Development and validation of the patient trust in community pharmacists (TRUST-Ph) scale: Results from a study conducted in Thailand

Development and validation of the patient trust in community pharmacists (TRUST-Ph) scale: Results from a study conducted in Thailand

Available online at www.sciencedirect.com Research in Social and Administrative Pharmacy 4 (2007) 272–283 Development and validation of the patient ...

185KB Sizes 0 Downloads 32 Views

Available online at www.sciencedirect.com

Research in Social and Administrative Pharmacy 4 (2007) 272–283

Development and validation of the patient trust in community pharmacists (TRUST-Ph) scale: Results from a study conducted in Thailand Surachat Ngorsuraches, Ph.D.a,*, Sanguan Lerkiatbundit, Ph.D.a, Shu Chuen Li, Ph.D.b, Charoen Treesak, Ph.D.c, Rachnida Sirithorn, M.S.a, Montree Korwiwattanakarn, M.S.a a

Department of Pharmacy Administration, Faculty of Pharmaceutical Sciences, Prince of Songkla University, Hatyai, Songkhla 90112, Thailand b Department of Pharmacy, Faculty of Science, National University of Singapore, Science Drive 4, Singapore 117543 c Social and Administrative Pharmacy & Clinical Pharmacy, Faculty of Pharmacy, Srinakharinwirot University, Ongkharak, Nakornnayok 26120, Thailand

Abstract Background: The quality of the pharmacist-patient relationship has been examined in various perspectives, for example patient satisfaction. Trust is another concept within which the quality of the relationship might be examined and is critical in contemporary pharmacy practice. Objective: To develop and validate a scale to measure patient trust in community pharmacists. Methods: A 5-dimension conceptual model of trust, which includes fidelity, competence, honesty, confidentiality, and global trust, was originally used for scale development. Candidate items were generated and revised using expert reviews, focus group discussion, and think aloud method in the first phase. The items and their revisions were tested in 2 successive phases. Data were collected from 2 convenience samples of 400 each in Songkhla and Yala provinces of Thailand. Factor analysis and item analysis were used to determine dimensions and refine items of the trust scale. Internal consistency and construct validity of the scale were determined. Results: At first, exploratory factor analysis showed that the most interpretable solution consisted of 2 factors labeled Benevolence and Technical Competence. The Technical Competence had unsatisfactory internal consistency (Cronbach’s alpha ! 0.80). An improvement was made in the later phase and another dimension, Communication, was identified. The final 30-item scale, namely TRUST-Ph, with 3 dimensions had good internal consistency (Cronbach’s alpha ¼ 0.84-0.91). The scale exhibited a strong positive association with satisfaction with pharmacy services (r ¼ 0.70, P ! .001). It also showed positive associations between the level of patient trust and other patient-pharmacist aspects. Conclusions: The developed scale (TRUST-Ph) to measure patient trust in community pharmacists had relatively high validity and reliability. It had 3 dimensions, which were Benevolence, Technical Competence, and Communication. The TRUST-Ph scale can be potentially used as a measure of

* Corresponding author. Tel./fax: 66 74 428167. E-mail address: [email protected] (S. Ngorsuraches). 1551-7411/08/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.sapharm.2007.10.002

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

273

patient-reported outcome for community pharmacist services. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Scale development; Trust scale; Trust in Pharmacist; TRUST-Ph

‘‘Let him not deceive himself by trusting what is worthless, for he will get nothing in return.’’ Job 15:31

Introduction Since the advent of clinical pharmacy in the 1960s, the philosophy of pharmacy practice has shifted from one that is product-oriented toward one that is more patient-oriented. Pharmacists play a key role in providing patient care. They are not only drug experts, but are also comprehensive pharmaceutical care providers to the patients. Furthermore, they are usually the last health care professionals whom the patients communicate with before leaving a health care facility. Hence, the quality of the pharmacist-patient relationship is one of the essential contributors to a successful patient-care plan and needs to be assessed. The pharmacist-patient relationship is multifaceted, and as a consequence, many variables can be examined as indicators of relationship quality. Patient satisfaction often has been used to assess the quality of relationship between pharmacist and patient and it also serves as an outcome measurement in patient care. However, trust recently has been recognized as another indicator of the quality of interpersonal relationship between health care providers and patients.1 Trust is distinct from satisfaction. Satisfaction is based on past experience, whereas trust is a psychological state that is primarily future-oriented.2 Nevertheless, trust and satisfaction are closely related. If a patient trusts a health care provider, he or she is likely to be more satisfied with the provider’s service.2 In consumer surveys, pharmacists were often cited as America’s most trusted professionals.3,4 American consumers deemed the honesty and ethical standards of pharmacists to be ‘‘high’’ or ‘‘very high.’’ Methods used in these surveys were briefly described but their measurements of trust have never been discussed. One study showed that patient trust in pharmacists was 1 of a few constructs of patient satisfaction with pharmacy services.5 Trust was not separately

investigated, and its specific measures have never been developed and validated. Fundamentally, trust is the acceptance of a vulnerable situation in which the truster believes that the trustee will perform a particular action in the truster’s best interests.1 As such, the establishment of trust cannot be separated from a feeling of vulnerability. In other words, if vulnerability does not exist, trust will not be necessary. Because a feeling of vulnerability among patients is common and often unavoidable in health care, the establishment of trust in and by health care providers is not only desirable but also necessary for an effective health care system. For instance, there is evidence demonstrating that patient’s trust in the information provided to them was positively related to patient adherence to medical treatment.6-8 Additionally, these studies showed that an increase in trust was more important than an increase in satisfaction to improve the patient adherence. Studies by Thom et al7 and Safran et al9 demonstrated that trust was a strong predictor of continuity with doctors. Recently, Piette et al also reported that patient trust in doctors moderated the impact of cost pressures on patient medication adherence.10 Besides trust in physicians, trust in other health care providers has also been examined. For example, trust was identified as an important component in the nurse-patient relationship and as a precondition for patient empowerment.11 Therefore, understanding the patient trust can help doctors and other health care providers to improve their behaviors for better quality of care. No published studies evaluating the relationship of trust between pharmacists and their patients were found, even though pharmacists especially at the community level have often been the first port of call for patients seeking medical advice. On the contrary, various measures of patient trust in doctors have been developed, and there is a large amount of empirical studies measuring its dimensions, levels, and determinants.12-17 Pearson and Raeke concluded that patient trust in doctor is a multidimensional construct.18 In general, there are 5 dimensions of doctor behavior on

274

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

which patients base their trust: fidelity, competence, honesty, confidentiality, and global trust.2,16,19-22 However, some of these studies finally ended up with a 1-dimensional scale. Fidelity is caring for a patient with his or her best interests or welfare in mind. It is related to the concepts of agency or loyalty, and is composed of caring, respect, advocacy, and avoiding conflicts of interests. Aspects of competence include avoiding mistakes, making the right decisions, producing best results with good practice, and interpersonal skills. Honesty means telling the truth and avoiding intentional falsehoods, and includes avoiding outright lies, half-truths, and deception by silence or omission. Misleading a patient about treatment or lacking informed consent to encourage the patient to comply with the treatment is considered dishonesty. Confidentiality involves proper use of sensitive information to protect patient privacy. Most patients usually presume confidentiality and trust their doctors when they obtain any kind of medical care. Finally, global trust combines all other aspects of trust that do not exclusively fit in any other dimensions. In Thailand, approximately 30% of all patients seek primary care, primarily at community pharmacies, most of which are independent and private. Because of the nature of the Thai health care setting, Thai community pharmacists play similar roles to those of doctors. They can prescribe and dispense most medications, including antibiotics. When patients come to community pharmacies, pharmacists will interview them, provide diagnosis, and then select medications for them. Within this setting, the pharmacist can develop a relationship with patients similar to doctors with their patients. Therefore, the Thai health care delivery system has a setup that provides a unique opportunity for the development of a scale to measure patient trust. Hence, the objective of this study was to develop and validate a scale to measure patient trust in community pharmacists. Methods The study was divided into 3 phases: scale development, scale testing, and scale improvement. The study was approved by the ethical review board at Faculty of Pharmaceutical Sciences, Prince of Songkla University. Phase I: Scale development To develop a measurement for patient trust in pharmacists, an initial multidimensional conceptual

model was used. It was decided to begin with the multidimensional concept, even though Hall et al found the measurement for trust in doctors to be a unidimensional scale.16 The choice of this conceptual model was based on the consideration that the patient trust in pharmacists may be different from patient trust in doctors. Furthermore, its dimensions were broader and each dimension represents a specific aspect of pharmacists. Following the decision on the conceptual model, questions were generated and adapted from existing scales which have wellestablished reliability (Cronbach’s alpha coefficients O 0.80) and validity. These scales included the Trust in Physician Scale, Primary Care Assessment Survey, Patient Trust Scale, and General Trust in Medical Profession Scale, which measure the patient’s trust in doctor.6,13,14,16 Initially, 5 dimensions of trust, namely, fidelity, competence, confidentiality, honesty, and global trust, were used as scale constructs. Items from the existing scales were adapted to a pharmacy context with all items reworded from ‘‘medical care’’ or ‘‘doctor’’ to ‘‘pharmacy service’’ or ‘‘pharmacist.’’ Positively and negatively worded items were used to avoid acquiescent responses.23 Response categories were strongly agree, agree, neutral, disagree, and strongly disagree. Because the scale items were obtained from various studies in the medical care field, 3 pharmacy faculty members, as independent experts, were asked to review the items and to confirm the relevance of each item to its dimension. According to a provided working definition of each dimension, the experts were invited to comment on the ambiguity and relevancy of individual items and suggest any other relevant items that were not yet in the scale. Based on these comments and suggestions, the scale was revised. Because the existing patient trust in doctors scales were not originally developed in a Thai context, 2 focus group discussions were conducted to address dimensions or items not sufficiently covered. A total of 13 participants were drawn from general population to participate in the focus groups. An incentive of 500 Baht (US$ 1: w37 Baht) was provided to each participant. During the focus group meetings, the definition of patient trust in pharmacists was opened for discussion. The candidate trust dimensions and items were also presented to the participants to discuss whether the dimensions covered all aspects of trust, whether the items reflected the patient trust in pharmacists, and whether the items were clear and easy to understand. Information from the focus groups was used to

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

create, delete, and adjust the candidate dimensions and items. A draft of a self-administered questionnaire was developed based on the information obtained from the literature review, the 3 experts, and the 2 focus groups. In addition, demographic characteristics were included in the questionnaire, and other variables, for example satisfaction related to trust in pharmacists were measured. Satisfaction with the pharmacy service was measured by a validated patient satisfaction scale (very good, good, fair, poor, and very poor) developed by Larson et al24 along with a 5-point scale. An overall satisfaction question was added to the scale. Each subject was asked about his or her past disagreement or dispute with a pharmacist, whether he thought of a pharmacist when he had a health problem, whether he had a preferred pharmacist to consult, whether he asked for the pharmacist’s service, and how often he followed the pharmacist’s recommendations. Response categories for the questions were ‘‘yes,’’ ‘‘no,’’ and ‘‘do not know,’’ except for the last question that used a 5-point scale (always, often, sometimes, rarely, and never). A message of informed consent was included in the questionnaire. The drafted questionnaire was reviewed for clarity and content validity by 3 pharmacy faculty members and 7 practicing pharmacists. Also, a qualitative interview, using ‘‘think aloud’’ technique with 5 people who had ever used pharmacy services before, was used to validate the whole questionnaire. From the information of interview, further revisions were made. The questionnaire was then piloted with a convenience sample of 30 people. Results from the pilot study were used to refine the questionnaire. The final questionnaire comprised 8 pages with 4 different sections: general information of the respondent, the past relationship between the respondent and pharmacist, a 40item trust scale, and a 21-item satisfaction scale. Phase II: Scale testing The sample for the scale validation was a convenience sample of subjects who had used a community pharmacist’s services once or more within the previous 3 months in Songkhla, a southern province of Thailand. The study participants were older than 18 years and could communicate in the Thai language. They were recruited at various public venues, such as shopping malls and bus stations. DeVellis suggested that 300 subjects might be an adequate number for scale validation

275

and fewer than 300 subjects might be sufficient for a pool of about 20 items.23 MacKeigan and Larson used 350 subjects to test a 45-item patient satisfaction scale.25 A sample of 10 patients per scale item was used in this scale validation. Therefore, a total sample of 400 patients was used in this study because the trust scale comprised 40 items. The questionnaire was directly distributed with an incentive of 30 Baht for questionnaire completion. The individual who distributed the questionnaire was identified as a research assistant to reduce social desirability response bias. Sample characteristics were descriptively analyzed. Response means and standard deviations were calculated to determine variability and symmetry in score distributions, which indicated discriminatory power. Exploratory factor analysis with principal components extraction was used to determine the dimensions of factors underlying the patient’s trust in pharmacist. The KaiserMeyer-Olkin measure of sampling adequacy and Bartlett’s test of sphericity were used to determine whether factor analysis was appropriate for analyzing the scale.26 The number of factors selected for Promax rotation was determined by using combined criteria of Scree test, cutoff eigenvalue at 1, factor loadings greater than 0.4, and interpretability of the trial factor solution. Internal consistency was assessed by Cronbach’s alpha. Redundant items or items with a low item-scale correlation (!0.3) were deleted.25 Phase III: Scale improvement The developed scale was revised in reference to the results from phase II. Some existing items were reworded to reduce ambiguity. Several new items were added to improve scale quality, primarily involving the ‘‘Technical Competence’’ dimension. The revised scale consisted of 47 items. Three pharmacy faculty members and 3 practicing pharmacists were asked to review the items for content validity and clarity. From the experts’ comments and suggestions, another revision was made and 7 items were deleted because of their ambiguity. A total of 40 items replaced those from the previous questionnaire. The final structure of the questionnaire comprised 8 pages of 4 different sections. The revised questionnaire was distributed to a convenience sample of 400 people in Yala, another southern province of Thailand. The previous inclusion criterions of subjects still applied. Data analyses were conducted following

276

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

procedures outlined for phase II. Construct validity of the scale was examined by the Pearson correlation between the ‘‘patient trust in pharmacists’’ and ‘‘patient satisfaction with pharmacy service’’ scores. Also, validity was assessed by concurrent association between the patient’s trust and its potentially related variables. Analysis of Variance (ANOVA) was used to compare trust scores across groups categorized by how often the subject followed pharmacist’s recommendations. Similarly, a 2-sample t-test was used to compare trust scores for each question with a dichotomous response format, including past disagreement or dispute with the pharmacist, whether the subject thought of the pharmacist when he or she had a health problem, whether he or she had preferred pharmacist, and whether he or she asked for the pharmacist’s service.

Results Phase I From various scales of the patient’s trust in a doctor, a total of 36 items were adopted and adapted to pharmacist context. The first 3 experts suggested refining some ambiguous items and recommended adding 2 new items about pharmacists considering patient’s safety and confidentiality to cover the aspect of trust more completely. All 38 items were presented to the 2 focus groups after the general discussion about the definition of patient trust in pharmacists. Based on the concerns of focus group participants, some candidate items were further refined. The focus group participants expressed trust-related concerns about treatment outcomes, pharmacist’s thoroughness, his or her compliance with law or ethics, and dispensing time. Four items addressing the participants’ concerns were added. After the revised questionnaire was reviewed by the second expert panel of 3 pharmacy faculty members and 7 practicing pharmacists, 2 items were deleted because they measured conditional trust (on how much information searched and thinking time) instead of general trust. In addition, the results of the qualitative interview using the ‘‘think aloud’’ technique suggested that an instruction specifying that the respondents should answer the questions based on their overall feeling about the pharmacist should be emphasized at the beginning of the questionnaire. The results of the pilot test with a convenience sample of 30 people showed that the respondents required

approximately 14 minutes for completing the questionnaire. Neither critical ambiguities nor difficulties with the questions were found. Phase II Table 1 shows respondent characteristics from both phases II and III. In phase II, the average age was approximately 30 years. Most of them were female (60.8%) and had either a high school degree or equivalent (52.8%) or a bachelor degree (31.3%). Approximately 70% of them earned less than 10,000 Baht per month and almost 30% of them worked in private sectors. Most respondents had either universal health insurance or social security health insurance coverage. In Thailand, these insurance schemes do not cover medication Table 1 Overall respondents’ characteristics (N ¼ 400) Number (%) Variables

Phase II

Age Mean  SD

Phase III

30.02  9.0 27.24  9.0

Sex Female

243 (60.8) 299 (74.8)

Highest education completed Elementary school or lower 64 (16.1) 31 (7.7) High school or equivalent 211 (52.8) 275 (68.8) University degree 125 (31.3) 94 (23.5) Occupation Government officials Private sector Others

20 (5.0) 41 (10.3) 106 (26.5) 72 (18.0) 274 (68.5) 287 (71.7)

Monthly income (Baht, US$ 1 w37 Baht) !10,000 281 (70.25) 336 (84.0) R10,000 119 (29.75) 64 (16.0) Type of health insurance coverage Universal coverage scheme 156 Social security scheme 152 Civil servant medical benefit 29 scheme Others 63 Perceived health status Very good Good Fair Poor Very poor

32 233 125 9 1

(39.1) 186 (46.5) (38.0) 139 (34.7) (7.3) 57 (14.3) (15.75) 18 (4.5) (8.0) 29 (7.2) (58.3) 201 (50.3) (31.3) 154 (38.5) (2.3) 13 (3.3) (0.3) 3 (0.7)

Number of community pharmacy visits in the last 3 mo 1-3 288 (72.0) 343 (85.7) O3 112 (28.0) 57 (14.3)

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

obtained from any community pharmacy. Almost 60% of subjects considered themselves in good health, whereas more than 70% had used community pharmacy services less than 4 times in the previous 3 months. In the 40-item trust scale, no item had more than 10% missing responses. The results of the Kolmogorov-Smirnov test indicated that responses to the trust scale were normally distributed. Item means varied from 2.32 to 4.11 on the 5-point Likert scale, with a range standard deviation of 0.70-1.13. None of the item responses were concentrated in only 1 or 2 categories. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.92 and the Bartlett’s test of sphericity was 3639.17 (df ¼ 435, P ! .001). These indicated that the variables shared enough common variance to be appropriate for a factor analysis.26 The Scree test from the exploratory factor analysis with principal components extraction indicated a possible break in the eigenvalue plot at a 10-factor solution. One factor consisted of exclusively 9 unfavorably worded items, which did not fit together logically. They were then deleted to avoid the problem of method factor.27,28 A total of 31 items were examined by exploratory factor analysis. Six factors had eigenvalues above 1. However, the Scree test indicated 2 possible breaks at 2 and 5 factors. Therefore, 2-6 factors were extracted and rotated. The most interpretable solution consisted of 2 factors and accounted for 55.96% of variance. These 2 factors were labeled ‘‘Benevolence’’ and ‘‘Technical Competence.’’ Seven items, from either ‘‘Benevolence’’ or ‘‘Technical Competence’’ factor, did not achieve the factor-loading criterion of 0.40 and were eliminated before item analysis. Table 2 shows item performance of the trust scale. In the item analysis, an item that had low item-total correlation (!0.3) or was weakly correlated with the overall factor was deleted from the scale. The final patient trust scale contained only 23 items and exhibited a 2-factor structure with relatively high Cronbach’s alpha (0.901 and 0.740, respectively), which could still be improved. Phase III For this phase, the average age of respondents was slightly less than 30 years. Almost 75% of them were female and approximately 70% had a high school degree or equivalent. More than 80% of them earned less than 10,000 Baht per month and only 18% of them worked in private

277

sectors. Similar to the respondents in phase II, most respondents in phase III had either universal health insurance or social security health insurance coverage. About one half of the respondents considered themselves in good health. More than 85% of them used community pharmacy services less than 4 times in the previous 3 months. All items in the revised 40-item trust scale had no more than 10% missing responses. Similar to the previous scale, the results of the KolmogorovSmirnov test indicated that the responses to the revised trust scale were normally distributed. Item means varied from 3.33 to 3.96, with a range standard deviation of 0.65-1.02. None of the item responses concentrated in only 1 or 2 categories. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.94 and the Bartlett’s test of sphericity was 6183.69 (df ¼ 780, P ! .001). They ensured that the variables shared enough common variance to be appropriate for factor analysis.26 Initially, the Scree test from the exploratory factor analysis with principal components extraction indicated a possible break in the eigenvalue plot at 7-factor solution with 59.55% of the variance explained. Similar to the results from phase II, the Scree test indicated 2 possible breaks at 2 and 5 factors. Therefore, 2-6 factors were rotated. The most interpretable solution consisted of 3 factors with 0.59-0.62 intercorrelations and accounted for 47.42% of variance. The factors were labeled ‘‘Benevolence,’’ ‘‘Technical Competence,’’ and ‘‘Communication.’’ Five items did not achieve the factor-loading criterion of 0.40. In addition, 5 items from all 3 dimensions did not fit logically with other items loading on their own factors. Therefore, they were eliminated before item analysis. One item had dual loadings on Technical Competency and Communication factors; it had difference in loadings less than 0.1. The item was assigned to Technical Competency based on the basis of logical fit before item analysis. Table 3 shows item performance of the revised trust scale. From item analysis, no item had low item-total correlation (!0.3) or was weakly correlated with the overall factor. In conclusion, the developed patient’s trust scale contained 30 items and exhibited a 3-factor structure with relatively high Cronbach’s alpha (0.909, 0.855, and 0.837 for Benevolence, Technical Competence, and Communication, respectively). The developed scale was named ‘‘TRUST-Ph.’’ For scale validation, the total trust score was significantly correlated with total satisfaction score (r ¼ 0.704, P ! .001). Table 4 shows

278

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

Table 2 Phase II item performance Dimension

Items

Mean  SD (N)

Factor loading

Item-scale correlation

Benevolencea

Pharmacists keep your sensitive medical information private. Pharmacists would admit if a mistake was made when dispensing. Pharmacists provide you with information on all potential medical options. Pharmacists put your health above the drug costs (profits). Pharmacists thoroughly ask you about your illness. Pharmacists do not do anything illegal or unethical. Pharmacists never mislead you about anything. Pharmacists suggest you to see a doctor when needed. Pharmacists are thorough and careful. Pharmacists choose the best treatment or medications for you. You can tell pharmacists anything. You talk with pharmacists even about embarrassing issues. Pharmacists care about you more than their own personal benefits. All in all, you trust pharmacists. You always follow pharmacists’ advice. Pharmacists offer the best things for you.

3.48  0.91 (379)

0.838

0.622

3.49  1.12 (374)

0.803

0.619

3.70  0.88 (393)

0.774

0.658

3.67  0.89 (385)

0.753

0.595

3.67  0.93 3.68  1.00 3.31  0.96 4.11  0.85 3.70  0.85 3.75  0.84

(396) (380) (377) (396) (385) (392)

0.727 0.702 0.674 0.601 0.600 0.541

0.637 0.499 0.653 0.528 0.629 0.604

3.31  1.00 (388) 3.71  0.92 (389)

0.497 0.492

0.427 0.440

3.43  0.95 (380)

0.479

0.547

3.78  0.70 (394) 3.68  0.81 (398) 3.76  0.81 (392)

0.463 0.440 0.439

0.550 0.575 0.597

3.38  0.87 (389) 3.58  0.94 (393)

0.827 0.824

0.545 0.419

3.99  0.70 (398) 3.83  0.77 (395)

0.726 0.549

0.441 0.446

3.73  0.68 (394) 3.72  0.84 (396) 3.70  0.77 (395)

0.546 0.495 0.458

0.475 0.286 0.415

3.59  0.74 (390)

0.446

0.468

Technical competenceb

a b

What pharmacists tell you is always right. Pharmacists should be the persons who make decision about your medications. Pharmacists offer you good quality medications. Pharmacists put your medications and advice above everything. You are confident in pharmacists’ dispensing. Pharmacists consider your needs and put them first. Pharmacists put highest efforts on decisions about your medications. You are confident in pharmacists.

Cronbach’s alpha ¼ 0.91. Cronbach’s alpha ¼ 0.74 (after an item that had low item-total correlation (!0.3) was deleted from the scale).

comparisons between patient’s trust in pharmacist and other relationships, including past disagreement or dispute with the pharmacist, whether he or she thought of pharmacist when he or she had health problem, whether he or she had preferred pharmacist to consult, whether he or she asked for pharmacist’s service, and how often he or she followed pharmacist’s recommendations. The results show significant associations in the predicted directions in all relationships between patient and pharmacist (P ! .05), except for the association between Technical Competency dimension and whether the patient ever disagreed with pharmacist about medication. For instance, higher patient trust scores were associated with

a lower likelihood of having a disagreement with a pharmacist, with a higher likelihood of turning to a pharmacist for assistance when the patient had problems with health and medications, with a higher likelihood of having a preferred pharmacist, a higher likelihood of asking for a pharmacist’s service, and with a higher likelihood of following a pharmacist’s recommendation.

Discussion To the best of our knowledge, this study is the first formal development of the scale measuring patient trust in pharmacists. At first, a 5-

279

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283 Table 3 Phase III item performance Dimension

Items

Mean  SD (N)

Factor loading

Item-scale correlation

Benevolencea

Pharmacists do not do anything illegal or unethical. Pharmacists care about your health. Pharmacists never mislead you about anything. Pharmacists keep your sensitive medical information private. Pharmacists care about you more than their own personal benefits. Pharmacists would admit if a mistake was made when dispensing. Pharmacists put your health above the drug costs (profits). Pharmacists pay attention to your problems.d Pharmacists sincerely serve you.d Pharmacists put your benefits as first priority.d You can tell pharmacists anything. Pharmacists choose the best treatment or medications for you.

3.88  0.65 3.58  0.78 3.40  0.86 3.71  0.86

0.756 0.731 0.728 0.700

0.669 0.677 0.642 0.672

3.33  0.91 (370) 0.652

0.650

3.52  1.02 (376) 0.636

0.637

3.38  0.95 (374) 0.592

0.633

3.63  0.74 3.66  0.81 3.55  0.88 3.34  0.93 3.66  0.83

0.554 0.539 0.527 0.495 0.485

0.690 0.683 0.635 0.490 0.654

Pharmacists put highest efforts on decision about your medications. Pharmacists are experts about drugs.d Pharmacists should be the persons who make decision about your medications. Pharmacists correctly notify you how to use drugs.d Pharmacists can help you with your illness.d Pharmacists can solve your medication problems.d Pharmacists carefully dispense you medications.d You are confident in pharmacists’ dispensing. Pharmacists offer you good quality medications. What pharmacists tell you is always right.

3.69  0.79 (388) 0.710

0.633

3.94  0.77 (391) 0.683 3.59  0.92 (393) 0.662

0.608 0.477

3.94  0.68 3.61  0.85 3.75  0.72 3.82  0.80 3.68  0.74 3.83  0.74 3.52  0.84

0.618 0.605 0.584 0.562 0.520 0.445 0.421

0.627 0.522 0.591 0.558 0.598 0.554 0.463

3.88  0.65 (396) 0.737

0.548

3.80  0.76 3.82  0.81 3.78  0.70 3.78  0.81 3.76  0.73 3.89  0.75 3.96  0.89

0.561 0.565 0.598 0.661 0.554 0.531 0.536

Technical competenceb

Communicationc

a b c d

You understand what pharmacists explain about how to use drugs. Pharmacists provide you opportunity to ask questions. Pharmacists clearly write medication labels. You are confident in pharmacists’ counseling. Pharmacists are willing to talk or answer your questions. You always follow pharmacists’ advice. Pharmacists use easy language for counseling. Pharmacists suggest you to see a doctor when needed.

(396) (380) (364) (364)

(389) (386) (373) (380) (388)

(391) (392) (391) (394) (390) (392) (389)

(393) (394) (390) (396) (392) (397) (391)

0.734 0.636 0.611 0.599 0.474 0.472 0.440

Cronbach’s alpha ¼ 0.91. Cronbach’s alpha ¼ 0.86. Cronbach’s alpha ¼ 0.84. New items, compared with phase II result.

dimension concept of trust applied from the measure of patient’s trust in medical professionals was used. Several attempts, incorporating expert and patient opinions, were used to qualitatively adjust and refine the scale. The scale had acceptable discriminatory power because the scores were not distributed within only 1 or 2 categories, which reflected its ability to separate various levels

of patient trust in pharmacists. The results also provided support for the assumption of a normal distribution in later statistical analysis. Initially, the unfavorably worded items were included for validity reasons. However, the factor analysis showed that they formed a separate dimension, even though they were not logically tied together. One reason for this could be that the unfavorably

280

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

Table 4 Comparison of Patient’s Trust Scores across Other Patient-Pharmacist Relationships Benevolence N

Mean  SD

Technical competency N

Mean  SD

1. Have you ever disagreed with pharmacists about your medications? Yes 38 40.87  9.50 41 36.41  6.94 No 216 43.60  6.86 239 37.87  4.92 P ¼ .035 P ¼ .102

Communication

Total trust score

N

Mean  SD

N

Mean  SD

44 244

29.68  5.10 31.30 þ 3.88 P ¼ .016

38 207

107.45  19.27 112.92  13.73 P ¼ .036

2. Have you thought of pharmacists when you have problems with your health or medications? Yes 199 44.06  6.93 222 38.29  5.15 232 31.66  3.86 192 No 72 40.79  7.58 77 35.36  5.23 82 29.66  4.03 65 P ¼ .001 P ! .001 P ! .001

114.20  14.18 105.83  14.52 P ! .001

3. Have you had preferred pharmacists? Yes 110 44.86  6.59 No 183 41.43  6.46 P ! .001

119 206

4. Have you asked for pharmacists to service? Yes 202 43.59  7.40 216 No 84 41.19  5.94 96 P ¼ .009 5. How often have you Rarely or sometimes Often Always

followed pharmacists’ 66 40.77  6.81 95 42.98  6.08 154 43.64  8.09 P ¼ .028

38.84  5.48 36.45  4.95 P ! .001

126 215

32.10  3.50 30.20  4.26 P ! .001

103 177

115.82  13.47 108.36  15.01 P ! .001

38.15  4.87 35.93  5.19 P ! .001

231 99

31.51  4.02 29.83  3.53 P ! .001

194 81

113.58  14.39 107.02  12.48 P ! .001

84 103 182

28.82  3.74 31.11  3.92 31.52  4.11 P ! .001

64 90 147

105.33  13.37 111.60  13.01 113.59  15.79 P ¼ .001

recommendations? 85 35.47  4.58 97 37.52  4.74 169 38.17  5.57 P ! .001

Mean of summated scores (according to number of items in the scale, full scores for Benevolence, Technical Competency, Communication, and Total Trust Scores are 60, 50, 40, and 150, respectively). Independent t-test with 1-tailed P-value, except question 5 which is ANOVA.

worded items in Thai language made no sense to the respondents. Therefore, these items were defined as a method factor and were deleted from the scale.27,28 From the factor selection criteria, the exploratory factor analysis showed an ambiguous number of underlying dimensions in phase II. The most interpretable number of dimensions was 2, labeled as Benevolence and Technical Competence, containing 16 and 7 items, respectively. Both dimensions included items of all 5 domains from the original concept: fidelity, competence, confidentiality, honesty, and global trust. There were twice as many items in the pharmacist trust scale compared with the scale of trust in the medical profession. However, it was not considered laborious as the respondents spent less than 20 minutes to complete all questionnaire sections, including demographic data and satisfaction measure. The Benevolence dimension had high internal validity (Cronbach’s alpha O 0.80), whereas the internal validity of the Technical Competence dimension was only moderate. A

reason could be that the Technical Competence dimension either had too few items or really had low internal consistency. Also, some items in each dimension were still ambiguous, which suggested that the developed scale could be improved. In phase III, revision of some existing items to reduce ambiguity and addition of several new items to improve the quality of the developed scale initially resulted in a revised scale with a higher number of items, but the revised scale was reduced to the same length as the previous version after expert review. Similar to the previous scale, the revised scale in this phase had an acceptable discriminatory power and could separate various levels of patient’s trust in pharmacist. Item scores were normally distributed which was good for later statistical analysis. Similar to phase II, the exploratory factor analysis showed an ambiguous number of underlying dimensions. However, the most interpretable solution suggested 3 factors: ‘‘Benevolence,’’ ‘‘Technical Competence,’’ and ‘‘Communication.’’

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

The results were consistent with findings from the studies of trust in the medical profession. The 3 dimensions of the developed scale covered all 5 domains of the original concept. Basically, the Benevolence dimension represented 4 domains of fidelity, honesty, confidentiality, and general trust. Technical Competence and Communication reflected whether pharmacists can make right decisions and produce best results with good practice. A pharmacist’s Technical Competence and Communication skills seem to be the key performance indicators that patients used to determine the relationship because they appeared in both the developed patient trust scale and the patient satisfaction scale developed by Larson et al.24 During this phase of scale improvement, a new dimension ‘‘Communication’’ was identified. Evidence suggests that communication was a part of the technical competence dimension of patient’s trust in doctor.2 Because patients could not directly evaluate their doctors’ competency at points of services, they assessed the doctors’ communication ability instead. Previous scales measuring patient trust in doctors showed that the communication ability has always been included in the Technical Competence dimension. However, this study showed that items measuring the domain of communication were separately retained in the final patient trust in pharmacist scale. One reason for this could be that a pharmacist’s communication skill is a key performance of pharmacy practice and it strongly encourages the patient-pharmacist relationship. Therefore, from the patient’s perspective, it is possible that the Communication dimension is important enough to influence trust in pharmacists and it is separated from general technical skills. Finally, the scale in this phase contained 7 items more than the previous version. The addition of these items had improved the internal validity of the scale causing all scale dimensions to have Cronbach’s alpha values more than 0.80. The revised scale also demonstrated good construct validity by exhibiting a strong positive association with satisfaction. Intuitively, a patient who was satisfied with a pharmacist would be more likely to trust the pharmacist. The results were consistent with previous studies in the medical profession-patient relationship, showing that trust and satisfaction were closely related but that they had different time orientations.16 The revised scale also showed consistency between the level of the patient’s trust and other patient-pharmacist relationships. The results were intuitive

281

and affirmed the validity of the scale. Only the association between the past disagreement between patient and pharmacist and the Technical Competence dimension was not statistically significant, even though the respondents who used to have a past disagreement scaled their trust in a pharmacist’s technical competence lower than the ones who never had any disagreement. One reason for this could be that the meaning of a past disagreement or dispute with the pharmacist might be ambiguous. Also, the disagreement experienced by the patient might have varying levels. A low level of disagreement between the patient and pharmacist might not have much effect on the patient’s trust in a pharmacist’s technical competence. The study results should be viewed within the context of several limitations. First, our findings were based on a relatively small convenience sampling of populations with specialized characteristics. Also, there were some missing values. With respect to factor analysis, after the list-wise deletion of missing values, only 292 and 284 cases remained in phases II and III, respectively. Even though Comrey stated that a sample size of 200 should be adequate in ordinary factor analysis which involves no more than 40 items, larger and more diverse samples are encouraged in further studies.29 Second, because people naturally choose to remain with the pharmacists they trust most, a selection bias might occur in the study because this group of people would tend to rate high trust in pharmacists. Third, a limitation of this study is the problem of social desirability bias because the questionnaires were directly handed and collected from the samples. However, the questionnaire distributor was identified as a research assistant at the beginning to reduce the social desirability bias. Furthermore, no significant relationship between socioeconomic indicators and the differential of patient trust in pharmacists was found. Therefore, the social desirability bias was likely not crucial in this study. Fourth, the number of items at the beginning of phases II and III might increase the possibility of artifact factors. Because this study was the first formal development of a trust scale in pharmacy area and it was intended to ensure scale completeness, a number of items were then initially included. However, the resulting factors were intuitive and provided good validity. Fifth, most questions determining the patient-pharmacist relationships were singleitem measures, thus limiting their validations. However, they were successfully used in a previous

282

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283

study.16 Another limitation is that trust in pharmacists may differ significantly from 1 country to another because of cultural and religious differences. The scope of pharmacy practice and legal authority of pharmacists with respect to dispensing medications in Thailand can create different perspectives of patient trust in pharmacists. Because of these limitations, further study is needed to refine the scale measuring patient trust in community pharmacists and to test the scale in a broader population and across countries. To reduce burden and cost from completing the scale, an abbreviated version of the scale should be developed.30 Also, the application of the developed scale should be examined, such as the association between patient’s trust in pharmacists and pharmacy service, or pharmacist intervention, or patient adherence, or continuity with pharmacist. Because patient trust has future orientation, which is another facet of the pharmacist-patient relationship, it will provide a more complete picture of the relationship and will be useful for improving the quality of care in pharmacy practice.

3.

4.

5.

6.

7.

8.

9.

Conclusions A scale for the measurement of patient trust in community pharmacists was developed. The final scale, TRUST-Ph, had 3 dimensions: ‘‘Benevolence,’’ ‘‘Technical Competence,’’ and ‘‘Communication.’’ It had relatively good validity, and appears to be able to measure community pharmacist performance and the quality of the patient-pharmacist relationship. If patient trust in community pharmacists plays a similar role to patient trust in doctors in terms of relationships with patient adherence and promoting continuity,6-8 its use may contribute positively to pharmacy practice and may have utility as a measure of patient-reported outcomes, or as a predictor of such outcomes.

10.

11. 12. 13.

14.

15.

Acknowledgment This research project was supported by a grant from Prince of Songkla University through Contract No. 2840.

16.

17.

References 1. Thom DH, Hall MA, Pawlson LG. Measuring patients’ trust in physicians when assessing quality of care. Health Aff (Millwood) 2004;23:124–132. 2. Hall A, Dugan E, Zheng B, Mishra AK. Trust in physicians and medical institutions: what is it? Can

18.

it be measured? And does it matter? Milbank Q 2001;79:613–639. [The Gallup Poll web site] The Gallup Organization. Honesty/ethics in professions. Available at: http:// www.galluppoll.com/content/?ci¼1654&pg1; December 8–10, 2006. Accessed 05.06.07. [Ethics Newsline Web site] The Institute for Global Ethics, Medical Practitioners. Top list of most trusted professions. Available at: http://www.global ethics.org/newsline/members/issue.tmpl?articleid¼ 11219922225391; November 22, 1999. Accessed 05. 06.07. Volume CI, Farris KB, Kassam R, Cox CE, Cave A. Pharmaceutical care research and education project: patient outcomes. J Am Pharm Assoc 2001;41:411– 420. Safran DG, Kosinski M, Alvin R, et al. The primary care assessment survey: test of data quality and measurement performance. Med Care 1998;36:728–739. Thom DH, Ribisl KM, Stewart AL, Luke DA. Further validation and reliability testing of the trust in physician scale. Med Care 1999;37:510–517. Trachtenberg F, Dugan E, Hall MA. How patients’ trust relates to their involvement in medical care. J Fam Pract 2005;54:344–352. Safran DG, Montgomery JE, Chang H, Murphy S, Rogers W. Switching doctors: predictors of voluntary disenrollment from a primary physician’s practice. J Fam Pract 2001;50:130–136. Piette JD, Heisler M, Krein S, Kerr EA. The role of patient-physician trust in moderating medication nonadherence due to cost pressures. Arch Intern Med 2005;165:1749–1755. Johns JL. A concept analysis of trust. J Adv Nurs 1996;24:76–83. Thom DH, Campbell B. Patient-physician trust: an exploratory study. J Fam Pract 1997;44:169–176. Kao AC, Green DC, Davis NA, Koplan JP, Cleary PD. Patients’ trust in their physicians. J Gen Intern Med 1998;13:681–686. Kao AC, Green DC, Zaslavsky AM, Koplan JP, Cleary PD. The relationship between method of physician payment and patient trust. JAMA 1998;280: 1708–1714. Hyman MR. An improved scale for assessing patients’ trust in their physician. Health Mark Q 2001; 19:23–42. Hall MA, Camacho F, Dugan E, Balkrishnan R. Trust in the medical profession: conceptual and measurement issues. Health Serv Res 2002;37: 1419–1439. Keating NL, Green DC, Kao AC, Gazmararian JA, Wu VY, Cleary PD. How are patient’s specific ambulatory care experiences related to trust, satisfaction, and considering changing physicians? J Gen Intern Med 2002;17:29–39. Pearson SD, Raeke LH. Patient’s trust in physicians: many theories, few measures, and little data. J Gen Intern Med 2000;15:509–513.

Ngorsuraches et al./Research in Social and Administrative Pharmacy 4 (2008) 272–283 19. Emanuel EJ, Dubler NN. Preserving the physicianpatient relationship in the era of managed Care. JAMA 1995;273:323–329. 20. Mechanic D, Schiesinger M. The impact of managed care on patients’ trust in medical care and their physicians. JAMA 1996;275:1693–1697. 21. Gray BH. Trust and trustworthy care in the managed care era. Health Aff (Millwood) 1997;16:34–49. 22. Mechanic D. The functions and limitations of trust in the provision of medical care. J Health Polit Policy Law 1998;23:661–686. 23. DeVellis RF. Scale Development, Theory and Applications. In: In: Applied Social Research Methods Series, Vol. 26. Thousand Oaks, CA: Sage Publications; 1991. p. 88–90. 24. Larson LN, Rovers JP, MacKeigan LD. Patient satisfaction with pharmaceutical care: update of a validated instrument. J Am Pharm Assoc 2002;42:44–50. 25. MacKeigan LD, Larson LN. Development and validation of an instrument to measure patient

26.

27.

28.

29.

30.

283

satisfaction with pharmacy services. Med Care 1989;27:522–536. Grimm LG, Yarnold PR. Reading and Understanding Multivariate Statistics. Washington, DC: American Psychological Association; 1998. p. 104. Grick P, Fiske ST, Mladinic A, et al. Beyond prejudice as simple antipathy: Hostile and benevolent sexism across cultures. J Pers Soc Psychol 2000;79: 763–775. Grick P, Lameiras M, Fiske ST, et al. Bad but bold: Ambivalent attitudes toward men predict gender inequality in 16 nations. J Pers Soc Psychol 2004;86: 713–728. Comrey AL. Factor analysis methods of scale development in personality and clinical psychology. J Consult Clin Psychol 1988;56:754–761. Dugan E, Trachtenberg F, Hall MA. Development of abbreviated measures to assess patient trust in a physician, a health insurer, and the medical profession. BMC Health Serv Res 2005;5:64.