Patient satisfaction evaluations in different clinic care models: Care stratification under a national demonstration project

ARTICLE IN PRESS Health & Place 16 (2010) 85–92

Contents lists available at ScienceDirect

Health & Place journal homepage: www.elsevier.com/locate/healthplace

Patient satisfaction evaluations in different clinic care models: Care stratification under a national demonstration project Blossom Yen-Ju Lin a,, Cheng-Chieh Lin a,b,c,d, Yung Kai Lin e a

School and Graduate Institute of Health Services Administration, College of Public Health, China Medical University, Taichung, Taiwan Department of Family Medicine, China Medical University Hospital, Taichung, Taiwan c Department of Family Medicine, College of Medicine, China Medical University, Taichung, Taiwan d Institute of Health Care Administration, College of Health Science, Asia University, Taichung, Taiwan e Division of Cardiovascular Surgery, Taichung Veterans General Hospital, Taichung, Taiwan b

a r t i c l e in fo

abstract

Article history: Received 14 April 2009 Received in revised form 17 July 2009 Accepted 29 August 2009

Primary Community Care Networks (PCCNs) were the product of primary care health reform in Taiwan. Under the PCCN intervened nationwide as a demonstration project, there were three types of service contexts for clinic patients: (1) member patients in PCCN clinics; (2) non-member patients in PCCN clinics; and (3) patients in non-PCCN clinics. A multi-site, cross-sectional validated survey of 3143 outpatients receiving care in clinics was conducted to investigate quality of care delivered to these three distinct clinic patients. It revealed that member patients indicated a higher level of satisfaction with the care quality of several physician–patient relationships and an increased willingness to recommend their clinics over non-member patients in PCCN clinics. However, no differences were found in the care quality evaluation measures between PCCN member patients and non-PCCN patients. Health policy implications were discussed for primary health reforms in clinic service contexts in this study. & 2009 Elsevier Ltd. All rights reserved.

Keywords: Clinic patient satisfaction Physician–patient relationship Health reform Integrated health networks Primary care

1. Introduction The SARS epidemic challenged Taiwan public health and the healthcare system in Spring 2003. Citizens’ freedom to voluntarily select their own medical providers complicated matters for the National Health Authority, an organization that sought to control and trace the direction and progression of the epidemic. This event precipitated the Taiwanese National Health Authority to reconsider what transpired with respect to the traditional, fragmented healthcare providers in Taiwan. ‘‘Primary Community Care Network (PCCN) Demonstration Project’’ was one of the resulting health reforms, a nationwide healthcare financing program funded by the Bureau of National Health Insurance (BNHI) in 2003. The PCCN served as a new model for the Taiwanese government to redefine the role of clinic physicians in the healthcare delivery system (Bureau of National Health Insurance in Taiwan, 2009). A PCCN consists of a group of clinic physicians whose medical jobs are viewed as family care, and these clinics have to cooperate with at least one hospital for patients’ secondary or tertiary care. A PCCN consists of 5–10 clinics: half of them should offer the services of general medicine, internal medicine, surgery, obstetrics

 Corresponding author. Tel.: + 886 4 22053366; fax: + 886 4 22076923.

E-mail address: [email protected] (B.Y.-J Lin). 1353-8292/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.healthplace.2009.08.008

and gynecology, pediatric, or family medicine. And the specialty clinics, usually handling the outpatients with mild illnesses, less complicated symptomologies than hospital specialties, are also allowed to join the PCCN demonstration project, including those who practice Otolaryngology, Ophthalmology, Rehabilitation Medicine, Dermatology, and Psychiatry (Bureau of National Health Insurance in Taiwan). One of the major tasks for clinic physicians in a PCCN is to recruit their patients to become PCCN patient members. The PCCN patient members have the extra benefits from their clinic physicians, including filed personal and family medical/health information for further health maintenance assists and suggestions; accessed 24-h a day, 7-day a week medical consultation telephone lines when their family physicians are off; approaching free medical brochures, health or medical lectures; reminded timing of health examinations; provided health education for the chronic disease management, and so on. Under the implementation of the PCCN demonstration project in Taiwan, outpatients seeking clinic services can be categorized to three types of service contexts: member patients in PCCN clinics (Patient Type I); nonmember patients in PCCN clinics (Patient Type II); and patients in non-PCCN clinics (Patient Type III) (see Fig. 1). It has been 5 years since the Taiwan Health Authority launched the health reform of the PCCN demonstration project and the project is still going on. However, few studies have attempted to examine the possible effects of clinic services delivered in the

ARTICLE IN PRESS 86

B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

Taiwan clinics Participating PCCN demonstration project yes

no

PCCN clinics

Non-PCCN clinics

PCCN clinic physicians recruited their patients into PCCN demonstration project yes

Patients in non-PCCN clinics Member patients in PCCN clinics (Patient Type I )

(Patient Type III ) no

Non-member patients in PCCN clinics

(Patient Type II ) Fig. 1. Categories of clinic outpatients under the implementation of the PCCN demonstration project in Taiwan.

three different service contexts. Our study was aimed to investigate whether there exist different patterns of quality of care delivered to three distinct clinic patients under the implementation of PCCN demonstration project. The findings would provide health policy implication about whether the PCCNs were organizational innovations for patients, worthy of increased diffusion, and meriting further exploration.

This study was aimed to evaluate and compare how patients in the three different care service contexts ranked with regard to service satisfaction, applying an ecological design.

pre-tested for 25 patients who experienced clinic visits. The wordings and meanings of each question item were revised to ensure content validity. Ten items were loaded into one common factor through a factor analysis, and the Cronbach a value was 0.889. Patient demographics and socioeconomic factors – all of which were verified to be associated with patient satisfaction (Bonds et al., 2004; Fan et al., 2005; Jung et al., 2003) – were included in the questionnaire. These factors comprised: gender, age, education, and whether the patients visited the surveyed clinics or not when they presented similar symptomatologies (i.e., frequent patients). In addition, the visiting clinic characteristics were also collected, including clinics’ location (non-urban vs. urban), specialty, and area competition (i.e., counted as number of the clinics in the county/city level).

2.1. Survey instrument development: patient satisfaction to primary care clinics’ service quality

2.2. Study subjects

In order to compare the three different service contexts of clinic care – member patients in PCCN clinics, non-member patients in PCCN clinics, and patients in non-PCCN clinics, the question items on the survey could not be limited to the characteristics and values specified in the PCCN clinics in this study. Rather, the general dimensions commonly described in service quality of primary care were included in the questionnaire. The structured questionnaires were first drafted from a thorough review of the previous literature (Institute of Medicine, 1996; Scheffler et al., 1978; Starfield, 1992; Baker et al., 2003, 2002; Borowsky et al., 2002; Polluste et al., 2000; Razzouk et al., 2004; Sampson et al., 2004; Wensing et al., 1998; Zebiene et al., 2004) and then examined by two academic professors and two clinic physicians to assure their logic, accuracy, and feasibility. To be practical, we chose ten items to cover as many dimensions of clinic care quality as possible. In addition, we expect the findings to serve as a framework for policy makers and healthcare providers to examine each of them for their potential contributions toward quality improvement. As a result, ten items – satisfaction for wait time, employee courtesy, physician competency, the humaneness of understanding and explanation to patients, and the concept of modern preventative and chronic illness management – were included, using a 5-point Likert scale. Overall patient satisfaction (also using a 5-point Likert scale) and patient willingness to recommend their surveyed clinics (using a ‘‘yes’’ or ‘‘no’’ response) were also measured. One pilot study was

We focus on all 416 participating clinics (i.e., PCCN clinics) located in the administrative areas of the middle branch of BNHI in this study. In order to effectively compare three different service contexts, it is necessary to identify the non-PCCN clinics relative to PCCN clinics. Two criteria were used to identify the respective non-PCCN clinics: (1) the respective non-PCCN clinics provided the same medical specialties as the PCCN clinics and (2) the respective non-PCCN clinics were located in the same market districts as the PCCN clinics. The studied PCCN clinics were excluded in our sample when their respective non-PCCN clinics with the same service lines and in the same district areas could not be identified. In addition, the studied PCCN clinics – those that were closed during the surveyed time period and those that declined to participate in our patient surveys – were also excluded. Finally, 324 PCCN clinics, out of total of 416 clinics, were included in this study. There were no significant differences noted with regard to the excluded and included PCCN clinics, according to geographical distribution (w2 = 0.214, p40.05). Among the selected PCCN clinics, 89 specialized in General Medicine, 26 specialized in Internal Medicine, 12 specialized in Surgery, 33 specialized in Obstetrics and Gynecology, 54 specialized in Pediatrics, 40 specialized in Family Medicine, 40 specialized in Otolaryngology, 15 specialized in Ophthalmology, 4 specialized in Rehabilitation Medicine, 9 specialized in Dermatology, and 2 specialized in Psychiatry; the same medical specialty distribution in the other 324 respective non-PCCN clinics.

2. Method

ARTICLE IN PRESS B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

2.3. Sampling strategy and questionnaire administration From each of the clinics included in this study, five patients were selected (see the appendix for sample estimation) by welltrained interviewers at the front entrance of the clinics after the patients had finished and were exiting. We let the interviewers stand outside the studied clinics to wait for the outpatients finishing their physician visits to avoid patients’ answers with the social desirability for clinic physicians. Consecutive sampling was used by all the interviewers until up to five patients were reached for each studied clinic. The questionnaires were handed to the respondents by the interviewers. Informed consent was obtained from each participant when the interviewers handed out the questionnaires in the study. Overall, the rejection rate in the surveying process was approximately 10%. The survey was conducted between April 21 and May 22, 2006. The final number of valid responses (valid survey feedback) totaled 1564 from the PCCN clinics and 1579 from the non-PCCN clinics. 2.4. Analytical techniques All the surveyed patients were classified into three groups, representing the three different clinic service contexts: member patients in the PCCN clinics (Patient Type I), non-member patients in the PCCN clinics (Patient Type II), and patients in the non-PCCN clinics (Patient Type III). The data were initially analyzed applying a descriptive approach using means and frequency counts (percentages) for patient background information and each survey item. Differences in personal background information and each survey item were analyzed across three patient types of clinic service contexts with w2 tests or ANOVA. With individual responding patients as unit of analysis, we performed the individual multiple regression analyses for each satisfaction question item and a logistic regression for patient promotion willingness for their visiting clinics. Multiple regression models and a logistic regression model were performed to explore how three different patient types in clinic service contexts ranked their satisfaction with regard to service quality and promotion willingness, respectively, controlling for patient background information (i.e., personal and visiting clinic characteristics). It deserves to mention that the data in this study could be analyzed by two-level or multilevel analysis if there are variations in clinics in terms of their demographic, organizational, and contextual differences. However, we have checked the Intraclass Correlation Coefficient (ICC) index for the individual dependent variables: individual patient satisfaction scores (all ICCso0.001) and patient willingness to recommend their visiting clinics (ICC= 0.014) – found that all the ICC values were less than 0.05, which is the minimum cut-point for performing multilevel analysis. And it suggested that the multilevel effects could be ignored in this dataset (Bliese, 2000; Cohen, 1988). All statistical analyses were performed using SPSS 12.0 software.

3. Results The number of sampled patients from the PCCN and non-PCCN clinics was approximately equal (i.e., 1564 vs. 1579). Of the 1564 patients from the PCCN clinics, 479 were classified as member patients, 1080 were classified as non-member patients, and 5 patients failed to identify their status as either member patients or non-member patients in the PCCN clinics. Females represented 56% of the respondents. The average age of those surveyed was 32.35 years, and 36% of them had educational levels of undergraduate and graduate. In terms of surveyed patients’ visiting

87

clinic characteristics, around 88% surveyed patient visits were located in urban clinics, around half of surveyed visits as internal medicines (see Table 1). The distributions of gender, age, frequent visiting type, and visiting clinic location differed among the three categories of surveyed patients. For example, female patients occupied a lower percentage of member patients of the PCCN clinics (Patient Type I), non-member patients in the PCCN clinics (Patient Type II) were older, and member patients of the PCCN clinics (Patient Type I) had higher percentages of frequent patients in the surveyed clinics. And higher percentage of member patients (around 94%) in the PCCN clinics (Patient Type I) was served on the urban areas (see Table 1). 3.1. Descriptive analysis of patient evaluation in clinic services As shown in Table 2, patients were relatively satisfied with all the surveyed items. The top three items which patients found to be most satisfactory were: trust in their physicians (item 8: mean 4.043), opportunities to describe diseases or symptoms (item 4: mean 4.028), and physicians’ abilities (item 10: mean 4.027). Similar patterns were found when analyzing three individual patient types. 3.2. Comparison of patient evaluations across three clinic service contexts Univariate analysis revealed that almost all patient evaluation items and patient willingness to recommend the clinics were ranked the highest by member patients in the PCCN clinics (Patient Type I), followed by patients in the non-PCCN clinics (Patient Type III), and then non-member patients in the PCCN clinics (Patient Type II) (p o0.05) (Table 2). With individual patient satisfaction evaluation items of service quality as dependent variables controlling patient backgrounds and visiting clinic characteristics, the regression analyses found that member patients in the PCCN clinics showed higher satisfaction than nonmember patients in the PCCN clinics (p o0.05) (Patient Type I vs. Patient Type II, as shown in Table 3) with the seven service items: courtesy of employees, patients’ needs cared by physicians, opportunities for patients to describe diseases or symptoms, physicians’ suggestions for patient health maintenance, patients’ trust in physicians, patients’ beliefs in physicians offering the best treatment(s), and physicians’ abilities. Overall patient satisfaction and patient willingness to recommend the visited clinics were also higher in the member patients than in the non-member patients in the PCCN clinics (Patient Type I vs. Patient Type II, as shown in Tables 3 and 4). Overall, there is a minor difference in service satisfaction between member patients in the PCCN clinics and patients in the non-PCCN clinics (Patient Type I vs. Patient Type III, also shown in Tables 3 and 4). However, member patients in the PCCN clinics were less satisfied with the waiting time than patients in the non-PCCN clinics (po0.01). In addition, patients who frequently visited the surveyed clinics showed higher satisfaction levels across all care dimensions tested in this study (see Table 3) and a stronger willingness to recommend their clinics (see Table 4). Older patients showed higher satisfaction levels with waiting times, physicians’ suggestion for health maintenance, and trusts. And female patients were more willing to recommend their clinics (see Tables 3 and 4). Pediatric and obstetric patients were more satisfied in opportunities to describe their diseases or symptoms, physicians’ suggestions on health maintenance and emergency cases, and trusts in physicians (items 4, 6–8, Table 3) than internal medicine patients. Furthermore, patients in non-urban areas showed higher satisfaction levels across all care dimensions tested in this study, save the

ARTICLE IN PRESS 88

B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

Table 1 Backgrounds of the responding patients (n= 3138). Variables

Frequency (%) Mean (SD)

Patient type I

Patient type II

Patient type III

Member patients in PCCN clinics Non-member patients in PCCN (n =479) clinics (n= 1080) Frequency (%) Gender Male Female

1352 (43.0%) 1756 (55.9%)

Mean (SD)

178 (37.2%) 297 (62.0%)

Frequency (%)

Mean (SD)

484 (44.8%) 590 (54.6%)

Patients in non-PCCN clinics (n =1579) Frequency (%)

Mean (SD)

687 (43.5%) 868 (55.0%)

v2 test: 8.382* Age (years) r 20 21–30 31–40 41–50 51–64 Z 65

822 707 622 479 321 173

(26.2%) (22.5%) (19.8%) (15.2%) (10.2%) (5.5%) 32.35 (18.15)

29.97 (17.20)

34.30 (18.33)

31.72 (18.18)

13.56 (5.51)

13.84 (5.78)

ANOVA: patient type II4I***; II 4III** Education (years) Elementary school and below Junior high school Senior high school Undergraduate Graduates

105 965 899 400 734

(3.3%) (30.7%) (28.6%) (12.7%) (23.4%) 13.72 (5.75)

13.73 (6.19)

ANOVA: no statistical significance for patient type I, II, and III Frequent patients Yes No

1470 (46.8%) 1650 (52.5%)

314 (65.6%) 164 (34.2%)

437 (40.5%) 639 (59.2%)

719 (45.5%) 847 (53.6%)

931 (86.2%) 149 (13.8%)

1397 (88.5%) 182 (11.5%)

v2 test: 85.355*** Visiting clinic locations + Urban Non-urban

2785 (88.6%) 358 (11.4%)

452 (94.4%) 27 (5.6%)

v2 test: 21.902*** Clinic specialty Internal MedicineJ Surgical medicine Pediatrics/obstetrics Other specialties&

1489(47.4%) 115 (3.7%) 854 (27.2%) 685 (21.8%)

216 16 150 97

(45.1%) (3.3%) (31.3%) (20.3%)

521 44 274 241

(48.2%) (4.1%) (25.4%) (22.3%)

748 55 430 346

(47.4%) (3.5%) (27.2%) (21.9%)

v2 test: 6.447 (p40.05) Note: (1) + Taichung and Changhwa cities and counties were recoded as urban; Nantou city and county were recoded as non-urban. Non-urban areas were defined as the areas which carry the population 300,000 persons or below in county/city level. The definition was based on the Ministry of The Interior in Taiwan (http://www.moi.gov.tw/ english/index.aspx). (2)JInternal Medicine included General Medicine, Internal Medicine, and Family Medicine. (3)&Other specialties included Otolaryngology, Ophthalmology, Rehabilitation Medicine, Dermatology, or Psychiatry. (4)*p o0.05; ** p o 0.01; *** p o0.001.

perception of waiting time. And surveyed patients in visiting clinics with higher competition levels in the county/city, were ranked higher service satisfactions across all care dimensions, save the perception of the two items: opportunities for patients to describe diseases or symptoms (item 5, Table 3), and patients’ beliefs in physicians offering the best treatment(s) (item 9, Table 3). However, no visiting clinic characteristics were found to have statistically significant relationships for patients’ willingness to promote their visiting clinics (see Table 4).

3.3. Determinants of overall patient service satisfaction across three clinic service contexts To understand the determinants of overall patient satisfaction for three primary care contexts, individually, three individual regression models were performed, controlling for patient personal and visiting clinic characteristics (see Table 5). It was found that all three types of patients in clinic service contexts were all cared about employee courtesy (item 2), patients’ beliefs to physicians for offering the best treatment (item 9), and

physicians’ abilities (item 10), related to overall patient satisfaction. In addition, it revealed that different service dimensions play a role in determining overall patient satisfaction for three patient types in clinic service contexts. For example, PCCN member patients (Patient Type I) cared more on opportunities to describe their symptoms or diseases (item 4) and their symptoms or diseases explained in details by physicians (item 5); non-member patients in PCCN clinics (Patient Type II) emphasized more about their needs cared by physicians (item 3); and non-PCCN clinic patients (Patient Type III) favored their needs cared by physicians (item 3), and physicians’ suggestions for them to maintain their health in their home routines (item 6).

4. Discussion In this study, three types of patients seeking treatment in different care contexts of clinics were queried using a patient survey to understand the quality of services provided. Controlling for patient background information (i.e., personal and visiting clinic characteristics), the satisfaction evaluation showed no

ARTICLE IN PRESS B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

89

Table 2 Descriptive analysis of quality satisfaction of primary care services. Question items

a

Overall

Patient type I

Patient type II

Patient type III

Mean (SD)

Member patients in PCCN clinics (n =479) Mean (SD)

Non-member patients in PCCN clinics (n =1080) Mean (SD)

Patients in nonPCCN clinics (n =1579) Mean (SD)

3.146 (0.815)

3.143 (0.830)

3.227 (0.833)

III 4II

3.950 (0.677)

3.771 (0.696)

3.845 (0.685)

I 4III 4II

4.029 (0.649)

3.838 (0.693)

3.929 (0.674)

I 4III 4II

4.126 (0.639)

3.971 (0.670)

4.036 (0.649)

I 4III 4 II

4.027 (0.686)

3.937 (0.684)

4.019 (0.675)

III 4II

4.004 (0.712)

3.856 (0.732)

3.904 (0.724)

I 4III 4II

3.989 (0.717)

3.841 (0.721)

3.890 (0.727)

I 4 II,. III

4.173 (0.657)

3.968 (0.683)

4.056 (0.679)

I 4III 4II

4.100 (0.662)

3.917 (0.681)

3.988 (0.666)

I 4III 4II

4.138 (0.632)

3.968 (0.647)

4.034 (0.656)

I 4III 4II

4.115 (0.618)

3.903 (0.671)

4.004 (0.609)

I 4III 4II

Patient Type I Member Patients in PCCN clinics (n =479) Frequency (%)

Patient Type II Non-member patients in PCCN clinics (n =1080) Frequency (%)

1. Satisfaction with waiting time

ANOVA testing for three patient types on items

3.185 (0.830) 2. Satisfaction with employee courtesy 3.836 (0.690) 3. Satisfaction with patient needs cared by physicians 4. Satisfaction with patient opportunities to describe their diseases or symptoms 5. Satisfaction with patient symptoms or diseases explained in details by physicians 6. Satisfaction with physicians’ suggestions for patients to maintain their health in their home routines 7. Satisfaction with physicians’ suggestions for patients who have prn conditions 8. Patient willingness to tell physicians for the health status with trusts 9. Patient belief to physicians for offering the best treatments 10. Physicians’ abilities to taking care of patients

3.913 (0.680) 4.028 (0.656) 3.992 (0.680) 3.902 (0.726)

3.888 (0.724) 4.043 (0.680) 3.980 (0.673) 4.027 (0.651)

11. Overall patient satisfaction 3.986 (0.636) Overall

Frequency (%) 12. Patient willingness to recommend their clinics

Yes 2530 (80.5%) No 590 (18.7%)

428 (89%) 50 (10%)

826 (76%) 251 (23%)

Patient Type III Patients in nonPCCN clinics (n =1579) Frequency (%) 1272 (81%) 589 (37%)

v2 testing for three patient types

36.003***

a The post-hoc tests of ANOVA were decided based on test of homogeneity of variances. Scheffe tests were performed when the survey item variables meet the assumptions of homogeneity of variances and Tamhane tests for violation of homogeneity of variances.

difference between the PCCN member patients (Patient Type I) and the non-PCCN patients (Patient Type III). However, there were lower rankings by the non-member patients in the PCCNs (Patient Type II). In addition, the results revealed that different service dimensions play a role in determining overall patient satisfaction for three patient types in clinic service contexts. It was an interesting finding that PCCN member patients (i.e., Patient Type I) showed no significant difference on service satisfactions from the patients in non-PCCN clinics (i.e., Patient Type III). This result could be explained by two possible arguments. First, at the onset of the PCCN demonstration projects, the recruiting criteria for the clinics were not based on the ‘‘quality’’ evaluation of participating clinics, but rather on ‘‘volunteerism’’. With practical phenomena in the background of healthcare reform, it was found that some clinics with good images or/and large outpatient volumes did not participate in or join into the demonstration project due to the cumbersome administrative paperwork and mandatory meetings. Although we may appreciate the efforts imposed by the PCCN clinic physicians, we cannot ignore those non-PCCN clinics that secure their service quality in routine business. Second, although the PCCN healthcare reform has major foci and requests on its PCCN clinics, we also expressed that the PCCN demonstration project ‘‘sells’’ the concept and value of ‘‘family physicians’’ and ‘‘patient-centered health services’’ nationwide to healthcare organizations in Taiwan. Therefore, the concept could be diffused by both the public and healthcare providers; and, at the same time, within the Taiwanese

healthcare industry. We could argue that the atmospherics of primary care reform have pervaded the healthcare industry irrespective of whether or not the clinic providers joined this national demonstration project. Too, this contention might lead to equal effects of service quality of non-PCCN patients and the PCCN member patients on satisfaction rankings and promotion willingness in the ‘‘post-test only’’ design study. Longitudinal observations could be continued to determine if actual effectiveness exists between the PCCN clinics and non-PCCN clinics, even after the demonstration project is terminated. Furthermore, continued observations could also provide insight and clues regarding whether PCCN clinic physicians maintain high quality in their organizational and interpersonal behaviors vis-a -vis their clinic patients. In the processing of the Taiwan PCCN demonstration project, the PCCN clinic physicians announced the spirit of the family physician and recruited their patients into the PCCNs using an informed consent method. The PCCN physicians found it was much easier to recruit their frequent patients into the member patients of the demonstration project. This phenomenon could be a result of patients trusting more in their physicians. We also found that a larger percentage of member patients were frequent patients (Table 1), and frequent patients showed higher satisfaction ratings across all items of service quality (Table 3). It has been found that physician–patient relational breakdowns and/or a dissolution of trust led to patient removals from doctors’ lists (Sampson et al., 2004). Of higher importance, trust was the

ARTICLE IN PRESS 90

B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

Table 3 Parameters of comparisons of service satisfaction for member and non-member patients adjusting patient and visiting clinic characteristics. Std. b

Reference: patient type IJ Member patients in PCCN clinics (n =479)

Adjusting Age

Gender (default: female)

Education Frequent patients

Patient type II Patient type III Non-member Patients patients in in nonPCCN clinics PCCN (n= 1080) clinics (n= 1579) 1. Satisfaction with waiting time 2. Satisfaction with employee courtesy 3. Satisfaction with patient needs cared by physicians 4. Satisfaction with patient opportunities to describe their diseases or symptoms 5. Satisfaction with patient symptoms or diseases explained in details by physicians 6. Satisfaction with physicians’ suggestions for patients to maintain their health in their home routines 7. Satisfaction with physicians’ suggestions for patients who have prn conditions 8. Patient willingness to tell physicians for the health status with trusts 9. Patient belief to physicians for offering the best treatments 10. Physicians’ abilities to take care of patients 11. Overall patient satisfaction

0.035

0.077**

0.053**  0.018

Clinic location + (default: nonurban)

Specialty (default: internal med) Surgical med

Ped/Obs med

Other med

 0.007

 0.046* 0.104***

0.028

0.132***

 0.038

 0.036*

Clinic # in locations (competition)

 0.074**

 0.038

0.017

 0.003

 0.009

0.249***

 0.229***

 0.026

0.015

 0.006

0.111***

 0.075**

 0.024

0.035

0.002

 0.005

0.284***

 0.190***

 0.012

0.023

 0.008

0.072*

 0.067**

 0.030

0.018

0.007

0.003

0.238***

 0.203***

 0.003

0.043*

0.025

0.069*

0.033

0.005

 0.003

0.005

0.243***

 0.157***

0.006

0.028

0.007

0.040

 0.013

0.200***

 0.208***

0.025

0.042*

 0.016

0.096***

 0.026

0.073*

 0.014

 0.060*

 0.038

0.049**  0.015

 0.060*

 0.038

0.027

0.004

0.013

0.203***

 0.179***

 0.014

0.045*

 0.091***

 0.048

0.018*

0.006

0.014

0.255***

 0.262***

 0.012

0.051**

 0.070**

 0.035

0.035

0.033

 0.009

0.259***

 0.172***

 0.014

0.043

 0.012

 0.056*

 0.026

0.009

0.014

 0.004

0.271***

 0.129***

 0.004

0.021

0.006

0.133***

 0.087***

 0.025

0.032

0.007

 0.024

0.322***

 0.132***

0.027

0.037

 0.013

0.118***

0.011

0.172***

0.160

Note: (1) JThree types of patients were recoded as two dummy variables and Patient type I, member patients in PCCN clinics (n= 479), as default one. (2) + Taichung and Changhwa cities and counties were recoded as urban; Nantou city and county were recoded as non-urban. Non-urban areas were defined as the areas which carry the population 300,000 persons or below in county/city level. The definition was based on the Ministry of The Interior in Taiwan (http://www.moi.gov.tw/english/index.aspx). (3) *p o 0.05; **po 0.01; ***p o 0.001.

Table 4 Parameters of comparisons of patient willingness to recommend their clinics for member and non-member patients adjusting patient and visiting clinic characteristics. Unstd. b

S.E.

Constant Patient type I: member patients in PCCN clinics Patient type II: non-member patients in PCCN clinics Patient type III: patients in non-PCCN clinics

0.505 (Reference)  0.440  0.219

0.290

3.038

1.657

0.180 0.176

5.951 1.548

0.644 0.803

Personal characteristics Age Gender (reference: female) Education Frequent patients (reference: no)

 0.001  0.349 0.011 2.434

0.003 0.101 0.009 0.148

0.048 12.024 1.480 272.060

0.999 0.706 1.011 11.400

0.020

0.254

0.006

1.020

 0.016  0.093  0.081 0.001

0.249 0.132 0.133 0.000

0.005 0.502 0.369 3.558

0.982 0.911 0.922 1.001

Visiting clinic characteristics Clinic location (reference: non-urban) Specialty (reference: internal medicine) Specialty: surgical Specialty: Ped/Obs Specialty: others Clinic # in locations (competition) Note: *po 0.05; **p o0.01; ***p o 0.001.

Wald

Odd’s ratio

*

*** ***

ARTICLE IN PRESS B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

91

Table 5 Determinants of overall service satisfactions across different patient types of clinic service contexts. (Determinants) std. b

Overall service satisfactions (dependent variable) Patient type I Member patients in PCCN clinics (n =479)

1. 2. 3. 4.

Satisfaction with waiting time 0.036 Satisfaction with employee courtesy 0.290*** Satisfaction with patient needs cared by physicians 0.019 Satisfaction with patient opportunities to describe their diseases or 0.124* symptoms 5. Satisfaction with patient symptoms or diseases explained in details 0.136** by physicians 6. Satisfaction with physicians’ suggestions for patients to maintain 0.092 their health in their home routines 7. Satisfaction with physicians’ suggestions for patients who have prn  0.076 conditions 8. Patient willingness to tell physicians for the health status with trusts  0.054 9. Patient belief to physicians for offering the best treatments 0.196*** 10. Physicians’ abilities to take care of patients 0.119* Personal characteristics Age Gender (reference: female) Education (years) Frequent patients (reference: no) Visiting clinic characteristics Clinic location (reference: non-urban) Specialty (reference: internal medicine) Specialty: surgical Specialty: Ped/Obs Specialty: others Clinic # in locations (competition)

Patient type II Non-member patients in PCCN clinics (n= 1080)  0.057 0.092** 0.167*** 0.055

Patient type III Patients in non-PCCN clinics (n= 1579) 0.008 0.115*** 0.078* 0.057

0.051

0.034

0.054

0.073

0.058

0.020

0.024 0.187*** 0.091*

0.052 0.111*** 0.145***

0.017 0.042 0.002 0.091*

 0.006  0.010  0.003 0.160***

0.017 0.006  0.025 0.133***

0.014

 0.023

 0.002 0.060  0.007 0.015

0.078*** 0.010  0.015 0.086*

0.007 0.008  0.005  0.009 0.033

*po 0.05, **po 0.01, ***po 0.001.

strongest predictor of patient satisfaction with family doctors (Baker et al., 2003); therefore, trust establishment could be the first step toward dedication and continuity in the spirit of family care. This should especially hold true for those non-member patients in the PCCN clinics. The National Health Authority should encourage physicians to enhance physician–patient interaction, such as enhancing communication quality and interpersonal treatment, via the formalization of sustained physician–patient partnership over time, using written documents, verbal exchanges of promises, other methods of formalizing accountabilities (Safran, 2003), and even telemedicine and e-health services (Matusitz and Breen, 2007); these might lead to the much needed, real health reform of primary care. Overall, we also found that patients in non-urban areas and higher competition areas ranked higher service satisfaction (see Table 3). We might argue that under the circumstances of higher competitive marketplaces might drive physician practices to distinguish their services from others through more efforts involved in the patient cares. Furthermore, we might appreciate the values of the non-urban providers devoted in the patient service quality and leaded to better perception about the healthcare providers. The patient perception of regarding communication and interactions with healthcare providers were examined and found that patients in urban areas reported poorer communication by their healthcare providers than non-urban patients (Wallace et al., 2008). Differences in perceived quality in this study might give the healthcare providers some hints for further examine for the services provided in urban and non-urban residence settings. In this study, the PCCN member patients (Patient Type I) were found to be more preferential toward the concept of patientcentered relationships than non-member patients in PCCN clinics (Patient Type II). For example, areas noticed include: patients’

needs cared by physicians (item 3 in Table 3), patients’ opportunities to describe their symptoms or diseases (item 4), patients’ willingness to tell the physicians for their health status with trusts (item 8), patients’ beliefs in physicians’ treatment choices (item 9), and physicians’ abilities (item 10). We agreed that it was probably PCCN clinic physicians who devoted themselves more to their PCCN member patients. Or, the nonmember patients in PCCN clinics held ‘‘perceptions’’ in different ways, as compared to the member patients in the clinics. We believe this condition may be related to PCCN physicians needing to devote more time and effort to consulting with their member patients – such as setting up 24/7 medical consultation telephone lines for providing urgent services onsite and for taking care of the patients whose family physicians’ practices are closed. We might also argue that the demonstration project accidentally disadvantaged non-members and set up a two-tier quality system within PCCNs’ clinics (see Tables 3 and 4). One of our findings might be a clue for PCCN clinic physicians to make up the disadvantage, as the results shown in Table 5. For example, different dimensions of perceived service satisfaction play a role in determining overall patient satisfaction across three patient types of clinic service contexts. In addition to employee courtesy (item 2), patients’ beliefs in physicians’ treatment choices (item 9), and physicians’ abilities (item 10) as the positive determinants for the overall patient satisfaction across all three service contexts, furthermore, non-member patients in PCCN clinics (Patient Type II) appreciated whether their needs were cared by physicians (item 3), allowing PCCN clinic physicians to make additional efforts toward establishing improved physician–patient relationships. Then, these patients could be recruited as member patients in the future. We might suggest, then, that the PCCN clinics’ physicians could propose additional recruitment planning: the establishment of better physician–patient relationships through

ARTICLE IN PRESS 92

B.Y.-J. Lin et al. / Health & Place 16 (2010) 85–92

community involvement; gatherings held to non-member patients with the experiences sharing and promotions of member patients; and enhancing financial incentives through deductible registration fees (Bureau of National Health Insurance in Taiwan; Lin and Wan, 2003). After all, the ultimate goal of this nationally launched demonstration project was to establish the family physician system and enable the citizens to have their own family physicians so as to improve the traditional fragmented healthcare industry and to care the people from the whole-person perspective. Several limitations need to be addressed at this stage. First, when we decided to include the clinics (PCCN clinics and nonPCCN clinics) in this study, we allowed our interviewers to use the ‘‘replacement method’’ to switch subjects (outpatients) who did not wish to participate in this study. Also, because of our use of the consecutive sampling method, the representative sample of the clinic population would be limited. Caveats should also be suggested to patients about the distinction between visible and invisible team care with primary care providers (Safran, 2003). For example, clinic physicians and their respective roles were much more visible and understandable by the patients; however, the roles of other practitioners involved in the patient care processes – clinic nurses, pharmacists, and administrators – were not clear to the patients. Thus, the performance of the primary care model tested in this study could only be focused on the surveyed items: the general items of quality of care (and much more from the physicians’ perspectives). Other possible gaps in the specific performance evaluation of the primary care team relationships in the PCCNs could be further studied according to different perspectives. For example, teamwork and clinic–hospital partnerships comprise a few options (Lin, 2007). In addition, future research on this topic could extend into patient evaluation of different clinic service contexts from the central Taiwan area to other geographical areas, a means to understand what the convergence and discriminatory potentialities are across different administrative areas in the PCCN demonstration program.

Acknowledgement We extend our appreciation to the interviewers at the Institute of Health Services Administration at China Medical University for joining this project. We also appreciate grant support from the Taiwan National Science Council (NSC93-2416-H-039-001-SSS; NSC94-2416-H-039-001-SSS).

Appendix. Estimates of sample sizes for each included clinic Because we designed the study based on an ecological perspective, we estimated the sample size to be as follows: (1) Based on the rule of the ratio of sample size and the number of study variables (question items in the survey), we assumed that one variable had at least 20 samples (outpatients), that is, 20 times as many respondents as observed variables (Baggaley, 1982; Marasculio and Levin, 1983): 16 variablesðin this survey; including the satisfaction survey questions and personal characteristicsÞ

 20 ¼ 320ðoutpatientsÞ :

This requires that each patient category should have at least 320 patients. (2) In the implementation of Taiwan PCCN demonstration project, the PCCN clinic physicians pointed out that one out of ten patients in one PCCN clinic could be recruited as

member patients in practice. Therefore, to secure the minimum 320 outpatients in our study for the category of member patients in the PCCN clinics, we estimated that the overall needed outpatient size for survey purposes should be 320  10 ¼ 3200ðtotal outpatients for three types of

clinic service contextsÞ :

(3) As mentioned earlier, there were 324 PCCN clinics and 324 non-PCCN clinics included in our study. As such, there should be 3200/(324 +324)ffi 5 outpatients for each studied clinic.

References Baggaley, A.R., 1982. Deciding on the ratio of the number of subjects to number of variable in factor analysis. Multivariate Experimental Clinical Research 6, 81–85. Baker, R., Mainous III, A.G., Gray, D.P., Love, M.M., 2003. Exploration of the relationship between continuity, trust in regular doctors and patient satisfaction with consultations with family doctors. Scandinavian Journal of Primary Health Care 21 (1), 27–32. Beck, R.S., Daughtridge, R., Sloane, P.D., 2002. Physician–patient communication in the primary care office: a systematic review. Journal of the American Board of Family Practice 15 (1), 25–38. Bliese, P., 2000. Within-group agreement, non-independence, and reliability. In: Klein, K., Kozlowski, S. (Eds.), Multi-level Theory, Research, and Methods in Organizations. Jossey-Bass, San Francisco, CA, pp. 349–381. Bonds, D.E., Foley, K.L., Dugan, E., Hall, M.A., Extrom, P., 2004. An exploration of patients’ trust in physicians in training. Journal of Health Care Poor Underserved 15 (2), 294–306. Borowsky, S.J., Nelson, D.B., Fortney, J.C., Hedeen, A.N., Bradley, J.L., Chapko, M.K., 2002. VA community-based outpatient clinics: performance measures based on patient perceptions of care. Medical Care 40 (7), 578–586. Bureau of National Health Insurance in Taiwan. /http://www.nhi.gov.tw/S (accessed 03.09). Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences. Lawrence, Hillsdale, NJ. Fan, V.S., Burman, M., McDonell, M.B., Fihn, S.D., 2005. Continuity of care and other determinants of patient satisfaction with primary care. Journal of General Internal Medicine 20 (3), 226–233. Institute of Medicine, 1996. Primary Care: America’s Health in a New Era. National Academy Press, Washington, DC. Jung, H.P., Baerveldt, C., Olesen, F., Grol, R., Wensing, M., 2003. Patient characteristics as predictors of primary health care preferences: a systematic literature analysis. Health Expectations 6 (2), 160–181. Lin, B.Y.J., 2007. Integration in primary community care networks (PCCNs): examination of governance, clinical, marketing, financial, and information infrastructures in a national demonstration project in Taiwan. BMC Health Services Research 7 (1), 90. Lin, B.Y.J., Wan, T.T.H., 2003. Creating Values for Health Care Organizations: A Guide toward Successful Integrated Care. Yeh Yeh Book Gallery, Taiwan. Marasculio, L.A., Levin, J.R., 1983. Multivariate Statistics in the Social Sciences. Brooks/Cole, Monterey, CA. Matusitz, J., Breen, G.M., 2007. Telemedicine: its effects on health communication. Health Communication 21 (1), 73–83. Polluste, K., Kalda, R., Lember, M., 2000. Primary health care system in transition: the patient’s experience. International Journal for Quality in Health Care 12 (6), 503–509. Razzouk, N., Seitz, V., Webb, J.M., 2004. What’s important in choosing a primary care physician: an analysis of consumer response. International Journal of Health Care Quality Assurance Incorporating Leadership in Health Services 17 (4–5), 205–211. Safran, D.G., 2003. Defining the future of primary care: what can we learn from patients?. Annals of Internal Medicine 138 (3), 248–255. Sampson, F., Munro, J., Pickin, M., Nicholl, J., 2004. Why are patients removed from their doctors’ lists? A comparison of patients’ and doctors’ accounts of removal. Family Practice 21 (5), 515–518. Scheffler, R.M., Weisfeld, N., Ruby, G., Estes, E.H., 1978. A manpower policy for primary health care. New England Journal of Medicine 298 (19), 1058–1062. Starfield, B., 1992. Primary Care: Concept, Evaluation, and Policy. Oxford University Press, New York. Wallace, L.S., DeVoe, J.E., Bennett, I.M., Roskos, S.E., Fryer Jr., G.E., 2008. Perceptions of healthcare providers’ communication skills: do they differ between urban and non-urban residents?. Health and Place 14 (4), 653–660. Wensing, M., Jung, H.P., Mainz, J., Olesen, F., Grol, R., 1998. A systematic review of the literature on patient priorities for general practice care. Part 1: description of the research domain. Social Science and Medicine 47 (10), 1573–1588. Zebiene, E., Razgauskas, E., Basys, V., Baubiniene, A., Gurevicius, R., Padaiga, Z., Svab, I., 2004. Meeting patient’s expectations in primary care consultations in Lithuania. International Journal for Quality in Health Care 16 (1), 83–89.