; as a result, many maintain a predominant focus on traditional product supply and dispensing services.1–4 Facilitators of practice change in community pharmacies have been explored using frameworks of organizational change strategies and work system models that are commonly used in the business literature.5–16 In the context of community pharmacy, good relationships with general practitioners (GPs) and other health care professionals, high patient expectation, the availability of remuneration, appropriate pharmacy infrastructure, adequate staff, as well as external support and assistance from other pharmacists, professional organizations and the government have been identified as essential to the practice change implementation.5–13,15,16 Because pharmacists’ attitudes toward community pharmacy practice change have also been reported as key influences,17–27 behavioral theories such as the Theory of Reasoned Action, the Theory of Planned Behavior (TPB), the Theory of Trying and the Theory of Goal-Directed Behavior have been applied as frameworks in pharmacy practice research.28–38 The use of a theoretical framework in pharmacy practice studies serves to enrich the value and interpretability of research findings.39 In numerous studies, some or all of the TPB constructs have been shown to predict community pharmacists’ behavioral intention to provide general clinical services, not specific to a condition.31– 33,35,37 Compared to the “pharmacists’ attitudes” and “perceived behavioral control” constructs, the “subjective norm” construct has consistently been found to be the strongest predictor.31,32,35

Pharmacists’ behavioral intention has also been shown to predict the delivery of general clinical services.37 In other studies, some behavioral constructs were found to be directly associated with provision of pharmacists’ clinical services.28,30 The term “perceived behavioral control” in the TPB literature represents pharmacists’ perceptions of their ability, the ease/difficulty, and the degree of control over the performance of behaviors.28–37 In this case, the behaviors refer to clinical pharmacy services. Indeed, pharmacists’ clinical knowledge and skills in communication could facilitate the provision of such general clinical services.6,12,16,17,19,22,40–43 These findings have highlighted the need for appropriate and relevant training and continuing pharmacy education for community pharmacists.6,12,16,17,21,24,40–43 The TPB “subjective norm” construct may be defined as pharmacists’ perceptions of others’ beliefs about their behaviors.28,29,31–37 This could include beliefs of patients, GPs, pharmacy managers, colleagues, professional organizations, and the government of pharmacists performing clinical services.34,35,37 In fact, pharmacists have frequently reported patients’ and GPs’ negative attitudes toward, poor knowledge of, and low expectations of their service provision as significant barriers.20,40–47 Furthermore, unfavorable attitudes have adversely affected the pharmacist-GP collaborative practice,20,29,41,48–52 a key to providing clinical pharmacy services.4 Many researchers have concluded that the TPB is a useful theoretical framework to predict levels of clinical pharmacy services.28,31–33,35 However, there is contrary evidence regarding the utility of the TPB.29 Although DeMik et al found high scores for attitudes, perceived behavioral control and subjective norms, there was no significant correlation between these TPB scores and levels of pharmacy services.29 Social desirability bias may be a possible explanation for this, whereby respondents gave desired or acceptable answers that were inconsistent with their personal

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

attitudes, preferences and beliefs.29 Moreover, the scales in TPB instruments have been worded globally, lacking items referring to specific components of clinical pharmacy services.28,29,35 Additionally, the TPB focuses on beliefs and attitudes with limited considerations on environmental factors that have been identified as key factors influencing pharmacists’ practice and their likelihood of adopting practice change.5–13,15,16 In Australia, community pharmacists have demonstrated capacity to provide various clinical services in prevention and management of cardiovascular disease (CVD), asthma, sleep disorders and mental health.10,19,53–55 With regard to CVD, community pharmacists have demonstrated a potential role in both primary and secondary prevention, i.e. reducing cardiovascular risk factors (hypertension, diabetes, dyslipidaemia) and preventing recurrences of CVD such as coronary heart disease.55 Based on a recent Australian study, the scope of community pharmacists’ support in the secondary prevention of CVD included counseling and monitoring about the disease, medicines, and lifestyle modifications.56 The determinants of these activities were classified into “people” (influenced by attitudes and beliefs of pharmacists, clients and GPs), “environment” (including pharmacy environment, support from professional organizations, and government policies), and “outcome” factors (e.g. client health

3

outcomes, job satisfaction, and pharmacy investment).57 Although the study57 produced results which corroborate the findings of previous research exploring factors influencing community pharmacists’ delivery of clinical services in general,5–13,15–31,35–37 little is known about the extent to which these factors affect community pharmacists’ activities in CVD secondary prevention. Therefore, the objectives of the current study were to determine how: (1) pharmacists’ attitudes to delivering CVD support and (2) environmental factors, influence the provision of CVD care in the community pharmacy setting. A structural equation modeling approach was used to test the hypothesized model.

Methods Theoretical background A theoretical model was developed from the TPB58 with additional variables addressing environmental factors, and used as a framework for a survey instrument (Fig. 1). The findings of a qualitative study on community pharmacists’ services in caring for clients with established CVD56,57 were used to develop core constructs for the survey instrument. The core construct “attitudes toward pharmacists’ support” was derived from pharmacists’ beliefs and outcome evaluations, the core construct “perceived

Fig. 1. Theoretical model illustrating factors influencing community pharmacists’ support to clients with CVD, modified from the Theory of Planned Behavior.58

4

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

behavioral control” was derived from pharmacists’ control beliefs and power, and the core construct “subjective norms” was derived from perceived clients’ and GPs’ beliefs as well as pharmacists’ motivation to comply with clients’ and GPs’ expectations. Additionally, environmental factors as previously described: “pharmacy environment” (including infrastructure and workload) and “government policies” (referring to structured programs and remuneration)57 were proposed to have a direct association with the provision of CVD support, while the

availability of training for pharmacists was hypothesized to predict pharmacists’ perceived behavioral control. Survey instrument development Based on the theoretical model (Fig. 1), a four-section instrument was developed (Fig. 2, Preliminary Survey Instrument). A question about community pharmacists’ intention to provide CVD support was included in Section A. Sixteen items were listed to assess respondents’ behaviors in CVD support, using a four-point scale

Fig. 2. Survey instrument development process.

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

reflecting the proportion of clients for whom they have provided CVD support in the previous four weeks. A four-point scale was also applied to measure the frequency of working with other health care professionals in relation to CVD support. In Section B, 28 attitudinal items were included (based on three TPB constructs: attitudes toward pharmacists’ support, perceived behavioral control, and subjective norms), using a fivepoint, Likert ¼ type scale. The last two sections comprised questions about pharmacy and pharmacist characteristics. The preliminary version of the instrument was distributed to 12 community pharmacists to examine face and content validity. These pharmacists were asked to comment on the appropriateness of the questions and response options as well as the layout of the instrument. Following the feedback, minor revisions were made to the survey instrument. Pilot study The revised survey instrument was tested with 85 community pharmacists in New South Wales (NSW), Australia (Fig. 2, Pilot study). Almost all respondents (98.8%) indicated their intention to provide CVD support. As this item was not deemed as a measure of CVD support, this item was excluded from further analyses. Based on a multiple linear regression analysis, the three attitudinal constructs (attitudes toward pharmacists’ support, perceived behavioral control, and subjective norms) were not found to predict the provision of CVD support. Therefore, a principal components analysis (PCA) of the attitudinal items was utilized to re-examine the underlying latent components influencing CVD support. Following PCA (the KMO Measure of sampling adequacy index of 0.785, Bartlett’s test of sphericity c2(378) ¼ 985.9, P ! 0.001), four latent components were identified, indicating that the items were clustered based on attitudes to specific activities, i.e. counseling and monitoring about aspects of CVD management (cardiovascular conditions, medicines and lifestyle modifications). Subsequently, a decision was made to retain 13 attitudinal items, modify six items, and add one item for the national survey instrument. Meanwhile, items representing pharmacists’ behaviors in CVD support were simplified to include seven core activities. To minimize recall bias, respondents were asked to indicate their CVD support in the two weeks prior to survey completion. No changes were made to the last two sections (Fig. 2, National Survey).

5

Survey sample A sample of Australian community pharmacies was randomly selected stratified by state/territory based on lists obtained from Pharmacy Councils in each state and territory, supplemented by Yellow Pages telephone directory. NSW community pharmacies that had been invited to participate in the pilot study were excluded. The sample size was determined using two approaches, i.e. the standard error of proportions equation59 and the ratio of subjects to items.60 For the former approach, based on an estimation of 16% of pharmacies providing advanced cardiovascular services,61 at a 5% degree of precision, a minimum number of 206 pharmacies was required. For the latter approach, a minimum sample of 200 pharmacies was calculated based on 10 cases for each of the 20 attitudinal items to be factor analyzed. As a minimum sample of 206 satisfied both approaches, with an estimation of 15% response rate based on the pilot study, a sample size of 1350 was determined. Data collection Each survey package consisting of a selfadministered questionnaire with a cover letter and a reply paid envelope was sent to each selected pharmacy in March 2014. The pharmacist-onduty was asked to complete the questionnaire. Each questionnaire was marked with a code to allow for a reminder mailing(s). Approximately four and eight weeks after the initial mailing, another survey package was sent to nonrespondents. To encourage participation, all respondents who completed the questionnaire were entered into a draw for a chance to win one of five AUD$100 gift cards. Ethics approval was obtained from the University’s Human Research Ethics Committee. Data analysis All data were coded and entered into a database in SPSS 21.0.62 Frequency distributions were compiled and examined for responses to all questions. Univariate and multivariate analyses were used to identify pharmacy/pharmacist characteristics that predict dependent variables.63 The dependent variables were categorized into primary and secondary variables. The primary dependent variable was the provision of CVD support. There were two secondary dependent variables: the frequency of working with GPs

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

6

and the level of pharmacies’ involvement in providing enhanced services. The results of univariate and multivariate analyses showed that pharmacy location (metropolitan vs. rural), infrastructure (a private area and documentation), workload (i.e. number of prescriptions dispensed, number of pharmacists and dispensary assistants on duty), and pharmacists’ resource adequacy (referring to Continuing Professional Development (CPD) points and completion of CVD course(s)) were found to predict both

secondary dependent variables. The secondary dependent variables and documentation predicted the primary dependent variable.63 Table 1 presents these variables. In the current study, pharmacy/ pharmacist characteristics and attitudinal items were classified as independent variables. Exploratory factor analysis (EFA) was conducted to model the 20 attitudinal items using maximum likelihood (ML) estimation. Internal consistency of latent factors was assessed using Cronbach’s alpha.

Table 1 Variables included in the model (excluding attitudinal items) Variables Dependent variables Primary

Secondarya

Themes

Response options

Provision of CVD supportb: (1) Educating about cardiovascular condition(s) (2) Counseling on clients’ cardiovascular medicines (3) Advising on lifestyle modifications (4) Monitoring medicine-related problems (5) Monitoring for changes in lifestyle (6) Ongoing monitoring clinical parameters (7) Making therapeutic recommendations

1 ¼ yes, 0 ¼ no (for each support, a score “0” was given to respondents with missing data)

Frequency of working with general practitioners

1 ¼ nil, 2 ¼ 1–2 contacts, 3 ¼ 3–4 contacts, 4 ¼ 5 or more contacts (a score “1” was given to respondents with missing data) 1 ¼ yes, 0 ¼ no (for each service, a score “0” was given to respondents with missing data)

Provision of enhanced pharmacy servicesc: (1) Chronic disease management (2) MedsCheck and diabetes MedsCheck services (3) Home medicines review (by accredited HMR pharmacist(s) who worked at the pharmacy) (4) Healthy lifestyle support (i.e. smoking cessation, weight management) Independent variables Pharmacy-related factors

Pharmacist-related factors

a b c

Infrastructure: (1) A private area (2) Documentation system(s)

1 ¼ yes, 0 ¼ no 1 ¼ yes, 0 ¼ no

Workload: (1) Number of prescriptions dispensed (2) Number of pharmacists (3) Number of dispensary assistants

1 ¼ yes, 0 ¼ no 1 ¼ yes, 0 ¼ no 1 ¼ yes, 0 ¼ no

Location (based on the pharmacy’s postcode, referring to the Australian Standard Geographic Classification64) Resource adequacy: (1) CPD points in the last 12 months (2) An educational course(s) about CVD in the last 12 months

1 ¼ metropolitan, 2 ¼ rural

Open response 1 ¼ yes, 0 ¼ no

Each secondary dependent variable was treated as an independent variable for the primary dependent variable. The sum of respondents’ score for seven items (a total score range of “0” to “7”). The sum of respondents’ score for four items (a total score range of “0” to “4”).

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

Subsequently, the structural model of the attitudinal latent factors combined with the findings of univariate and multivariate analyses63 were used to guide the development of an empirically testable model to describe influences on the provision of CVD support by community pharmacists. This model was tested using structural equation modeling (SEM) Mplus 6.1. To deal with the categorical nature of pharmacy/pharmacist characteristics, the weighted least squares with mean and variance (WLSMV) estimation procedure was used.65 The model was specified using theta parameterization and the fit of the model was assessed using the following statistics: the normed chi-square, which is the ratio of chi-square to its degrees of freedom (c2/df), the root mean-square error of approximation (RMSEA), the comparative fit index (CFI), the Tucker–Lewis index (TLI), and the weighted root mean-square residual (WRMR).66 Relative chi-square determines the fit of the data to the model adjusted for the complexity of the model. Adequate fit is obtained when c2/df % 2.66,67 RMSEA with 90% confidence level (CI) assesses absolute fit of the model. For categorical data, good model fit is obtained when RMSEA !0.06.66,68 Both CFI and TLI, which are incremental fit indexes, measure the proportionate improvement in fit. Recommended cut-off points for CFI and TLI are 0.95.68 For the WRMR, a value of !0.90 is preferable.66 Finally, factor loadings and modification indices (MIs) were examined to identify variables representing mis-specified parameters so that the model fit could be improved.65,66 Results are reported following guidelines as described by Schreiber.66

Results Of the 1350 mailed questionnaires, 30 were undeliverable and 209 were usable, giving a 15.8% response rate. Most of the participating community pharmacies were from metropolitan areas (53.1%), independent pharmacies (52.6%), at least in 25% of their business hours per week employed two or more pharmacists (59.3%) and dispensary assistants (49.8%), and dispensed at least 150 prescriptions per day (51.7%). More than 80% of respondents reported that a private area and documentation were available. Most respondents (53.6%) were female and 83.7% had completed undergraduate pharmacy training in Australia. The respondents’ median age was 32 (IQR 28–47) years. The majority (80.9%)

7

reported that they have not completed courses about CVD in the last 12 months. The median CPD points they have gained in the last 12 months was 45 (IQR 40–60). With regards to the provision of CVD support, most respondents reported that they provided at least five activities (23.9%) and offered at least three enhanced pharmacy services (28.7%). In relation to activities in supporting clients with CVD, the majority (86.2%) reported at least one contact with a GP. Exploratory factor analysis Extracting 20 attitudinal items using ML estimation resulted in a KMO value of 0.864, and the Bartlett’s test of sphericity reached statistical significance (c2(190) ¼ 1334.1, P ! 0.001). However, inspection of the correlation matrix for these items revealed a number of items with coefficients !0.3. Moreover, several items yielded communalities !0.3. Based on the scree plot, a two-factor solution was considered most appropriate. Items were further evaluated for inclusion and exclusion based on factor loadings, correlations, and the interpretability of the item in relation to the extracted factor. After careful examination, the factor structure indicated that six items made up Factor 1 and three items made up Factor 2. The two factors were found to explain 50.5% of the variance, and the Cronbach’s alpha coefficients were above 0.70, indicating good internal consistency (Table 2). Five out of six items in Factor 1 were derived from the TPB construct “subjective norms”. Therefore, Factor 1 was labeled as “subjective norms of pharmacists’ role in CVD support.” Because all three items in Factor 2 were derived from the TPB construct “pharmacists’ beliefs,” Factor 2 was then labeled “pharmacists’ perceived responsibilities in CVD support.” Table 3 presents correlations, means and standard deviations of the attitudinal items. Structural equation modeling The structural model was developed by including the two attitudinal latent factors and all potential influences on the provision of CVD support (Table 1). Specifically, the provision of CVD support was expected to have: direct relationships with the frequency of working with GPs, the level of enhanced pharmacy services, documentation, and the attitudinal latent factor “subjective norms of pharmacists’ role in CVD

8

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

Table 2 Distribution and factor loadings of attitudinal items included in the model Factors and attitudinal items

SAg and Ag n (%)a

Factor 1: Subjective norms of pharmacists’ role in CVD support F1.1: GPs expect me to monitor 28 my clients for changes in lifestyle needed to help reduce their cardiovascular risk (n ¼ 205) F1.2: GPs believe that I should 32 monitor my clients’ clinical parameters (e.g. blood pressure) (n ¼ 206) F1.3: My clients believe that 38 ongoing monitoring of their cardiovascular condition(s) is my responsibility (n ¼ 206) F1.4: GPs believe that 75 counseling my clients with CVD about lifestyle modifications is part of my professional role (n ¼ 206) F1.5: My clients expect me to 88 explain lifestyle changes needed to prevent further cardiovascular events (n ¼ 206) F1.6: Monitoring my clients’ 63 clinical parameters (e.g. blood pressure) is my responsibility (n ¼ 206) Factor 2: Pharmacists’ perceived responsibilities in CVD support 63 F2.1: Assisting my clients with CVD to manage their body weight is my responsibility (n ¼ 206) 84 F2.2: It is my responsibility to monitor my clients for changes in lifestyle needed to improve cardiovascular health (n ¼ 206) F2.3: It is my responsibility to 109 counsel my clients about all aspects of CVD management (n ¼ 206)

Neutral n (%)a

SDg and Dg n (%)a

Cronbach’s Percentage Factor alpha of variance loadingsb Factor 1 Factor 2 0.836

44.49

(13.6) 69 (33.7) 108 (52.7)

0.800

0.017

(15.5) 76 (36.9)

98 (47.6)

0.708

0.102

(18.4) 83 (40.3)

85 (41.3)

0.682

0.036

(36.4) 69 (33.5)

62 (30.1)

0.583

0.119

(42.7) 66 (32)

52 (25.2)

0.582

0.146

(30.6) 77 (37.4)

68 (32)

0.484

0.289

0.768

5.97

(30.5) 79 (38.3)

64 (31.1)

0.039

0.866

(40.8) 71 (34.5)

51 (24.7)

0.038

0.727

(52.9) 55 (26.7)

42 (20.3)

0.271

0.421

SAg ¼ strongly agree, Ag ¼ agree, Dg ¼ disagree, SDg ¼ strongly disagree. The purpose of giving bold in the table was only highlighting explanations provided in the text under sub-heading “Explanatory factor analysis.” a Valid percent (excluding missing data). b Extraction method: Maximum Likelihood, Rotation method: Oblimin with Kaiser Normalization (Rotation converged in 6 iterations).

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

9

Table 3 Correlations, means, and standard deviations (SD) of attitudinal items Correlation F1.1 F1.1 F1.2 F1.3 F1.4 F1.5 F1.6 F2.1 F2.2 F2.3

0.471 0.458 0.473 0.561 0.462 0.407 0.440 0.521

F1.2

0.384 0.417 0.222 0.430 0.209 0.223 0.225

F1.3

0.438 0.433 0.509 0.338 0.291 0.276

F1.4

0.336 0.447 0.446 0.378 0.362

F1.5

0.299 0.455 0.335 0.456

F1.6

0.350 0.487 0.455

F2.1

0.628 0.520

Meana

SD

2.56 2.63 2.76 3.06 3.25 2.96 2.98 3.17 3.40

0.88 0.92 0.86 0.96 0.92 0.98 0.89 0.91 0.96

F2.2

0.462

F1 ¼ factor 1 (subjective norms of pharmacists’ role in CVD support), F2 ¼ factor 2 (pharmacists’ perceived responsibilities in CVD support). a Based on a five-point Likert scale (“1” ¼ strongly disagree, “2” ¼ disagree, “3” ¼ neutral, “4” ¼ agree, and “5” ¼ strongly agree).

support”; indirect associations with the attitudinal latent factor “pharmacists’ perceived responsibilities in CVD support” and independent variables. After running a model as specified, the model indices indicated poor fit: c2/df ¼ 2.917, RMSEA ¼ 0.099 (90% CI ¼ 0.088–0.110), CFI ¼ 0.679, TLI ¼ 0.616 and WRMR ¼ 1.453, with MIs suggested a direct path from the level of enhanced pharmacy services to the frequency

of working with GPs. The refined model (Fig. 3) had good fit: c2/df ¼ 1.403, RMSEA ¼ 0.047 (90% CI ¼ 0.031–0.062), CFI ¼ 0.962, TLI ¼ 0.955 and WRMR ¼ 0.838. Table 4 shows standardized and unstandardized regression coefficients and squared standardized coefficients of variables in the model. The provision of CVD was directly associated with the attitudinal latent factor “subjective norms of pharmacists’ role in

Fig. 3. Structural model illustrating factors influencing pharmacists’ support to clients with CVD. For clarity, error terms have been excluded in this figure. Values next to lines are standardized regression coefficients (bold lines represent significant !0.01, solid lines represent significant !0.05, and dashed lines are not significant). Italicized values are squared standardized coefficients. Items F1.1 and F2.2 had their loadings constrained to 1 to ensure the model was identified, so no significance test was conducted for these loadings.

Standardized regression coefficient F1

F2

IS1

IS2

EPS

FGP

0.79

SUP

F1

F2

0.42

0.76 0.53 0.64 0.68 0.62 0.69 0.04

IS1

IS2

FGP

SUP 0.47, 0.10*

0.00 0.08* 0.11* 0.10* 0.10* 0.11*

0.58 0.28 0.40 0.46 0.39 0.48

0.26

0.22, 0.08y

0.76 0.73 0.71

EPS

0.98, 0.20* 1.00, 0.54, 0.70, 0.79, 0.69, 0.82,

0.22

R2

Unstandardized regression coefficient, standard error

0.05, 0.09

0.29, 0.10y

1.10, 0.26* 1.00, 0.00 0.94, 0.20* 0.32

0.05, 0.05

0.04 0.31 0.11 0.23 0.10 0.12 0.34 0.51

0.04 0.09

0.39

0.58 0.53 0.50

0.10 0.18 0.26

0.00, 0.03 0.29, 0.02, 0.66, 0.00, 0.00, 0.48, 0.69,

0.08y 0.05 0.26z 0.00 0.00 0.16y 0.16*

0.00, 0.00 0.00, 0.00

0.32, 0.07*

0.13, 0.12 0.17, 0.08z 0.29, 0.08*

0.05 0.10 0.50 0.25 0.42

Statistically significant values: *P value !0.001, yP value !0.01; zP value !0.05. F1 ¼ factor 1 (subjective norms of pharmacists’ role in CVD support), F2 ¼ factor 2 (pharmacists’ perceived responsibilities in CVD support), IS1 ¼ infrastructure 1 (private area), IS2 ¼ infrastructure 2 (documentation), EPS ¼ enhanced pharmacy services, FGP ¼ frequency of working with GPs, SUP ¼ provision of CVD support, WL1 ¼ workload 1 (number of prescriptions dispensed), WL2 ¼ workload 2 (number of pharmacists), WL3 ¼ workload 3 (number of dispensary assistants), LOC ¼ location, RA1 ¼ resource adequacy 1 (CPD points), RA2 ¼ resource adequacy 2 (CVD courses).

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

F1 F1.1 F1.2 F1.3 F1.4 F1.5 F1.6 F2 F2.1 F2.2 F2.3 WL1 WL2 WL3 LOC RA1 RA2 IS1 IS2 EPS FGP SUP

10

Table 4 Regression coefficients (standardized and unstandardized with standard error) and squared standardized coefficients (R2)

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

CVD support” (P ! 0.001) and the frequency of working with GPs (P ! 0.001). The level of enhanced pharmacy services predicted CVD support either directly (P ¼ 0.044) or indirectly through the frequency of working with GPs (P ! 0.001). The attitudinal latent factor “pharmacists’ perceived responsibilities in CVD support” was found to predict the frequency of working with GPs (P ¼ 0.003) and the availability of private area (P ¼ 0.005). A private area predicted the level of enhanced pharmacy services (P ¼ 0.003), along with some independent variables: documentation (P ! 0.001), number of pharmacists (P ¼ 0.001) and pharmacy location (P ¼ 0.011). Several independent variables (i.e. the number of prescriptions dispensed, of dispensary assistants on duty, of CPD points, and pharmacists’ attendance at CVD courses) with weak coefficients were retained in the model because they contributed to the overall model fit.

Discussion The results of structural equation modeling demonstrated the interrelationships of community pharmacists’ attitudes and environmental factors on the provision of CVD secondary prevention. The TPB framework was valuable for identification of “subjective norms” and “pharmacists’ beliefs” as key constructs of community pharmacists’ attitudes. In the current study, the two constructs were labeled as “subjective norms of pharmacists’ role in CVD support” and “pharmacists’ perceived responsibilities in CVD support.” The former construct was directly associated with CVD support provision, while the latter construct was predicted to influence pharmacists in creating a favorable pharmacy environment (referring to infrastructure) for the delivery of CVD support. Other pharmacy-related factors such as workload and location along with pharmacist-related factors (referring to their involvement with CPD and attendance at CVD courses) were also found to predict CVD support provision. Pharmacists’ attitudes toward CVD support Mirroring evidence from several pharmacy practice research using the TPB framework,28,31,32,35,38 “subjective norms” were a strong predictor of pharmacists’ behavior, in this case in providing CVD care. Looking at items that made up this construct, it may be interpreted that GPs’ and clients’ beliefs and expectations about

11

pharmacist’ counseling on lifestyle modifications and monitoring on clients’ changes in lifestyle and cardiovascular conditions influenced pharmacists’ performance on CVD support. Interestingly, one item derived from the TPB construct “pharmacists’ beliefs” (Fig. 1) was also clustered in the “subjective norms” construct of the structural model (Fig. 3). This indicates that pharmacists themselves may be less supportive of their role in advanced health care services such as CVD secondary prevention, which is consistent with the findings of previous studies.18–21,23–26 Thus, beliefs and expectations of pharmacists themselves as well as GPs and clients should become a focus in the development of advanced pharmacy services. As pointed out in the introduction to this paper, working with GPs is key to providing clinical services,4 while accommodating a private area is a prerequisite for assuring confidential delivery of such services.63 Based on the results of the current study, these key elements are likely to be in place if pharmacists believe that it is their responsibility to provide advanced clinical services in addition to traditional product supply and dispensing services. Environmental factors Environment within community pharmacy including infrastructure (i.e. a private area and documentation) and workload were found to predict CVD care provision by community pharmacists. The relationship of documentation to CVD care may reflect the importance of documentation in supporting chronic disease management.6,17,56,57 Of the three variables used to assess pharmacy workload (Table 1), the number of pharmacists on duty was the strongest predictor. As pharmacist-related factors with respect to CPD points and their attendance at CVD courses also contributed to the model fit, the findings imply that the provision of such an advanced clinical service requires adequate number of pharmacists working with appropriate and relevant knowledge.63 Another pharmacy-related factor that was estimated to influence CVD care provision is pharmacy location. In this case, community pharmacies in rural areas were more likely to have a higher level of enhanced pharmacy services than their metropolitan counterparts,63 which may reflect their greater orientation toward delivery of health care. Community pharmacies in areas with limited access to other primary care

12

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

services play a valuable role in providing clinical services,11,69 including CVD secondary prevention. The role of enhanced pharmacy services and collaboration with GPs Based on the current study’s structural model, enhanced pharmacy services and collaboration with GPs demonstrate the operationalization of CVD care provision. In this context, support to clients with established CVD is provided when community pharmacists deliver enhanced pharmacy services and work with GPs. Additionally, bearing in mind three of four enhanced pharmacy services listed on the survey were government funded programs, enhanced pharmacy services may represent support from the government in the form of structured remunerated pharmacy practice programs, confirming the proposed theoretical model (Fig. 1). Working with GPs has been reported as a big challenge for community pharmacists in implementing patient-centered services.48–52 A direct path from the level of enhanced pharmacy services to the frequency of working with GPs as shown in the structural model (Fig. 3) may suggest that structured enhanced pharmacy services supported by the government are needed to foster pharmacist-GP collaborative practice. Limitations Although the sample of the current study was adequate for performing statistical analyses required, with a low response rate, the findings may not be transferable to the Australian community pharmacy population. The low response rate did not allow us to validate the structural model being developed. Moreover, theoretical justification made from the structural model was based on cross-sectional data; hence, any claims of causality should be interpreted with caution. Finally, based on the non-respondent analysis presented elsewhere,63 the results may be biased toward community pharmacists who provided enhanced pharmacy services. Despite this, the respondents were typical of Australian community pharmacists70 in terms of gender, age and pharmacy training country.

Conclusion To the best knowledge of the authors, this is the first study to provide a unified explanatory

model of CVD secondary prevention in the community pharmacy setting incorporating both attitudinal and environmental predictors. Pharmacists’ attitudes that illustrate their subjective judgment of GPs’ and clients’ beliefs and expectations about their CVD care provision as well as pharmacists’ own beliefs about performing CVD care appeared to be the strongest predictor of service provision. Pharmacists’ beliefs about their responsibilities in providing CVD care were found to influence pharmacists in working with GPs and in accommodating an appropriate pharmacy environment for the delivery of CVD care. Importantly, community pharmacies would be able to provide such an advanced clinical service if they employed sufficient numbers of pharmacists with appropriate and relevant knowledge. Additionally, support from the government in the form of the availability of structured remunerated enhanced pharmacy services appear to be an enabling factor in driving extended provision of health care services by pharmacists. Acknowledgments The authors thank pharmacists who participated in the pilot study and the national survey. This project was funded by Faculty of Pharmacy, the University of Sydney, Australia. References 1. Jacobs S, Ashcroft DM, Hassell K. Culture in community pharmacy organisations: what can we glean from the literature? J Health Organ Manag 2011;25: 420–454. 2. Hassell K, Seston EM, Schafheutle EI, Wagner A, Eden M. Workload in community pharmacies in the UK and its impact on patient safety and pharmacists’ well-being: a review of the evidence. Health Soc Care Community 2011;19:561–575. 3. Schafheutle EI, Seston EM, Hassell K. Factors influencing pharmacist performance: a review of the peer-reviewed literature. Health Policy 2011; 102:178–192. 4. Willink DP, Isetts BJ. Becoming ‘indispensable’: developing innovative community pharmacy practices. J Am Pharm Assoc 2005;45:376–389. 5. Doucette WR, Kreling DH, Schommer JC, Gaither CA, Mott DA, Pedersen CA. Evaluation of community pharmacy service mix: evidence from the 2004 national pharmacist workforce study. J Am Pharm Assoc 2006;46:348–355. 6. Doucette WR, Nevins JC, Gaither C, et al. Organizational factors influencing pharmacy practice change. Res Social Adm Pharm 2012;8:274–284.

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15 7. Feletto E, Wilson L, Roberts A, Benrimoj S. Flexibility in community pharmacy: a qualitative study of business models and cognitive services. Pharm World Sci 2010;32:130–138. 8. Feletto E, Wilson LK, Roberts AS, Benrimoj SI. Building capacity to implement cognitive pharmaceutical services: quantifying the needs of community pharmacies. Res Social Adm Pharm 2010;6: 163–173. 9. Feletto E, Wilson LK, Roberts AS, Benrimoj SI. Measuring organizational flexibility in community pharmacy: building the capacity to implement cognitive pharmaceutical services. Res Social Adm Pharm 2011;7:27–38. 10. Feletto E, Lui GWY, Armour C, Saini B. Practice change in community pharmacy: using changemanagement principles when implementing a pharmacy asthma management service in NSW, Australia. Int J Pharm Pract 2013;21:28–37. 11. Gadkari AS, Mott DA, Kreling DH, Bonnarens JK. Pharmacy characteristics associated with the provision of drug therapy services in nonmetropolitan community pharmacies. J Rural Health 2009;25: 290–295. 12. Gastelurrutia MA, Benrimoj SI, Castrillon CC, de Amezua MJC, Fernandez-Llimos F, Faus MJ. Facilitators for practice change in Spanish community pharmacy. Pharm World Sci 2009;31:32–39. 13. Hopp TR, Klinke BO, Sørensen EW, Herborg H, Roberts AS. Implementation of cognitive pharmaceutical services in Danish community pharmacies – perceptions of strategists and practitioners. Int J Pharm Pract 2006;14:37–49. 14. Roberts AS, Hopp TR, Sørensen EW, et al. Understanding practice change in community pharmacy: a qualitative research instrument based on organisational theory. Pharm World Sci 2003;25:227–234. 15. Roberts AS, Benrimoj SI, Chen TF, Williams KA, Hopp TR, Aslani P. Understanding practice change in community pharmacy: a qualitative study in Australia. Res Social Adm Pharm 2005;1:546–564. 16. Roberts AS, Benrimoj SI, Chen TF, Williams KA, Aslani P. Practice change in community pharmacy: quantification of facilitators. Ann Pharmacother 2008;42:861–868. 17. AbuRuz S, Al-Ghazawi M, Snyder A. Pharmaceutical care in a community-based practice setting in Jordan: where are we now with our attitudes and perceived barriers? Int J Pharm Pract 2012;20:71– 79. 18. Al Hamarneh YN, Rosenthal M, McElnay JC, Tsuyuki RT. Pharmacists’ perceptions of their practice: a comparison between Alberta and Northern Ireland. Int J Pharm Pract 2012;20:57–64. 19. Berbatis CG, Sunderland VB, Joyce A, Bulsara MK, Mills C. Enhanced pharmacy services, barriers and facilitators in Australia’s community pharmacies: Australia’s national pharmacy database project. Int J Pharm Pract 2007;15:185–191.

13

20. Bryant LJM, Coster G, Gamble GD, McCormick RN. General practitioners’ and pharmacists’ perceptions of the role of community pharmacists in delivering clinical services. Res Social Adm Pharm 2009;5:347–362. 21. Grindrod KA, Rosenthal M, Lynd L, et al. Pharmacists’ perspectives on providing chronic disease management services in the community – part I: current practice environment. Can Pharm J 2009; 142(5):234–239. 22. Latif A, Boardman HF. Community pharmacists’ attitudes towards medicines use reviews and factors affecting the numbers performed. Pharm World Sci 2008;30:536–543. 23. Mak VSL, Clark A, Poulsen JH, Udengaard KU, Gilbert AL. Pharmacists’ awareness of Australia’s health care reforms and their beliefs and attitudes about their current and future roles. Int J Pharm Pract 2012;20:33–40. 24. Rosenthal M, Grindrod KA, Lynd L, et al. Pharmacists’ perspectives on providing chronic disease management services in the community – part II: development and implementation of services. Can Pharm J 2009;142(6):284–288. 25. Rosenthal M, Austin Z, Tsuyuki RT. Are pharmacists the ultimate barrier to pharmacy practice change? Can Pharm J 2010;143(1):37–42. 26. Rosenthal MM, Breault RR, Austin Z, Tsuyuki RT. Pharmacists’ self-perception of their professional role: insights into community pharmacy culture. J Am Pharm Assoc 2011;51:363–367. 27. Urbonas G, Jakusovaite I, Savickas A. Pharmacy specialists’ attitudes toward pharmaceutical service quality at community pharmacies. Medicina (Kaunas) 2010;46:686–692. 28. Amin MEK, Chewning B. Predicting pharmacists’ adjustment of medication regimens in Ramadan using the theory of planned behavior. Res Social Adm Pharm 2015;11:e1–e15. 29. DeMik DE, Vander Weg MW, Lundt ES, Coffey CS, Ardery G, Carter BL. Using theory to predict implementation of a physician-pharmacist collaborative intervention within a practice-based research network. Res Social Adm Pharm 2013;9): 719–730. 30. Farris KB, Schopflocher DP. Between intention and behavior: an application of community pharmacists’ assessment of pharmaceutical care. Soc Sci Med 1999;49:55–66. 31. Fleming ML, Barner JC, Brown CM, Shepherd MD, Strassels S, Novak S. Using the theory of planned behavior to examine pharmacists’ intention to utilize a prescription drug monitoring program database. Res Social Adm Pharm 2014;10:285–296. 32. Gavaza P, Brown CM, Lawson KA, Rascati KL, Wilson JP, Steinhardt M. Examination of pharmacists’ intention to report serious adverse drug events (ADEs) to the FDA using the theory of planned behavior. Res Social Adm Pharm 2011;7:369–382.

14

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

33. Gavaza P, Fleming M, Barner JC. Examination of psychosocial predictor of Virginia pharmacists’ intention to utilize a prescription drug monitoring program using the theory of planned behavior. Res Social Adm Pharm 2014;10:448–458. 34. Hagemeier NE, Murawski MM, Lopez NC, Alamian A, Pack RP. Theoretical exploration to Tennessee community pharmacists’ perceptions regarding opioid pain reliever abuse communication. Res Social Adm Pharm 2014;10:562–575. 35. Herbert KE, Urmie JM, Newland BA, Farris KB. Prediction of pharmacist intention to provide Medicare medication therapy management services using the theory of planned behavior. Res Social Adm Pharm 2006;2:299–314. 36. Kennelty KA, Chewning B, Wise M, Kind A, Roberts T, Kreling D. Barriers and facilitators of medication reconciliation processes for recently discharged patients from community pharmacists’ perspectives. Res Social Adm Pharm; 2015. in press. 37. Odedina FT, Hepler CD, Segal R, Miller D. The pharmacists’ implementation of pharmaceutical care (PIPC) model. Pharm Res 1997;14):135–144. 38. Pradel FG, Obeidat NA, Tsoukleris MG. Factors affecting pharmacists’ pediatric asthma counseling. J Am Pharm Assoc 2007;47:737–746. 39. Brown CM. Conceptualizing research. In: Aparasu RR, ed. Research Methods for Pharmaceutical Practice and Policy. London: Pharmaceutical Press; 2011. 40. Blake KB, Madhavan SS. Perceived barriers to provision of medication therapy management services (MTMS) and the likelihood of a pharmacist to work in a pharmacy that provides MTMS. Ann Pharmacother 2010;44:424–431. 41. Hatah E, Braund R, Duffull S, Tordoff J. General practitioners’ perceptions of pharmacists’ new services in New Zealand. Int J Clin Pharmacol 2012; 34:364–373. 42. Ibrahim A, Scott J. Community pharmacists in Khartoum State, Sudan: their current roles and perspectives on pharmaceutical care implementation. Int J Clin Pharmacol 2013;35:236–243. 43. Tordoff J, Chang SY, Norris PT. Community pharmacists’ perceptions of services that benefit older people in New Zealand. Int J Clin Pharm 2012;34: 342–350. 44. Cavaco AM, Sousa Dias JP, Bates IP. Consumers’ perceptions of community pharmacy in Portugal: a qualitative exploratory study. Pharm World Sci 2005;27:54–60. 45. Peterson GM, Jackson SL, Hughes JD, Fitzmaurice KD, Murphy LE. Public perceptions of the role of Australian pharmacists in cardiovascular disease. J Clin Pharm Ther 2010;35:671–677. 46. Sabry NA, Farid SF. The role of clinical pharmacists as perceived by Egyptian physicians. Int J Pharm Pract 2014;22:354–359.

47. Wilbure K, Beniles A, Hammuda A. Physician perceptions of pharmacist roles in a primary care setting in Qatar. Global Health 2012;8:12. 48. Alkhateeb FM, Unni E, Latif D, Shawaqfeh MS, Al-Rousan RM. Physician attitudes toward collaborative agreements with pharmacists and their expectations of community pharmacists’ responsibilities in West Virginia. J Am Pharm Assoc 2009;49: 797–800. 49. Kucukarslan S, Lai S, Dong Y, Al-Bassam N, Kim K. Physician beliefs and attitudes toward collaboration with community pharmacists. Res Social Adm Pharm 2011;7:224–232. 50. Rubio-Valera M, Jove AM, Hughes CM, GuillenSola M, Rovira M, Fernandez A. Factors affecting collaboration between general practitioners and community pharmacists: a qualitative study. BMC Health Serv Res 2012;12:188. 51. Van C, Costa D, Abbott P, Mitchell B, Krass I. Community pharmacist attitudes towards collaboration with general practitioners: development and validation of a measure and a model. BMC Health Serv Res 2012;12:320. 52. Van C, Costa D, Mitchel B, Abbott P, Krass I. Development and validation of a measure and a model of general practitioner attitudes toward collaboration with pharmacists. Res Social Adm Pharm 2013;9:688–699. 53. Fuller JM, Wong KK, Grunstein R, Krass I, Patel J, Saini B. A comparison of screening methods for sleep disorders in Australian community pharmacies: a randomized controlled trial. PLoS One 2014;9:e101003. 54. O’Reilly CL, Wong E, Chen TF. A feasibility study of community pharmacists performing depression screening services. Res Social Adm Pharm 2015 May-Jun;11(3):364–381. 55. George J, McNamara K, Stewart K. The roles of community pharmacists in cardiovascular disease prevention and management. Australas Med J 2011;4(5):266–272. 56. Puspitasari HP, Aslani P, Krass I. Australian community pharmacists’ awareness and practice in supporting secondary prevention of cardiovascular disease. Int J Clin Pharmacol 2013;35: 1218–1228. 57. Puspitasari HP, Aslani P, Krass I. Challenges in the care of clients with established cardiovascular disease: lessons learned from Australian community pharmacists. PLoS One 2014;9(11):e113337. 58. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process 1991;50:179–211. 59. Rao PSRS. Sampling Methodologies with Applications. Boca Raton: CRC Press; 2000. 60. DeVellis RF. Scale Development: Theory and Applications. 3rd ed. Thousand Oaks: SAGE Publications; 2012. 61. Oladapo AO, Rascati KL. Review of survey articles regarding medication therapy management (MTM)

Puspitasari et al. / Research in Social and Administrative Pharmacy j (2015) 1–15

62.

63.

64.

65.

services/programs in the United States. J Pharm Pract 2012;25:457–470. Ho R. Handbook of Univariate and Multivariate Data Analysis and Interpretation with SPSS. Boca Raton: CRC Press; 2006. Puspitasari HP, Aslani P, Krass I. The influence of pharmacy and pharmacist characteristics on the provision of secondary prevention of cardiovascular disease. Int J Clin Pharm; 2015. in press. Australian Government Department of Health. Australian standard geographic classification – remoteness area (ASGC-RA). Available at: http://www. health.gov.au/internet/otd/publishing.nsf/Content/loca tor#; Accessed 16.01.14. Bryne BM. Structural Equation Modeling with Mplus: Basic Concepts, Applications, and Programming. New York: Taylor & Francis Group; 2012.

15

66. Schreiber JB. Core reporting practices in structural equation modeling. Res Social Adm Pharm 2008;4: 83–97. 67. Bollen K. Structural Equations with Latent Variables. New York: John Wiley; 1989. 68. Hu L, Bentler PM. Cutoff criteria for fit indexed in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 1999; 6:1–55. 69. Hag JD, Stratton TP. Patient care services in rural Minnesota community pharmacies. J Am Pharm Assoc 2010;50:508–516. 70. Australian Institute of Health and Welfare. Allied Health Workforce 2012. National Health Workforce Series No. 5. Cat. No HWL 51. Canberra: Australian Institute of Health and Welfare; 2013.