Exploring the formation of patient satisfaction in rural community telepharmacies

Exploring the formation of patient satisfaction in rural community telepharmacies

Research Exploring the formation of patient satisfaction in rural community telepharmacies Daniel Friesner and David M. Scott Received August 25, 20...

514KB Sizes 0 Downloads 32 Views

Research

Exploring the formation of patient satisfaction in rural community telepharmacies Daniel Friesner and David M. Scott

Received August 25, 2008, and in revised form December 22, 2008. Accepted for publication December 31, 2008.

Abstract Objectives: To apply a previously validated patient satisfaction questionnaire within rural community telepharmacies in an effort to identify the underlying factors determining satisfaction with those services and to assess whether the latent structure(s) of patient satisfaction varies depending on delivery mode or communityspecific factors. Design: Descriptive, nonexperimental, cross-sectional study. Setting: Eight rural community telepharmacy sites (seven in North Dakota and one in Minnesota) in fall 2005. Patients: 400 potential participants in rural communities (response rate 24% [n = 96]) whose primary community pharmacy is a telepharmacy site. Intervention: Patients visiting a pharmacy to have at least one prescription filled were asked to complete a survey and mail responses to the investigators. The survey contained 37 questions, the first 20 of which were adapted from a well-established, validated survey instrument. Main outcome measure: Patient satisfaction with rural community telepharmacy services; patient responses to 20 questions in the survey were used as main outcome variables. Results: Applying factor analysis to the data yielded a single dimension of patient satisfaction. Conclusion: A previous application of this instrument in a traditional community pharmacy setting yielded two interrelated latent constructs (“friendly explanation” and “managing therapy”). Our analysis suggests that the formation of patient satisfaction in rural community telepharmacies is much simpler in that patients form a single construct exhibiting high mean and median values. Anecdotal evidence from the literature suggests that the formation of a single construct reflects patients’ desire to retain a point of access to health care in their communities. Keywords: Telepharmacy, factor analysis, health care services, rural setting. J Am Pharm Assoc. 2009;49:509–518. doi: 10.1331/JAPhA.2009.08110

Journal of the American Pharmacists Association

www.japha.org

Daniel Friesner, PhD, is Associate Professor of Pharmacy Practice, and David M. Scott, MPH, PhD, is Professor of Pharmacy Practice, North Dakota State University, Fargo. Correspondence: Daniel Friesner, PhD, Department 2660, P.O. Box 6050, North Dakota State University, Fargo, ND 58108-6050. Fax: 701-231-7606. E-mail: daniel.friesner@ndsu. edu Disclosure: The authors declare no conflicts of interest or financial interests in any product or service mentioned in this article, including grants, employment, gifts, stock holdings, or honoraria. Acknowledgments: To Charles Peterson and Howard Anderson Jr. for critically reviewing the manuscript to improve its readability and practical policy implications; to Ann Rathke for providing detailed background information about telepharmacies in North Dakota; to Dave Peterson, Gary Boehler, Tim Weippert, and Dave Rueter for providing information about the distribution of their patients by age and gender; to the pharmacists, pharmacy owners, and patients who participated in the study; and to the JAPhA associate editor (Donald Harrison) and several anonymous reviewers for valuable comments that improved the manuscript. Funding: Supported by grant 4D1B TM 00051-03 from the Office for the Advancement of Telehealth, Health Resources and Services Administration, U.S. Department of Health and Human Services.

J u l /A u g 2009 • 49:4 •

JAPhA • 509

Research

Telepharmacy patient satisfaction

A

considerable portion of the recent social and administrative sciences literature has attempted to identify those aspects of pharmaceutical care that patients value.1–5 The practical value of this work to pharmacists, pharmacy administrators, and pharmaceutical entrepreneurs is both extensive and straightforward. Patient satisfaction (which is often highly correlated with, and in some cases equivalent to, the quality of care) is a highly subjective and multifaceted phenomenon.6,7 It is also a primary determinant of repeat purchases and/or use by customers and, by extension, a crucial determinant of a company’s long-term financial viability.8,9 Thus, understanding the global themes that guide the formation of satisfaction and adapting current practices to meet those demands are crucial to the survival of the firm and, from a pharmaceutical policy perspective, crucial to providing high-quality care. This is especially true for community pharmacies operating on very slim profit margins, which are usually between 2% and 4%.10 For these businesses, financial viability can only be attained by using the volume of new and repeat patients to offset these low margins. Community pharmacies operating in rural areas face even

At a Glance

Synopsis: A total of 96 patients of eight rural community telepharmacy sites (seven in North Dakota and one in Minnesota) responded to a survey assessing patient satisfaction. Most patients were very satisfied with the level of telepharmacy service. Exploratory factor analysis of survey data revealed a single dimension of patient satisfaction, which was defined as “telepharmaceutical care,” suggesting that the formation of patient satisfaction in rural community telepharmacies is much simpler compared with “traditional” community pharmacy settings in that patients form a single construct exhibiting high mean and median values. Analysis: Previous research identified two related components of patient satisfaction in a traditional community pharmacy setting (i.e., “friendly explanation” and “managing therapy”). This study’s findings implied that determining patient satisfaction is much simpler in the context of community telepharmacy, as only one latent factor was observed. Assuming that patients behave in a manner consistent with their values and that other challenges (including declining reimbursement and a shortage of available pharmacists) can be met successfully, telepharmacies have the potential to remain profitable enterprises. One study estimated that the total economic effect of telepharmacies in North Dakota was approximately $7.5 million. These telepharmacies have the potential to accumulate $500,000 in economic activity in towns of 1,000 people or fewer. Therefore, as long as patients continue to be satisfied with and use these telepharmacies, the communities in which they exist will also continue to flourish.

510 • JAPhA • 4 9 : 4 • J u l / A u g 2009

www.japha.org

greater challenges in generating volume and thus require an even greater understanding of patient satisfaction. The small (and often declining) population base11,12 makes these pharmacies more reliant on repeat purchases, which are almost exclusively tied to patient satisfaction. Many younger residents in rural communities may “bypass” the local pharmacy and seek those same services in larger, more distant communities by bundling activities (e.g., combining shopping and recreational activities with the need to fill medications) as a means to justify the extra travel costs.13 Finally, recent changes in Medicare Part D policies have resulted not only in decreased pharmacy reimbursement but also encouraged some older residents to use mail order prescription services as opposed to patronizing the local community pharmacy.14 This combination of forces, in conjunction with other factors, such as the ability to attract young pharmacists to work in rural communities, has resulted in dozens of rural community pharmacy closures during the previous few years.11,12 Without an understanding of the factors that form patient satisfaction, many more community pharmacies will continue to lose customers and ultimately be forced out of business. Researchers typically use questions from validated survey instruments to capture various facets of patient satisfaction and exploratory factor analysis (EFA)1,3 to identify whether and how those facets are related by reducing these facets to a smaller number of characteristics. One of the most well-known patient satisfaction surveys for pharmaceutical services was developed by Larson and MacKeigan.1,3,5 During the previous 2 decades, numerous studies have validated this survey in a variety of community practice settings, including traditional community pharmacies and pharmacies with mail order service.1,4 The vast majority of these studies have come to two general conclusions. First, most components of patient satisfaction, especially in a community pharmacy setting, can be reduced to a small number of latent factors (usually between two and five).1,7 Second, the number and composition of these factors have been found to vary slightly depending on the type of services offered. While the survey’s content is generally considered valid, applying the survey within the context of specific practice settings, in order to fully understand the nuances associated with the formation of patient satisfaction, is important. To date, one aspect of patient satisfaction with pharmacy services that has not been developed is community telepharmacy. Community telepharmacy is defined as a full-service community pharmacy, providing all aspects of traditional community pharmaceutical care using distance communication technology.15 Despite this relatively straightforward definition, a number of different ways to provide community telepharmacy services exist. In this study’s application of telepharmacy, licensed pharmacists work at a “central site,” which is typically a community pharmacy in a rural and/or a medically underserved area. Pharmacists use audio and visual computer equipment to supervise licensed technicians at other rural locations, which are known as “remote sites” and are usually located in very small communities between 30 to 75 miles from the central site.15–18 Several characteristics of patient satisfaction in community Journal of the American Pharmacists Association

Telepharmacy patient satisfaction

telepharmacy make it both unique and interesting to study.15–18 First, although community telepharmacy services are designed to provide the same core set of deliverables as traditional community pharmacy services, the mechanism by which those services are delivered is, in many instances, quite different. Patients may have very different perceptions about the nature and quality of counseling services delivered in person versus an electronically interactive format. Second, several means exist by which community telepharmacy services are delivered, ranging from central sites (where pharmacists are physically present) to remote sites (where pharmacists supervise technicians through audio–visual links).15–18 As such, patient satisfaction may vary based on the type of telepharmacy services offered.

Objectives We sought to apply Larson and MacKeigan’s basic patient satisfaction survey5 in rural community telepharmacy settings in an effort to empirically identify latent factors characterizing patient perceptions of quality. Consistent with previous literature, this analysis uses EFA to extract and characterize those latent factors.19,20 We also sought to contrast our findings with those from previous studies to determine whether patient satisfaction is formed consistently across community practice settings. A secondary objective was to assess whether the latent structure(s) of patient satisfaction varies depending on delivery mode or community-specific factors. Telepharmacy services can be delivered in a variety of ways, two of which are explored in this article. Patients may prefer one delivery method (which is inclusive of both the telepharmacy technology itself and the

Research

staff’s use of the technology) over the other. Similarly, community-specific characteristics, including (but not limited to) population size, age, gender, chronic disease prevalence, and sociological conditions may affect the formation of patient satisfaction. To the extent that the data allow, testing whether significant differences in patient satisfaction exist based on these characteristics is possible.

Methods Survey instrument

This study measures patient satisfaction using a 37-item survey (Appendix 1 in the electronic version of this article, available online at www.japha.org). The first 20 items are virtually identical to those developed by Larson et al.1 The scale design and ordering of these items in our survey is identical to that of Larson et al., with only small changes in wording to adapt the survey to a telepharmacy setting. The next two questions ask respondents to identify their gender and approximate age. The remaining questions (which are not used in the study) contain 13 close-ended and 2 open-ended questions addressing basic respondent demographics and patient perceptions about the general nature of telepharmacy. Surveys were coded such that, although patient anonymity was maintained, tracking which pharmacy provided services to a respondent was possible. To ensure that the survey was appropriate for use in a community telepharmacy setting, a preliminary version of the survey was pilot tested in a community telepharmacy site that is similar to (but not one of) those used in this study. Feedback from the pilot test was used to refine the survey.

Table 1. Telepharmacy and community characteristics Panel A: Telepharmacy and community (incorporated town) characteristics Community population Telepharmacy Characteristic 2004 2005 2006 1 Exclusively remote 726 713 695 2 Exclusively remote 755 749 743 3 Exclusively remote 482 478 479 4 Exclusively remote 535 531 530 5 Central and remote site 1,054 1,045 1,032 6 Central and remote site 897 866 845 7 Central and remote site 2,227 2,214 2,190 8 Central and remote site 1,849 1,845 1,816 Panel B: 2005 county-level characteristics for each telepharmacy

Telepharmacy 1 2 3 4 5 6 7 8

Total population 4,807 2,424 6,934 13,632 5,780 4,420 5,780 5,554

Percentage of population aged: ≤44 45–64 ≥65 48.72 28.73 22.55 50.17 28.88 21.29 66.09 21.21 12.69 69.20 20.88 9.91 45.35 25.22 20.12 48.12 27.44 24.43 45.35 25.22 20.12 53.94 25.01 21.05

Journal of the American Pharmacists Association

www.japha.org

Median age Years Men 44.7 44.2 31.6 26.8 40.5 44.2 40.5 39.5

Women 46.9 46.3 31.0 31.6 42.7 47.7 42.7 43.2

J u l /A u g 2009 • 49:4 •

JAPhA • 511

Research

Telepharmacy patient satisfaction

Data collection and sample

Eight pharmacies (seven in North Dakota and one in northwestern Minnesota) were chosen for the study. Several considerations guided our selection of these sites. First, these sites were among the first community telepharmacies implemented in the upper midwestern United States and, at the time this survey was administered (fall 2005), had been operating for more than 2 years. As a result, patients frequenting these pharmacies had extensive exposure to the concept of telepharmacy (as practiced in their community) and the experience necessary to provide knowledgeable responses to the survey questions. Second, these sites use one of two modes of delivery. Although these modes were not entirely distinct, they were the most commonly used in North Dakota and Minnesota. The distinction between sites was not intended to function as a set of control/test groups but rather to be representative of the telepharmacy types commonly practiced in the region. Third, although all eight pharmacies were located in rural communities, these communities vary by population, age, and gender. To illustrate this, Table 1 contains some population-, age-, and gender-related demographic information collected from the U.S. Census Bureau (www.census.gov/popest/estimates.php) for each of these communities. The Census Bureau provides total population estimates for each town over time (2004–2006) and cross-sectional estimates (in 2005, the year of our survey) by age and gender at the county level. Although data on race and ethnicity are also provided, the homogeneity of the target populations makes these data of little use and, thus, they are omitted from the analysis. Each of these communities is relatively small, ranging from a few hundred to a few thousand. These eight communities have stagnant (and slightly decreasing) populations with subtle population differences (especially the distribution by age), which may imply that each community has different pharmacy services needs. Women in these communities also tend to live as much as 3 years longer than men. Because seven of these eight pharmacies were located in geographically large counties, the populations they served were relatively isolated from other sources of outpatient pharmacy services. Although these differences may seem insubstantial, minor changes in these already small populations make supporting a traditional pharmacy extremely difficult. Thus, these communities depend on customer satisfaction to remain profitable. Four of the pharmacies were part of a small community chain. These pharmacies were all remote sites that were staffed by technicians, and all pharmacist verification and counseling was conducted via telepharmacy technology using audio–visual equipment. The remaining four sites were independently owned and jointly operated by a group of pharmacists and were equipped to function either as central or remote sites. These four pharmacies were staffed by three licensed pharmacists and several technicians who alternated across the four sites as staffing and other conditions dictated. On any given day, any one of these four sites may have been staffed in person or remotely by a pharmacist. Consequently, patients visiting these four pharmacies were likely to receive a mix of traditional and 512 • JAPhA • 4 9 : 4 • J u l / A u g 2009

www.japha.org

telepharmacy services. Because patients at all eight sites had exposure to telepharmacy-based care and because the survey asked patients to evaluate only their experiences with telepharmacy services, we expected responses across the two groups of pharmacies to be relatively homogenous, especially if the technology was used consistently across locations. As noted in Table 1, the first group of remote pharmacies was more likely to be located in smaller communities compared with the second group, which may function as either remote or central sites. Because remote sites can survive on lower volumes of prescription service (which one would expect to occur in smaller communities, holding all else constant), this result was not surprising. A total of 50 surveys were provided to each pharmacy (400 total). Pharmacists (or in the case of remote sites, pharmacy technicians) were asked to distribute the surveys to patients 18 years of age or older who spoke English and had one or more prescription filled at that location. Given the racial and ethnic homogeneity of these communities, we did not expect the exclusion of non–English-speaking patients to bias the study. However, the requirement that respondents be at least 18 years of age may have induced parents or legal guardians to respond on behalf of their children. Patients were asked to complete the survey and mail it back to the authors’ home institution in a prepaid return envelope. Patients were not asked to reveal their medication use but were asked to identify their gender and approximate age. Because the use of mail surveys is not believed to represent a threat to human participants and because participation in a voluntary survey is considered informed consent, the authors applied for and were granted exempt status from the North Dakota State University Institutional Review Board. Data analysis

All data analysis was conducted using SPSS version 16.0 (SPSS, Chicago). Descriptive statistics for each of the 20 questions were computed, including means (±SD) and medians. Distributions of responses by gender and age were also computed to characterize the typical respondent in our sample. To analyze the latent structure of patient satisfaction, this study follows the literature and uses EFA. EFA proceeds in several steps, the first of which is to determine whether data are appropriate for the technique.19,20 Several heuristic measures have been frequently used to this end. The first is the sample size, especially as it relates to the number of variables in the analysis. At a minimum, at least 50 observations should appear in the sample and the ratio of observations to variables in the analysis should be 5 to 1. If possible, at least 100 observations should appear and the ratio of observations to variables should be closer to 10 to 1. A second heuristic is the Kaiser-Meyer-Olkin measure of sampling adequacy. This is a measure bounded on the unit interval, with values above 0.9 representing samples that are excellent candidates for factor analysis and values below 0.6 describing samples that are inappropriate for factor analysis. Generally speaking, most samples used in factor analysis studies lie in the 0.7 to 0.8 range.19 A third metric is the chi-square test of significance, when applied to the correlation matrix of the variables included in the analysis (also known as the Bartlett test Journal of the American Pharmacists Association

Telepharmacy patient satisfaction

of sphericity). A fundamental assumption of factor analysis is that sufficient correlation exists across the variables to be able to combine them into a smaller number of latent factors. As such, the Bartlett test analyzes the null hypothesis that the set of correlations are jointly not significantly different from zero. Rejecting this null at conventional significance levels (we use a level of 0.05 in this study) implies that the data are appropriate for factor analysis. The second step in factor analysis is determining how to extract the factors, retain the significant factors, and subsequently rotate the significant factors.19,20 This study applies the standard practice of using principal components to extract the factors. This procedure extracts the latent factors as the eigenvalues of the data matrix and by definition extracts the same number of eigenvalues as the number of variables in the data set. The eigenvalue-greater-than-1 rule, in conjunction with the scree test, is used to retain only those factors that are significant.19,20 Finally, the varimax method is used to rotate the factors and generate factor scores. Factor scores are deemed significant only if those factor scores are greater than 0.6 in magnitude. The final step in EFA is to examine the internal consistency of the scales and, if consistent, to create scale scores representing reliable, valid, and consistent measures of patient satisfaction. Internal consistency is measured using Cronbach’s alpha. As was done in Larson et al.,1 scale scores are created by taking the mean of all variables that load significantly onto a particular factor.

Research

Robustness tests

Having used EFA to characterize patient satisfaction scores, determining whether our two groups of telepharmacies (those that are solely remote sites versus those that may be either remote or central sites) have fundamentally different levels of patient satisfaction was possible. Testing whether significant differences exist across each of the eight locations (which would account for general socioeconomic conditions specific to each community), as well as by respondent age and gender, was also possible. To test these conjectures within the context of our study, this analysis operated under the null hypothesis that mean (and median) satisfaction scores did not differ by the type of pharmacy, pharmacy location, or age or gender of participants. Under this null, pooling the data from all eight telepharmacies and using EFA to create the satisfaction measure(s) was reasonable because the null implies that no distinct differences exist (at the average) across telepharmacies or across respondent characteristics. Rejection of the null hypothesis implies that the formation of patient satisfaction was specific to particular communities, pharmacies, or age groups or by gender. Future research would then be necessary to gain a complete understanding of why these differences exist. All analyses were conducted using both parametric (oneway analysis of variance) and nonparametric (Kruskal-Wallis and sign tests for the median) hypothesis tests. In all cases, a significance level of 0.05 was used.

Table 2. Demographics of respondents to survey assessing satisfaction with rural community telepharmacy services Panel A: Respondent demographic descriptive statistics Respondents Characteristic No. (%) Gender Men 26 (27.08) Women 70 (72.92) Age (years) ≤20 1 (1.04) 21–44 6 (6.25) 45–64 45 (46.88) 65–74 23 (23.96) 75–84 19 (19.79) ≥85 2 (2.08) Panel B: Selected CI calculations and comparison with those from Larson et al.1 Current study Characteristic % (95% CI) Gender Men 27.08 (18.19­–35.97) Women 72.92 (64.03–81.81) Age (years) ≤44 7.29 (2.09–12.49) 45–64 46.88 (36.89–­56.86) ≥65 43.75 (33.83–53.67)

Characteristic Gender Men Women Age (years) ≤39 40–69 ≥70

Larson et al.1 study % (95% CI) 29.90 (25.56–34.24) 69.90 (65.55–74.25) 6.70 (4.33–­9.07) 47.90 (43.17–­52.63) 45.40 (40.68–50.11)

All estimates are calculated as proportions and subsequently transformed into percentages.

Journal of the American Pharmacists Association

www.japha.org

J u l /A u g 2009 • 49:4 •

JAPhA • 513

Research

Telepharmacy patient satisfaction

Results Of the 400 questionnaires distributed to these pharmacies, 107 were returned, 96 of which provided a complete set of information germane to this study (usable response rate 24%). Although the usable sample size (and the response rate that generated it) is somewhat low compared with general patient satisfaction studies, it is generally consistent with numerous patient satisfaction studies conducted in other areas of telehealth, including telemedicine.21 Of these 96 responses, 38 came from remote pharmacies and 58 from the central-site pharmacies. Given that the communities with remote sites are nearly 50% smaller (and thus have lower volumes of patients to fill out and return a survey) than the remaining sites, this discrepancy is not surprising. Table 2 contains a summary of respondent characteristics. Nearly 73% of respondents were women. Approximately 47% of respondents were between 45 and 64 years of age, and more than 90% were between 45 and 84 years of age. The low overall response rate, combined with the high proportion of women and elderly respondents, suggested that the small sample may have resulted from nonresponse bias among men or younger individuals. Although the absence of a validation sample made it impossible to address this issue directly, by comparing our demographic estimates with those in other published studies of patient satisfaction in community pharmacies, most notably that of Larson et al.,1 it can be addressed indirectly. To accomplish this, 95% CIs were calculated for each of our demographic groups, which are shown in Table 2. If the sample was consistent with those of other studies (which have high response rates and representative samples), then the point estimates and 95% CIs provided in previous studies for a particular demographic group should overlap with the new (rural telepharmacy) CI estimates. The analysis looks for CI overlap, as opposed to calculating test statistic and conducting formal hypothesis tests, because the two studies’ age categories were not identical; therefore, using a test statistic approach was inappropriate. Clearly, overlap of CIs occurred, suggesting (but in no way proving) that the two samples identified similar respondents. In every case, the point estimates generated by Larson et al. fall within our CI estimates. Concomitantly, the point estimates generated by this sample fell within CI estimates generated based on their results. Other studies also reported similar respondent characteristics and thus corroborate the analysis in Table 2.3,4 As a final robustness check, the owners of the remote-site pharmacies were contacted at a later date (without informing them of results) and asked to broadly identify the distribution of their patients by age and gender (D. Peterson, personal communication, November 2008). Their results (which admittedly apply to 2007–2008, as opposed to 2005 when our study was conducted) were quite similar to those presented in Table 2. More specifically, the owners reported having more female than male patients and that the percentage of patients 65 years or older ranged from 25% to just greater than 50%, depending on the location of the store. Table 3 contains the list of 20 questions and basic descriptive statistics for each question. This table shows that most patients 514 • JAPhA • 4 9 : 4 • J u l / A u g 2009

www.japha.org

were very satisfied with their level of service—a finding that is highly consistent with previous studies of patient satisfaction in community pharmacies.1,3,22 Mean scores fell between 4 and 5 (5, extremely satisfied with services rendered). The two questions with the highest mean scores were those identifying the courtesy of the staff and the overall quality of service. These questions ranked second and fifth highest in the study by Larson et al.1 SDs were also small in magnitude; in most cases, values were between 0.5 and 0.9, indicating little variation across patient perceptions. Median scores were generally above the mean, implying that the distributions were skewed to the left. In general, the heuristics discussed previously imply that the data were appropriate for factor analysis. The Kaiser-Meyer-Olkin measure was 0.951 and the Bartlett test statistic 2,171.3 (P < 0.001), both of which indicate that the data were extremely conducive to EFA. The only slightly concerning metric was that the ratio of the sample size to the number of variables was at the minimum of 5 to 1. However, given the affirmative information contained in the other three metrics, concluding that the data could be used in an EFA was reasonable. Principal component extraction yielded one primary factor (eigenvalue 14.191) that explained 70.956% of the variation in our 20 variables. The next largest factor was substantially less than 1 (0.890) and explained only 4.448% of the variation in our variables. The scree test indicated a very strong, clear break between the first and second eigenvalues as well. Further details of the extraction results are available from the lead author upon request. Having extracted this factor, generating factor scores (Table 4) was possible. It should be noted that, because only one latent factor was extracted, the choice of rotational methods was moot.19,20 All factor scores exceeded 0.7 (range 0.720–0.898). The survey questions that most closely aligned with the underlying satisfaction measure were “The pharmacist’s efforts to help you stay healthy” (0.898), “The pharmacist’s interest in your health” (0.894), and “How well that pharmacist answers your questions” (0.892). One implication from this finding was that all 20 items used from the survey were significant and thus can be used to create a single patient satisfaction metric. Given the nature of the three survey questions with the highest factor loadings, concluding that the latent satisfaction measure was truly global and not focused on a particular aspect of care, such as “managing therapy” or “friendly explanation,” was also possible. We defined this metric as “telepharmaceutical care.” After the latent factor was identified, using Cronbach’s alpha to characterize the internal consistency of the scale was possible. The value provided by Cronbach’s alpha was 0.977, indicating that an extremely high degree of internal consistency and reliability existed. Given the high mean levels of satisfaction (Table 3) and the large number of variables used in the analysis, this result was not surprising. As mentioned earlier, the latent metric was characterized by taking a simple mean across each of the 20 items. Summary statistics for the final patient satisfaction metric are provided in Table 5, panel A. Consistent with the individual satisfaction proxies, the overall metric exhibited high patient satisfaction (mean 4.464, median 4.65) and low variation (SD 0.597). Journal of the American Pharmacists Association

Telepharmacy patient satisfaction

Research

Table 3. Survey questions and descriptive statistics Survey question The professional appearance of the pharmacy The availability of the pharmacist to answer your questions The pharmacist’s professional relationship with you The pharmacist’s ability to advise you about problems that you might have with your medications The promptness of prescription drug services The professionalism of the pharmacy staff How well the pharmacist explains what your medications do The pharmacist’s interest in your health How well the pharmacist helps you manage your medications The pharmacist’s efforts to solve problems that you have with your medications The responsibility that the pharmacist assumes for your drug therapy How well the pharmacist instructs you about how to take your medications Your pharmacy services overall How well the pharmacist answers your questions The pharmacist’s efforts to help you improve your health or stay healthy The courtesy and respect shown you by the pharmacy staff The privacy of your conversations with the pharmacist The pharmacist’s efforts to ensure that your medications do what they are supposed to do How well the pharmacist explains possible adverse effects The amount of time the pharmacist offers to spend with you

Mean ± SD 4.64 ± 0.58 4.58 ± 0.61 4.51 ± 0.70

First quartile 4 4 4

% rated excellent 67.71 64.58 60.42

4.56 ± 0.59 4.57 ± 0.63 4.60 ± 0.57 4.49 ± 0.63 4.33 ± 0.82 4.32 ± 0.75

4 4 4 4 4 4

5 5 5 5 5 4

5 5 5 5 5 5

61.46 63.54 64.58 56.25 52.08 47.92

4.34 ± 0.77 4.31 ± 0.80 4.53 ± 0.70 4.66 ± 0.58 4.55 ± 0.68 4.25 ± 0.82 4.66 ± 0.61 4.33 ± 0.80

4 4 4 4 4 4 4 4

5 4.5 5 5 5 4 5 5

5 5 5 5 5 5 5 5

51.04 50.00 63.54 70.83 64.58 44.79 71.88 51.04

4.40 ± 0.75 4.33 ± 0.85 4.30 ± 0.91

4 4 4

5 5 5

5 5 5

54.17 55.21 53.13

Median Third quartile 5 5 5 5 5 5

Table 4. Factor scores regarding patient perceptions of quality Survey question The professional appearance of the pharmacy The availability of the pharmacist to answer your questions The pharmacist’s professional relationship with you The pharmacist’s ability to advise you about problems that you might have with your medications The promptness of prescription drug services The professionalism of the pharmacy staff How well the pharmacist explains what your medications do The pharmacist’s interest in your health How well the pharmacist helps you manage your medications The pharmacist’s efforts to solve problems that you have with your medications The responsibility that the pharmacist assumes for your drug therapy How well the pharmacist instructs you about how to take your medications Your pharmacy services overall How well the pharmacist answers your questions The pharmacist’s efforts to help you improve your health or stay healthy The courtesy and respect shown you by the pharmacy staff The privacy of your conversations with the pharmacist The pharmacist’s efforts to ensure that your medications do what they are supposed to do How well the pharmacist explains possible adverse effects The amount of time the pharmacist offers to spend with you Journal of the American Pharmacists Association

www.japha.org

Factor score 0.737 0.790 0.882 0.833 0.720 0.825 0.859 0.894 0.869 0.889 0.885 0.854 0.854 0.892 0.898 0.818 0.789 0.858 0.859 0.813 J u l /A u g 2009 • 49:4 •

JAPhA • 515

Research

Telepharmacy patient satisfaction

After identifying our global satisfaction measure, checking whether this variable was robust to various pharmacy, community, and respondent characteristics was possible. These tests are shown in Table 5, panels B through E. The tests failed to

reject our null hypothesis (P > 0.05), indicating a lack of significant differences in patient satisfaction across any of these factors. Although not conclusive proof, these tests indicated that the formation of patient satisfaction was generated con-

Table 5. Hypothesis tests Panel A: Basic descriptive statistics Variable Final satisfaction metric Panel B: Analysis by type of pharmacy Type of pharmacy Remote telepharmacy Central-site telepharmacy Hypothesis test One-way ANOVA (F-test) Kruskal-Wallis test (chi-square) Sign test of the median (chi-square) Panel C: Analysis by pharmacy Pharmacy no. 1 2 3 4 5 6 7 8 Hypothesis test One-way ANOVA (F-test) Kruskal-Wallis Test (chi-square) Sign test of the median (chi-square) Panel D: Analysis by respondent gender Gender Men Women Hypothesis test One-way ANOVA (F-test) Kruskal-Wallis test (chi-square) Sign test of the median (chi-square) Panel E: Analysis by respondent age Age (years) ≤20 21–44 45–64 65–74 75–84 ≥85 Hypothesis test One-way ANOVA (F-test) Kruskal-Wallis test (chi-square) Sign test of the median (chi-square)

Mean ± SD (median) 4.464 ± 0.597 (4.650) Mean ± SD (n) 4.399 ± 0.675 (38) 4.507 ± 0.541 (58) df 1, 94 1 1

Test statistic value 0.753 0.227 0.064

P 0.388 0.634 0.965

Mean ± SD (n) 4.430 ± 0.748 (10) 4.442 ± 0.422 (6) 4.396 ± 0.747 (12) 4.345 ± 0.726 (10) 4.476 ± 0.598 (21) 4.458 ± 0.557 (13) 4.750 ± 0.232 (5) 4.511 ± 0.540 (19) df 7, 88 7 7

Test statistic value 0.252 1.092 1.805

P 0.970 0.993 0.970

Mean ± SD (n) 4.567 ± 0.465 (26) 4.426 ± 0.638 (70) df 1, 94 1 1

Test statistic value 1.068 0.875 0.341

P 0.304 0.350 0.723

Mean ± SD (n) 3.950 ± NA (1) 4.192 ± 0.917 (6) 4.460 ± 0.627 (45) 4.557 ± 0.529 (23) 4.453 ± 0.532 (19) 4.464 ± 0.597 (2) df 5, 90 1 1

Test statistic value 0.548 3.352 2.092

P 0.740 0.646 0.836

Abbreviation used: ANOVA, analysis of variance; df, degrees of freedom; NA, not applicable. All two-by-two sign test results reflect a Yates Continuity Correction.

516 • JAPhA • 4 9 : 4 • J u l / A u g 2009

www.japha.org

Journal of the American Pharmacists Association

Telepharmacy patient satisfaction

sistently across telepharmacy sites.

Discussion Several implications can be drawn from this analysis. The primary objective was to identify the formation of patient satisfaction in rural community telepharmacies. That is, we sought to identify the underlying, latent factors that patients truly value, that induce them to repeatedly seek pharmacy services from these businesses. We found that a single underlying measure, defined simply as “telepharmaceutical care,” determined patient satisfaction. All of the survey questions appeared to contribute significantly to the formation of this latent factor. Second, this analysis suggested that the underlying constructs of patient satisfaction were slightly different in telepharmacies compared with other settings, such as traditional community pharmacies. Evidence presented by Larson et al.,1 for example, identified two related components of patient satisfaction in a traditional community setting, namely “friendly explanation” and “managing therapy.” This study’s findings implied that determining patient satisfaction is much simpler in the context of community telepharmacy, as only one latent factor was observed. Moreover, the three survey questions aligning most closely with overall patient satisfaction metric in this study load onto different latent factors in the Larson et al. study. This implies that the patient satisfaction measure derived in the current work was truly a composite of the two found previously. Lastly, the results in Table 5 imply that the underlying patient satisfaction construct does not differ based on the type of telepharmacy (central site or remote site), based on the location of the telepharmacy (which also included community-specific characteristics), across respondent ages, or by gender. Thus, although patients may form opinions of satisfaction differently between traditional pharmacies and telepharmacies, they did not appear to view the formation of satisfaction differently between the two types of telepharmacies used in this study. Although these findings are intriguing from a statistical perspective, understanding what they mean from the perspective of the patients frequenting rural community telepharmacies, the pharmacists owning and operating these establishments, and the economic vitality of the communities in which both the patients and pharmacists reside is even more important. Clearly, patient satisfaction with rural community telepharmacy services was universally high, implying that these telepharmacies were providing services that their patients valued. This finding was highly consistent with recent studies measuring patient satisfaction with telemedicine. For example, Gustke et al.23 surveyed patient satisfaction among 495 clinical consultations and found that average patient satisfaction exceeded 98%. Assuming that patients’ behavior is consistent with their values and that other challenges (including declining reimbursement and a shortage of available pharmacists) can be met successfully, these pharmacies have the potential to remain profitable enterprises. The use of telepharmacy technology also provided an opportunity for rural pharmacies to expand their scope of operations to other smaller surrounding communities and further enhance profitability.11,12 Similar implications exist from the perspective of the comJournal of the American Pharmacists Association

Research

munity. Many rural communities struggle to preserve their economic vitality and access to health care. Again, assuming that patients act consistently with their valuations of telepharmacy services, telepharmacy is one way to meet both of these ends simultaneously. One study estimated that the total economic effect of telepharmacies in North Dakota (where seven of our eight telepharmacies operate) was approximately $7.5 million. In many cases, these telepharmacies have the potential to accumulate $500,000 in economic activity in towns of 1,000 people or less.12 Therefore, as long as patients continue to be satisfied with and use these telepharmacies, the communities in which they exist will also continue to flourish. However, the reason for patients valuing telepharmacy services far higher and more consistently than traditional community services remains unresolved, particularly because patients tend to have high levels of satisfaction with their local community pharmacies.1 One likely explanation relates to the changing demographics of the communities themselves. As these communities age, patients are more likely to develop chronic conditions, which increases the demand for pharmacy services11 and reduces an individual’s ability and/or willingness to travel to receive medical care. This combination of forces also induces residents to value having access to a local pharmacy within the community, especially for those who prefer not to receive medication therapy management (MTM) over the telephone or via e-mail, as would be the case with a mail order pharmacy.23 Younger adults (especially those who have young children) may also value the convenience and personal service offered by lower-volume, rural telepharmacies. To a lesser extent, these perceptions may be reinforced by stories in the popular press chronicling rural pharmacy closures and other economic contractions in local economics (some of which may be in their own communities).12,24 In summary, people in these rural communities simply value access to all of the services provided in a local telepharmacy, inclusive of both dispensing and MTM activities.25 The fact that no significant differences existed in overall patient satisfaction based on the community, the type of telepharmacy model used, age, or gender suggested that the majority of individuals in each of these communities placed similar values on the access provided by community pharmacies. This may be driven by the fact that all of these communities were relatively small and thus faced similar economic challenges.25 Despite being located in different counties, these communities still shared many sociocultural characteristics, which would also help explain why access to care was valued across respondents consistently.

Limitations This study was subject to several limitations. The first limitation was the low response rate (24%) of the sample. Although not unacceptably low, this response rate was lower than similar studies1 and limited the sample size at our disposal. Ideally, a response rate closer to 50% would have doubled the size of the sample and increased the ratio of the sample size to the number of variables to a more acceptable level. As one reviewer noted, even if the sample was representative of the population, drawing reliable conclusions about the underlying population based on a www.japha.org

J u l /A u g 2009 • 49:4 •

JAPhA • 517

Research

Telepharmacy patient satisfaction

sample of less than 100 observations was difficult. In any case, our results should be considered as exploratory and highlight a need for future work in this area. A second limitation is the nature of community telepharmacy services examined in this study. This analysis was limited to rural, remote sites with technicians on staff. Other telepharmacy models and/or those in different settings (e.g., hospital inpatient telepharmacies) may exhibit differences in patient satisfaction compared with those observed in the current study. Urban community telepharmacies (e.g., those using automated dispensing machines to reduce the strain on the pharmacy staff during peak hours) may also exhibit differences in patient satisfaction compared with the current work. A third limitation arose because of the simplicity of our results. Because we identified only one global measure of patient satisfaction, the manner by which patient satisfaction should be assessed, especially for other rural community telepharmacies not covered in this study, could not be elucidated. Future work that establishes a comparison group, either of the same telepharmacies at a different point in time and/or a different set of rural community telepharmacies using a similar delivery model would allow for a more detailed and practical understanding of how to assess and improve patient satisfaction.

Conclusion This study investigated whether patient satisfaction with rural community telepharmacies was formed in the same manner as that for pharmacy settings that do not use community telepharmacy services. Using data from eight rural North Dakota and Minnesota sites, this study found that, like other practice settings, patient satisfaction in community telepharmacies can be characterized by a small number of underlying factors. However, patient satisfaction was characterized much more simply in community telepharmacies because, unlike “traditional” community pharmacy settings, only a single underlying construct was found for the rural telepharmacies sampled. We attributed this difference to the fact that patients in rural communities valued accessing pharmacy services locally at telepharmacies rather than having to travel outside of the community. However, our assertion, though corroborated by empirical evidence, was not proven conclusively. As such, future work is needed in this area. Understanding why patients who use rural community telepharmacies appear to be more satisfied than patients who visit “traditional” community pharmacies is important. More information in this area will allow pharmacists to better adjust their current practice models and thus maximize the quality of care provided to patients. References 1. Larson LN, Rovers JP, MacKeigan LD. Patient satisfaction with pharmaceutical care: update of a validated instrument. J Am Pharm Assoc. 2002;42:44–50.

4. Johnson JA, Coons SJ, Hays RD, et al. A comparison of satisfaction with mail versus traditional pharmacy services. J Manag Care Pharm. 1997;3:327–36. 5. Larson LN, MacKeigan LD. Further validation of an instrument to measure patient satisfaction with pharmacy services. J Pharm Mark Manage. 1994;8:125–39. 6. Donabedian A. The quality of health care in a health maintenance organization: a personal view. Inquiry. 1983;20:218–22. 7. Johnson JA, Coons SJ, Hays RD. The structure of satisfaction with pharmacy services. Med Care. 1998;36:244–50. 8. Nelson E, Rust R, Zahorik A, et al. Do patient perceptions of quality relate to hospital financial performance? J Health Care Mark. 1992;12:6– 13. 9. Powers T, Bendall-Lyon D. The satisfaction score. Mark Health Serv. 2003;23:28–32. 10. Desselle S, Zgarrick D. Pharmacy management: essentials for all practice settings. New York: McGraw Hill; 2005:229–64. 11. Traynor AP, Sorensen TD, Larson T. The main street pharmacy: becoming an endangered species. Rural Minnesota Journal. 2007;2:83– 100. 12. Dart L. Digital doses: telepharmacies save people in small towns and rural areas from having to drive hundreds of miles to fill a prescription. Rural Electric. 2005;64(1):28–30. 13. Liu J, Bellamy G, Barnet B, Weng S. Bypass of local primary care in rural counties: effect of patient and community characteristics. Ann Fam Med. 2008;6:124–30. 14. Ver Steegh A. Pharmacy closures threaten rural communities. Accessed at www.mn2020.org/index.asp?Type=B_ BASIC&SEC=%7B6405D470-68EB-46F6-9721-B02262AE42DB%7D, November 6, 2008. 15. North Dakota State University College of Pharmacy, Nursing and Allied Sciences. What is telepharmacy? Accessed at http://telepharmacy.ndsu.nodak.edu, August 22, 2008. 16. Young D. Telepharmacy project aids North Dakota’s rural communities. Am J Health Syst Pharm. 2006;63:1776, 1779–80. 17. Peterson CD, Anderson HC Jr. The North Dakota telepharmacy project: restoring and retaining pharmacy services in rural communities. J Pharm Technol. 2004;20:28–39. 18. Khan S, Snyder HW, Rathke AM, et al. Is there a successful business case for telepharmacy? Telemed J E Health. 2008;14:235–44. 19. Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate data analysis. 5th ed. New York: Macmillan; 1998:87–140. 20. Johnson RA, Wichern DW. Applied multivariate statistical techniques. 5th ed. Upper Saddle River, NJ: Prentice Hall; 2002:477–542. 21. Mair F. Systematic review of studies of patient satisfaction with telemedicine. BMJ. 2000;320:1517–20. 22. Ried LD, Wang F, Young H, Awiphan R. Patients’ satisfaction and the perception of the pharmacist. J Am Pharm Assoc. 1999;39:835–42. 23. Gustke SS, Balch DC, West VL, Rogers LO. Patient satisfaction with telemedicine. Telemed J. 2000;6:5–13.

2. Schommer JC, Kucukarslan SN. Measuring patient satisfaction with pharmaceutical services. Am J Health Syst Pharm. 1997;54:2721–32.

24. Vaughan C. North Dakota telepharmacy project fills need. Health Leaders Media. Accessed at: http://telepharmacy.ndsu.nodak.edu/ publications/HealthLeadersMedia.2doc.htm, November 6, 2008.

3. MacKeigan LD, Larson LN. Development and validation of an instrument to measure patient satisfaction with pharmacy services. Med Care. 1989;27:522–34.

25. Whitten P, Love B. Patient and provider satisfaction with the use of telemedicine: overview and rationale for cautious enthusiasm. J Post Grad Med. 2005;51:294–300.

518 • JAPhA • 4 9 : 4 • J u l / A u g 2009

www.japha.org

Journal of the American Pharmacists Association