OUP user menu

Patient satisfaction with and recommendation of a primary care provider: associations of perceived quality and patient education

Yu-Chi Tung, Guann-Ming Chang
DOI: http://dx.doi.org/10.1093/intqhc/mzp006 206-213 First published online: 3 March 2009

Abstract

Objective To identify whether attributes of perceived clinic quality and patient education are associated with patient satisfaction and recommendation of a primary care provider.

Design Data used in this study were obtained through a national telephone survey by random sampling.

Setting Clinics throughout Taiwan.

Participants A total of 1910 patients.

Main outcome measures Overall patient satisfaction and recommendation were measured by single item questions. Attributes of clinic quality were measured using 11 items: doctor's technical skill (four items), doctor's interpersonal skill (three items), staff care and access (four items). Patient education was measured on the basis of education provided on disease prevention and control during the visit.

Results With regard to clinic quality, doctor's technical skill was most related to overall satisfaction and recommendation, followed by doctor's interpersonal skill. Staff care and access were associated with overall satisfaction but were not associated with recommendation. Patient education was related to both overall satisfaction and recommendation.

Conclusion Doctor's technical skill is the most critical attribute of primary care quality for both overall satisfaction and recommendation, followed by doctor's interpersonal skill. Staff care and access are associated with improved overall satisfaction but not related to increasing the likelihood of recommending a clinic to relatives and friends. Doctor's technical and interpersonal skills rather than staff care and access can be the essence of quality competition in the primary care market. Providing patient education during the visit on how to prevent or control diseases may also relate to improved patient satisfaction and recommendation.

Keywords
  • health insurance
  • patient outcomes
  • patient satisfaction
  • performance management
  • primary care
  • quality management
  • recommendation

Introduction

Patient satisfaction is important because it is taken into account when decisions are made about changes and improvement in services, and it is used as a tool in determining reimbursement rates, especially in the context of competitive healthcare environments and consumerism [14]. Improving patient satisfaction also increases the likelihood that a patient will recommend a given provider to friends and relatives when those individuals are looking for an appropriate provider [3].

Recommendation from family or friends becomes an important source of information for selecting a provider when this information is limited in many countries [4]. Although patient satisfaction and recommendation are correlated measures, recent studies find that a hospital with a high percentage of patient satisfaction does not necessarily receive a similar level of recommendation. Additionally, patient satisfaction and recommendation of a hospital or emergency department are influenced by somewhat different factors according to logistic regression analysis [4, 5]. Therefore, it is suggested that patient satisfaction and recommendation should be analysed separately in the allocation of limited resources to achieve the highest possible levels of patient satisfaction and recommendation [4, 5]. To our knowledge, studies on factors associated with patient satisfaction and recommendation of a clinic are relatively scarce in the literature.

There have been many patient satisfaction studies, the earliest of which attempted to identify patient characteristics such as age, gender and geographical region to predict patient satisfaction levels [6, 7]. Recent research analyses healthcare quality attributes, such as physician care, staff care, access or others to identify attributes that influence overall patient satisfaction [810]. It has also been reported that the technical and interpersonal skills of healthcare providers are two unique attributes involved in patient assessment of health care [1014].

A patient education approach has the potential to be a low cost and effective means of improving overall patient satisfaction [15], but few articles have examined the relationship between patient education and overall satisfaction with primary care. Gallefoss and Bakke [16] reported that for patients with chronic obstructive pulmonary disease (COPD), patient education seemed to improve overall patient satisfaction with general practitioners, but this was not true for asthmatics. Fan and colleagues [17] reported that, for diabetes, patient education was associated with satisfaction with primary care providers. These findings may not be generalizable to other patients owing to their specific settings or small number of participants.

This study, using a national survey from Taiwan, explores the associations of perceived quality and patient education with patient satisfaction and recommendation of a primary care provider.

Taiwan's Healthcare System

In Taiwan, the Bureau of National Health Insurance (BNHI), which is the sole insurer, has implemented national health insurance (NHI) for the entire population since March 1995. Each enrollee pays a premium and then enjoys comprehensive benefits with a low copayment policy (NT $50 [US $1.5] for the primary care). Each enrollee is free to go to any hospital or clinic because almost all providers have contracts with the BNHI.

In an attempt to control medical expenditure, since July 2001, the BNHI has adopted a primary care global budgeting system designed to replace a fee-for-service payment system. A yearly expenditure cap for primary care is determined by a fee negotiation committee. Every year, the committee under the supervision of the Department of Health negotiates and sets the total primary care budget (fixed budget) to be spent in the following year.

Reimbursement to clinics is based on an existing fee-for-service schedule, which lists each reimbursement service and its points of worth. A clinic's revenue is the product of total service points and the dollar value for each point. The dollar value for each service point is equal to the predetermined expenditure divided by the total service points. A higher service volume would have a lower dollar value per service point. Thus, a dollar value per service point is floating [18] and decreased to 0.8464 in 2004 [19].

Similarly to other countries implementing global budgeting to cut reimbursement, health care organizations confronting financial stress may decrease operating expenses or increase other revenues to preserve profits. For example, they may cut investment in staffing and equipment, reduce levels and availability of services and make patients wait longer to see a doctor or pay higher out-of-pocket prices. As a result, such changes in service provision may adversely affect patient satisfaction [2023].

On the other hand, as in the adoption of global budgeting, the BNHI has also promoted patient education on disease prevention and control, to decrease medical utilization by patients. For example, during the visit, physicians or nurses can explain to patients how to prevent or control the disease. Patient education during the visit may be supplemented by written material (such as handouts). Clinics can provide printed summary information to patients at the end of a visit.

Therefore, given the high degree of freedom of patient choice, personal recommendation is practically the only information available to patients or their family. Patient education is promoted in an environment of managed competition or cuts in reimbursement. Taiwan's healthcare system provides an excellent opportunity to examine and compare the associations of perceived quality and patient education with patient satisfaction and recommendation of a particular clinic.

Methods

Data source

Data for this study came from the Access to and Quality of Primary Care Survey in Taiwan. A total of 3310 primary care patients were selected as the target sample via stratified random sampling. First, the BNHI stratified clinics according to city and sampled clinics randomly within cities. Second, the BNHI randomly sampled patients who saw a physician within sampled clinics on 8 June 2005. The patients' information provided by the clinics included names, telephone numbers and addresses. Telephone interviews were conducted between 17 and 30 June 2005.

A total of 1910 patients of all ages were interviewed successfully, with a response rate of 58%. The majority of the failed cases was attributed to ‘going out’ and ‘busy’ in one follow-up; the ‘refused to answer’ cases accounted for 14% of the target sample. The sample was weighted to take account of its size in the primary care patient population in terms of city of residence, gender and age. These variables were chosen because national data on their joint distribution in the study population were available. After weighting, comparison of respondent characteristics (gender, age and city of residence) with the study population characteristics showed no significant difference.

Variables

Dependent variables

Two dependent variables were used in this study to indicate overall satisfaction and recommendation. The question measuring patient satisfaction was, ‘Generally speaking, were you satisfied with the overall quality of the clinic?’ The question for patient recommendation was, ‘When your family, relatives or friends need to see a doctor, would you recommend this clinic?’ The instrument used a five-point Likert scale that ranged from very dissatisfied/unwilling (1) to very satisfied/willing (5).

Independent variables

As regards perceived clinic quality, 11 items were used to measure the quality of care and services of the visit. After reviewing several measurement tools developed by previous researchers [3, 4, 14], we used these items to measure three major attributes of clinic quality: doctor's technical skill (medical skills, medical equipment, treatment outcome or recovery, and carefulness of diagnosis and treatment), doctor's interpersonal skill (explanation of diagnosis and treatment, answering your questions and attitude and time with the physician) and staff care and access (staff's explanation of care, clinic's cleanliness and comfort, length of time spent waiting to see the physician and out-of-pocket charges). The instrument also used a five-point Likert scale that ranged from very dissatisfied (1) to very satisfied (5).

The score for each attribute was computed as the mean of the items that measured the attribute. Missing values in each of the 11 items (from 0.3% to 6.4% of the total number of respondents) were replaced by the mean scores in order to maximize the amount of useful information [4, 14].

The question, ‘During the visit, did physicians or staff educate you about preventing relapse or deterioration of the disease?’ was asked to assess whether patient education was provided. A single dichotomous (yes/no) variable was created.

Other variables that might affect overall satisfaction and recommendation were selected in the light of the literature review [47, 14, 2023]. These included respondent type (patient/proxy), gender, age, location area, education, presence of severe or chronic disease (yes/no), first visit to the clinic (yes/no), difficulty in seeing a doctor on weekends (yes/no), registration fee and payment of expenses other than the copayment and registration fee (yes/no).

If patients were too young, too old or too weak to answer the questionnaire themselves, proxies who accompanied them to see the doctor could answer it. Severe or chronic diseases were identified according to patients' possession of certificates issued by the BNHI. Difficulty in seeing a doctor on weekends meant that patients might want to see a doctor but did not do so because the clinic was closed.

Statistical analysis

We used multiple logistic regression analysis to identify whether attributes of ‘perceived clinic quality’ and ‘patient education’ were related to patient satisfaction and recommendation. When analysing dependent variables, we dichotomized the responses by treating ‘very satisfied/willing and satisfied/willing’ as one group (coded 1) and the remaining responses (neutral, dissatisfied/unwilling, very dissatisfied/unwilling and no answer) as the other group (coded 0) in a logistic regression model based on prior research [4, 14]. The relative influence of these dependent variables was given as odds ratios (ORs) with 95% confidence intervals. The data were analysed using the SPSS statistical package (Release 16.0).

Results

Characteristics of the sample

More than half of the patients (54.7%) were female, about one-third (33.5%) were less than 20 years old, nearly 30% (29.2%) were located in the Taipei area and 46.3% had received no more than 6 years of schooling. A total of 19.7% of patients had severe or chronic disease, 14.7% visited the clinic for the first time and 67.0% responded to the interview themselves. The percentage of respondents who reported that it was difficult to see a doctor on weekends was 14.2%. The average registration fee was NT $62 (US $1.9). Nearly 10% (9.7%) paid expenses other than the copayment and registration fee.

The mean scores of doctor's technical skill, doctor's interpersonal skill and staff care and access were 3.9, 4.0 and 4.0, respectively. The percentage of respondents who reported that the clinic provided ‘education on disease prevention and control’ was 76.3% and 81.0% claimed they were ‘very satisfied or satisfied’ with the clinic care they received. The proportion of clinic recommendation (‘very willing’ and ‘willing’) was 78.9%. The proportion of ‘no answer’ responses for the recommendation question was 4.7%, whereas only 0.8% responded with ‘no answer’ for the satisfaction question (Table 1).

View this table:
Table 1

Sample characteristics, patient satisfaction and recommendation

Total
n1910
Gender, female (%)54.7
Age group, years (%)
 19 or less33.5
 20–2910.7
 30–3912.1
 40–4913.4
 50–5911.3
 60 and over19.0
Location area (%)
 Taipei29.2
 Northern13.1
 Central21.7
 Southern16.2
 Kao-Ping17.3
 Eastern2.5
Education, years (%)
 6 or less (primary)46.3
 7–9 (junior high)14.1
 10–12 (senior high)22.7
 13–16 (college)16.1
 ≥ 17 (postgraduate)0.8
Severe or chronic disease (%)19.7
First visit to the clinic (%)14.7
Respondent type, patient (%)67.0
Difficulty in seeing a doctor on weekends (%)14.2
Registration fee, NT$ (mean ± SD)61.9 ± 31.6
Paying expenses other than the copayment and registration fee (%)9.7
Doctor's technical skill (mean ± SD)3.9 ± 0.5
Doctor's interpersonal skill (mean ± SD)4.0 ±0.5
Staff care and access (mean ± SD)4.0 ±0.5
Providing education on disease prevention and control (%)76.3
Overall satisfaction with the clinic (%)
 Very satisfied/satisfied81.0
 Neutral17.4
 Dissatisfied/very dissatisfied0.8
 No answer0.8
Recommending the clinic (%)
 Very willing/willing78.9
 Neutral11.5
 Unwilling/very unwilling4.9
 No answer4.7

To confirm the convergent and discriminant validity of perceived clinic quality, a structural equation measurement model was built. Confirmatory factor analysis was conducted to analyse the factor structure of perceived quality. The goodness-of-fit statistics showed a good model fit: chi-square (χ2) = 18.30 (P = 0.69), root mean square error of approximation (RMSEA) = 0.001 < 0.05 and goodness of fit index (GFI) = 0.93 > 0.90. The reliability of each attribute was examined by means of the Cronbach's alpha coefficient. The obtained values of the reliability estimates were all >0.70, which indicates a reasonable internal reliability among items in the same attributes. All lambda weights (factor loadings) were statistically significant at α = 0.05, which indicates that all selected items loaded highly on the corresponding constructs or attributes.

Univariate analysis for overall satisfaction and recommendation

Overall satisfaction was associated with all patient characteristics excluding patient gender, respondent type and paying expenses other than the copayment and registration fee. The likelihood of recommending the clinic was related to patient gender, age, whether it was first visit, respondent type and difficulty in seeing the doctor on weekends (Table 2).

View this table:
Table 2

Patient satisfaction and recommendation by respondent characteristics

Overall satisfaction with the clinic (%)P-valueRecommend the clinic (%)P-value
Total81.078.9
Gender
 Male81.40.78876.30.035
 Female80.980.4
Age group, years
 19 or less82.10.00182.10.016
 20–2969.374.1
 30–3981.078.4
 40–4982.479.7
 50–5983.882.3
 60 and over83.473.8
Location area
 Taipei78.90.00679.00.892
 Northern74.276.5
 Central82.180.4
 Southern85.979.7
 Kao-Ping83.978.2
 Eastern80.978.3
Education, years
 ≤ 682.90.01779.30.905
 7–982.978.8
 10–1275.377.6
 13–1682.179.0
 ≥ 1781.386.7
Severe or chronic disease
 Yes85.40.01778.50.814
 No80.079.0
First visit to the clinic
 Yes68.2<0.00165.4<0.001
 No83.381.2
Respondent type
 Patient80.00.09277.50.031
 Proxy83.281.8
Difficulty in seeing a doctor on weekends
 Yes73.90.00170.3<0.001
 No82.280.3
Registration fee, NT$a
 < 5083.30.04876.90.336
 50–9982.480.1
 ≥ 10077.977.6
Paying the other expenses
 Yes78.40.33080.00.700
 No81.378.8
  • n = 1910.

  • aThe variable measuring registration fee was categorized into three levels using the first and the third quartiles as cut-off points.

Patients with higher scores of perceived quality (doctor's technical skill, doctor's interpersonal skill and staff care and access) were more likely to report being satisfied with the visit and to recommend the clinic to others. Patients who had been provided with education on disease prevention and control during the visit were more likely to report being satisfied and to recommend the clinic than those who had not been provided (86.2% versus 64.4% and 83.9% versus 62.8%, respectively) (Table 3).

View this table:
Table 3

Patient satisfaction and recommendation by perceived quality and patient education

Overall satisfaction with the clinic (%)P-valueRecommend the clinic (%)P-value
Total81.078.9
Doctor's technical skilla
 Low36.5<0.00146.7<0.001
 Medium91.185.6
 High98.992.7
Doctor's interpersonal skilla
 Low45.4<0.00153.9<0.001
 Medium93.686.6
 High97.792.0
Staff care and accessa
 Low49.2<0.00159.2<0.001
 Medium84.578.2
 High97.692.4
Providing education on disease prevention and control
 Yes86.2<0.00183.9<0.001
 No64.462.8
  • n = 1910.

  • aVariables measuring doctor's technical skill, doctor's interpersonal skill and staff care and access were categorized into three levels using the first and the third quartiles as cut-off points.

Results of the logistic regression models

In terms of overall satisfaction on perceived clinic quality, ‘doctor's technical skill’ showed the largest OR (19.76), followed by ‘doctor's interpersonal skill’ (5.44), and ‘staff care and access’ (2.44). Patient education was also significantly related to overall satisfaction, with an OR of 1.59.

In terms of recommendation on perceived clinic quality, ‘doctor's technical skill’ and ‘doctor's interpersonal skill’ were significant factors, with ORs of 5.06 and 2.16, respectively. Patient education was also significantly associated with recommendation (OR = 1.69) (Table 4).

View this table:
Table 4

Factors associated with overall satisfaction and recommendation of a clinic from logistic regression models (n = 1910)

Overall satisfactionaRecommendationb
ORP-value95% Confidence intervalORP-value95% Confidence interval
Doctor's technical skill19.76<0.00110.91–35.795.06<0.0013.17–8.07
Doctor's interpersonal skill5.44<0.0013.15–9.402.160.0011.40–3.34
Staff care and access2.44<0.0011.54–3.871.240.2530.86–1.81
Providing education on disease prevention and control (reference: no)1.590.0121.11–2.291.69<0.0011.27–2.24
Male (reference: female)1.050.7860.74–1.480.710.0110.55–0.92
Age, years (reference: ≤19)
 20–290.480.0410.23–0.970.700.2250.39–1.24
 30–390.940.8590.45–1.940.690.1970.39–1.22
 40–490.900.7770.45–1.810.690.1910.40–1.20
 50–590.650.2630.31–1.380.750.3300.41–1.34
 60 and over0.550.0790.29–1.070.420.0010.26–0.70
Location area (reference: eastern)
 Taipei1.660.3750.54–5.121.560.3120.66–3.70
 Northern0.950.9270.30–3.021.260.6170.51–3.10
 Central1.400.5620.45–4.371.490.3710.62–3.58
 Southern1.230.7240.39–3.911.080.8600.45–2.60
 Kao-Ping1.760.3300.56–5.521.220.6550.51–2.92
Education, years (reference: ≤6)
 7–91.320.3440.74–2.371.110.6440.72–1.70
 10–120.890.6700.52–1.531.240.3160.81–1.89
 13–161.570.1770.82–3.031.220.4250.75–1.98
 ≥ 170.950.9500.17–5.142.090.3920.39–11.27
Severe or chronic disease (reference: no)1.180.5100.72–1.950.920.6480.63–1.33
First visit to the clinic (reference: no)1.070.7520.69–1.670.670.0190.48–0.94
Proxy (reference: patient)1.180.5930.65–2.131.090.7140.70–1.70
Difficulty in seeing a doctor on weekends (reference: no)0.870.5710.55–1.390.730.0730.51–1.03
Registration fee, NT$1.000.4461.00–1.011.000.7851.00–1.01
Paying expenses other than the copayment and registration fee (reference: no)0.820.4860.47–1.441.350.2050.85–2.15
  • aLikelihood ratio test for model: χ2 = 895.00; P < 0.001; R2Nagelkerke = 0.60 and C-index = 0.90. Hosmer and Lemeshow test: χ2 = 9.18; P = 0.327.

  • bLikelihood ratio test for model: χ2 = 452.34; P < 0.001; R2Nagelkerke = 0.33 and C-index = 0.84. Hosmer and Lemeshow test: χ2 = 12.89; P = 0.116.

Discussion

This study provided an opportunity to examine whether the attributes of perceived quality and patient education were associated with overall satisfaction and recommendation of a particular clinic. Our findings show that doctor's technical skill, doctor's interpersonal skill and patient education seemed to be critical factors in patient satisfaction and recommendation of a clinic.

First, we found that subjects were more hesitant in responding to ‘recommendation’ questions (4.7% ‘no answer’) than to the ‘overall satisfaction’ questions (0.8% ‘no answer’). This finding is consistent with that of Cheng and colleagues [4] regarding inpatient care. People may feel more responsible for recommending a healthcare provider so they tend to skip the question when they are uncertain about the quality of the provider [4].

Second, concerning individual attributes of clinic quality, we found that doctor's technical skill, doctor's interpersonal skill, staff care and access had different magnitudes of correlation with overall satisfaction. This finding is similar to that of Otani and colleagues [3], who found that physician care was most influential, followed by staff care and access. In primary care settings, the physician is the focus of the patient's experience, so the physician's role becomes key. Physician care that most directly affects the patient's medical concern or health recovery is most associated with patient satisfaction [3].

Further, more specifically, we found that doctor's technical skill is more related to overall satisfaction and recommendation of a clinic than doctor's interpersonal skill. The result is a little different from that of Cheng and colleagues [4] regarding inpatient care. They found that doctor's interpersonal skill was more correlated with overall satisfaction, but doctor's technical skill was a more critical factor in recommendation. The difference may arise from settings (clinics versus hospitals) or disease complexity (simple versus complex diseases). Primary care patients may directly assess whether the physician can meet their needs by making the right diagnosis, providing effective treatment and curing their illness. Therefore, doctor's technical skill is the most important factor in overall satisfaction and plays a critical role in patient recommendation of a clinic to relatives and friends. Moreover, staff care and access are associated with overall patient satisfaction but are not associated with recommendation. Although patient satisfaction is one important outcome indicator, attracting new patients to select a provider depends mainly on patient recommendation [14, 24]. In other words, factors related to selecting a given healthcare provider depend on its core competence (doctor's technical skill and interpersonal skill) rather than staff care and access.

Third, this study found that providing education on disease prevention and control was related to patient satisfaction. This finding is similar to that of Gallefoss and Bakke [16] and that of Fan and colleagues [17] regarding COPD and diabetes. Owing, however, to their use of just a few settings or small sample size, their findings cannot be generalized. Additionally, we also verified the association between patient education and recommendation. One possible explanation for our findings is that people are increasingly concerned about how to promote their health, so they would be eager to absorb knowledge and to increase understanding of prevention and control of their diseases [15, 17], especially when they are sick. Thus, if clinics can provide information on disease prevention and control to satisfy patients' need, patients are not only satisfied but also willing to recommend them.

This study has several limitations that should be noted. First, owing to the constraints of a multi-purpose questionnaire administered by telephone survey, only a limited number of questions can be employed. Although it is better to include more items or variables, this might be a good starting point to examine the associations of perceived quality and patient education with patient satisfaction and recommendation of a clinic. Second, despite our use of a national survey, the generalizability of this study is limited because of the low response rate. In addition, we are unable to compare the sample distribution with that of the study population on other key variables such as education level or the presence of chronic disease because the study population data are unavailable.

Our study showed the associations of perceived quality and patient education with overall satisfaction and recommendation. Doctor's technical skill is the most critical attribute of primary care quality in both overall satisfaction and recommendation, followed by doctor's interpersonal skill. Staff care and access are associated with improved overall satisfaction but not related to increasing the likelihood of recommending a clinic to relatives and friends. Therefore, doctor's technical skill and interpersonal skill rather than staff care and access can be the essence of quality competition in the primary care market. The most important way to improve patient satisfaction and recommendation implies enhancing doctor's technical skill as top priority.

Lastly, patient education is associated with patient satisfaction and recommendation, so providing written or oral information during the visit on how to prevent or control diseases may relate to improved patient satisfaction and recommendation as well as doctor's technical and interpersonal skills, particularly as the healthcare system moves into an environment of managed competition or cuts in reimbursement.

Acknowledgements

The study was supported by grants from the Department of Health (DOH94-HI-1005) in Taiwan.

References

View Abstract