Determinants of patient satisfaction: a study among 39 hospitals in an in-patient setting in Germany

International Journal for Quality in Health Care 2011; Volume 23, Number 5: pp. 503 509 Advance Access Publication: 29 June 2011 Determinants of patient satisfaction: a study among 39 hospitals in an in-patient setting in Germany TONIO SCHOENFELDER 1, JOERG KLEWER 2 AND JOACHIM KUGLER 1 1 Department of Public Health, Dresden Medical School, University of Dresden, Loescherstrasse 18, 01309 Dresden, Germany, and 2 Department of Public Health and Health Care Management, University of Applied Sciences Zwickau, Dr.-Friedrichs-Rings 2A, 08056 Zwickau, Germany Address reprint requests to: Tonio Schoenfelder, Department of Public Health, Dresden Medical School, University of Dresden, Loescherstrasse 18, 01309 Dresden, Germany. Tel: þ49-37421-20396; Fax: þ49-3761-186720; E-mail: tonio.schoenfelder@gmx.de Accepted for publication 30 May 2011 10.1093/intqhc/mzr038 Abstract Objective. To identify key determinants of patient satisfaction. Design. Data used were obtained through a self-administered, post-visit questionnaire by random sampling during the period of January 2009 to September 2009. Setting. Thirty-nine hospitals in Germany. Participants. A total of 8428 patients. Main Outcome Measure. Global patient satisfaction was measured by a single item question. Attributes of medical aspects of care were measured using 12 items, performance of service using 3 items and different dimensions of patient expectations using 12 items. Medical aspects of care and performance of service items were entered into logistic regression analysis to identify determinants of patient satisfaction. Results. The results of the analysis showed that there are 10 determinants of global patient satisfaction. The outcome of treatment was overall, the most salient predictor followed by nursing kindness as the second most important component. Items reflecting information receiving about the undergoing treatment do not have a major influence on patient satisfaction. Conclusion. The analysis identified key determinants that should be altered first in order to improve global patient satisfaction. The results also indicate that some aspects of the hospital stay are not seen as relevant by patients and therefore are unrelated to satisfaction ratings. The findings suggest that variables measuring patients perceptions of care are more important determinants of global patient satisfaction in comparison to demographics and visit characteristics. Results of the present study have implications for health providers aiming at improving the service quality and quality of care. Keywords: health services research, patient satisfaction, quality management, hospital medicine, survey research Introduction Satisfaction can be described as a patient s reaction to several aspects of their service experience [1]. Patients thereby evaluate the health-care services as well as the providers from their own subjective point of view [2, 3]. There are several motivations for surveying patient satisfaction. It may influence health-care utilization, can be a predictor of subsequent health-related behavior [1, 4] and whether patients are willing or not to recommend their health-care provider to others [1, 3]. Patient satisfaction is a useful measure in assessing patterns of communication. Even though patients may not be able to judge specific technical aspects, they provide the best source of accurate information regarding clarity of explanations, helpfulness of information patients are receiving, barriers to obtaining care or the physician s interpersonal behavior [5, 6]. Though numerous works have addressed this topic, patient satisfaction remains difficult to determine [7]. It is a multidimensional concept consisting of various aspects which all of which do not necessarily have to do with the actual quality of care and service experienced by the patient. One criticism of patient satisfaction ratings has been the inability to account for International Journal for Quality in Health Care vol. 23 no. 5 # The Author 2011. Published by Oxford University Press in association with the International Society for Quality in Health Care; all rights reserved 503

Schoenfelder et al. expectations about medical care, which may be influenced by prior experiences with the health-care system [8]. In addition, patient characteristics also have an impact: older patients and those with lower levels of education appear to be more satisfied [9]. Furthermore, psychosocial determinants play a role in the sense that patients report greater satisfaction than they actually feel because they fear negative consequences in case they give negative feedback [10]. In Germany, measuring satisfaction has been mandatory since 2005 as an element of quality management reports regarding service quality and quality of care which have to be published by all providers aiming at supporting patients with information in order to allow benchmarking of hospitals [11]. As a consequence, several questionnaires have been used to evaluate in-patient care. Dimensions of what is being measured differ between these instruments. Therefore, as long as hospitals do not use the same instrument, each published patient satisfaction data reflects only the service quality and quality of care of its respective hospital. As a solution, a patient satisfaction survey among all 39 hospitals of the administrative district of Dresden was conducted to measure the outcome of in-patient health services of the entire area and to identify key determinants that predict higher patient satisfaction. Method Research setting and data collection The data set used in this study was obtained through a selfadministered, post-visit questionnaire during the period of January 2009 to September 2009. The survey was conducted in the district of Dresden (Germany) and included all 39 hospitals of the area. Participants were recruited from six departments: surgery, internal medicine, urology, neurology, ophthalmology and pediatrics. In 2008, about 245 000 patients were hospitalized in these six departments and 39 hospitals. Survey participants were policy-holders of four statutory health insurances, which together have a market share of 85% of the region s total population. In order to include all area hospitals in the survey and to ensure all patients receive the same questionnaire, the statutory health insurances were chosen as source of contact because there was the potential risk some hospitals would deny participation due to the additional expenditure in organization and time. Study participants were randomly selected based on gender, age and the market share of their health insurance. Data source was the health insurances internal data pool. Sample selection criteria included only individuals who spent at least one night or more in the hospital throughout 2008. Participation was completely anonymous and voluntary. The questionnaire was accompanied by a cover letter informing the participants about the purpose of the study, and about their consent to participate when sending back the questionnaire. In total, 31600 post-paid surveys were distributed; 10 045 were finally completed, resulting in a response rate of 32%. Out of the total number of completed surveys, 1617 questionnaires were considered problematic due to excessive missing and ambiguous data. Thus, these questionnaires were excluded and only 8428 surveys were finally analyzed. Questionnaire design The questionnaire represents a modified version of a survey instrument used by one of the statutory health insurances, which in turn was developed on the basis of a review of patient satisfaction literature, published instruments and verbatim patient responses to questions about hospital quality. Translating foreign survey instruments seemed inappropriate because of substantial structural differences between healthcare systems. The original instrument has been used in two previous regional surveys by the statutory health insurance; however, results were not published. For the present study, the order of questions was modified to make the questionnaire fit on two pages and one question was added which clarified in which department the patient was treated. The modified questionnaire was submitted to 38 discharged inpatients to ensure that wording, format length and sequencing of the questions were appropriate. The 37-item questionnaire was drafted in German and enquired about patient demographics and visit characteristics of the hospital stay (Table 1). The instrument also collected data regarding medical and service attributes of the undergone treatment, patients expectations of forthcoming healthcare events, importance of hospital setting and the kind of recommendation on a hospital patients would prefer. A description of the items is shown in Table 2. One item collected information concerning global satisfaction with the hospital stay in general. Patients responded on a six-point scale ranging from excellent/absolutely necessary to very poor/completely unnecessary. For data analyses, the lowest rating was coded with 1, the highest with 6. The questionnaire did not contain a billing dimension, as it would not apply to the German system, in which patients are not required to pay for hospitalizations. Analysis To identify dimensions of the questionnaire, factor analysis with Varimax rotation was performed. The Kaiser-Meyer- Olkin measure of sampling adequacy was high (0.90), indicating the matrix was well suited for factor analysis [12]; 27 items were submitted. Factor loadings of 0.4 or greater were considered significant for defining the factors. Each item was attributed to the factor on which it had the highest loading. Six components (eigenvalues.1), which accounted for 63% of the total variance, were identified (Table 2). Cronbach alpha was used in order to estimate the internal consistency of the subscales resulting in scores of 0.60 0.92. The distribution of the satisfaction scores was skewed toward higher satisfaction. Therefore, the non-parametric Kruskal Wallis test was used to assess the effects of age and 504

Determinants of patient satisfaction Patient experience Table 1 Characteristics of the patient sample Variable Number of patients (%)... Gender 5774 (100) Male 2995 (51.9) Female 2779 (48.1) Missing 2654 Age 8373 (100),15 634 (7.6) 15 20 144 (1.7) 21 30 363 (4.3) 31 40 384 (4.6) 41 50 706 (8.4) 51 60 1168 (13.9) 61 70 1859 (22.1) 71 80 2270 (26.9) 80þ 845 (10) Missing 55 Quantity of hospitalizations 8205 (100) within previous 5 years 1 2 5091 (62) 3 5 2372 (29).5 742 (9) Missing 223 Department 7968 (100) Surgery 2699 (33.9) Internal medicine 2301 (28.9) Urology 1091 (13.7) Neurology 781 (9.8) Pediatrics 589 (7.4) Ophthalmology 507 (6.4) Missing 460 Source of admission 8381 (100) Specialist 3368 (40.2) General practitioner 2560 (30.5) Emergency 1881 (22.4) Self-admission 472 (5.6) Transfer from another clinic 100 (1.2) Missing 47 Length of stay 8243 (100) Too short 782 (9.5) Appropriate 5704 (69.2) Too long 509 (6.2) Do not know 1248 (15.1) Missing 185 Complications 8245 (100) Yes 919 (11) No 7326 (89) Missing 183 Use facility again? 8296 (100) Yes 6819 (82.2) No 440 (5.3) Do not know 1037 (12.5) Missing 132 gender on global satisfaction ratings. General associations between satisfaction ratings and visit characteristics were investigated with x 2 -tests and Fisher s exact test when cell counts were small. All associations were considered significant at P, 0.05 level. In order to assess determinants of patient satisfaction, logistic regression analysis with stepwise backwards likelihood ratio testing for model selection was performed. The dependent variable was the patient s global satisfaction rating for the complete hospital stay; it was dichotomized into excellent/good (reference) versus fair/acceptable/poor/ very poor. Independent variables were entered, which mirrored the patients experiences with the hospitals services; these were items of the factors medical aspects of care and performance of service. Therefore, variables addressing patient expectations were not entered. Additionally, all variables significant by bivariate analysis were included. The performance of the estimated model was tested on a validation sample. Therefore, the data set was divided into two groups: 70% of the cases were used for model development and the remaining data were used as validation sample for model testing. Because some of the analyzed questionnaires contained missing values, and therefore were excluded in the logistic regression, 4381 cases were used for model development and 1271 for validation. Data were analyzed using SPSS 15.0 for Windows (SPSS Inc., Chicago, IL). Results The majority of the study sample was male, aged 71 80 and reported 1 2 hospitalizations within previous 5 years. Approximately 3368 (40%) participants were sent to hospital by a specialist and 2560 (31%) by their GP. A minority of respondents (919, 11%) reported post-discharge complications and 5704 (69%) assessed length of stay to be appropriate. Approximately 6819 (82%) patients would use the facility again in the case of another hospitalization (Table 1). Satisfaction ratings In total, 6595 (80%) patients rated the global satisfaction with hospital stay related to all performed services either excellent or good. The grouped median score was 5.04. Kindness of the hospitals physicians (5.36) and nurses (5.34) reached the highest scores, while clear information about anesthesia (4.64) and medication (4.50) received the lowest ratings (Table 3). Associations between satisfaction and patient demographics and visit characteristics Gender, department, source of admission, and number of previous hospitalizations were not associated with global satisfaction. Patients age, however, was related to level of 505

Schoenfelder et al. Table 2 Results of the factor analysis and reliability testing Factors and abbreviated items (number of items) Factor loadings Variance explained (in %) Cronbach alpha... Factor I: Medical aspects of care (12) 29.16 0.92 Individualized medical care 0.84 Clear reply of inquiries by physicians 0.82 Clear information about undergoing operations 0.82 Clear information about medication 0.77 Discharge procedures and instructions 0.76 Clear information about anesthesia 0.70 Physician s knowledge of patient anamnesis 0.69 Kindness of the hospital s physicians 0.67 Organization of procedures and operations 0.66 Outcome of treatment 0.65 Efficiency of admitting procedure 0.56 Kindness of the hospital s nurses 0.51 Factor II: Service expectations (4) 13.54 0.82 Expectations of nurses kindness 0.83 Expectations of physicians kindness 0.82 Expectations of quality of food 0.74 Expectations of accommodation 0.73 Factor III: Performance of service (3) 6.71 0.71 Accommodation 0.76 Cleanliness 0.74 Quality of food 0.65 Factor IV: Setting (3) 4.81 0.77 Distance between clinic and place of residence 0.91 Location and accessibility 0.89 Clinic size 0.51 Factor V: Recommendation (3) 4.75 0.60 General practitioner 0.80 Specialist 0.79 Relatives and friends 0.48 Factor VI: Expectations of care (2) 3.92 0.68 High quality care 0.83 Clear information about undergoing treatment 0.78 satisfaction (P, 0.001). A comparison of the grouped median values of the global satisfaction score showed that younger patients of age groups,15, 15 20 and 21 30 rated this aspect with 4.43, 4.72, 4.66, respectively. Among the other patients, the global satisfaction score in the age group of 31 40 consistently decreased from 4.93 to 5.10 in the age group of.80 years. Study participants rating global satisfaction from fair to very poor reported significant lower willingness to be hospitalized in the same clinic again (P, 0.001). Patients reporting postdischarge complications were less satisfied with their complete hospital stay in comparison to patients without complications (P, 0.001, 4.49 with complications versus 5.16 without), were also less willing to be hospitalized again in the same clinic (P, 0.001) and reported length of stay to be too short (P, 0.001). Results of the multivariate logistic regression analysis The analysis revealed 10 determinants of global patient satisfaction (Table 4). The most influential determinants were outcome of treatment (OR 3.70) and kindness of the hospital s nurses (OR 2.78) and physicians (OR 1.96). Four variables of the medical aspect of care factor were not included in the regression model: clear reply of inquiries by physicians, clear information about medication, anesthesia and undergoing operations. Regarding the performance of service component, lower ratings of accommodation and quality of food resulted in decreasing global patient satisfaction. Cleanliness was the only service aspect not included in the model. The results of the model development sample were similar to those of the validation sample: 92.7% versus 506

Determinants of patient satisfaction Patient experience Table 3 Levels of global patient satisfaction, satisfaction with medical aspects of care and satisfaction with performance of service; 6.0 ¼ best, 1.0 ¼ worst Abbreviated item content Grouped median... Global satisfaction with hospital stay 5.04 Medical aspects of care Kindness of the hospital s physicians 5.36 Kindness of the hospital s nurses 5.38 Efficiency of admitting procedure 5.23 Outcome of treatment 5.17 Clear reply of inquiries by physicians 5.14 Individualized medical care 5.10 Organization of procedures and operations 5.04 Physician s knowledge of patient anamnesis 4.85 Clear information about undergoing operations 4.74 Discharge procedures and instructions 4.66 Clear information about anesthesia 4.64 Clear information about medication 4.50 Performance of service Cleanliness 5.34 Accommodation 5.14 Quality of food 4.98 92.2% of all cases were assigned correctly. The Nagelkerke-R 2 resulted in 0.72. Table 4 Results of the multivariate logistic regression analysis: determinants of global patient satisfaction Variable OR (95% CI) P-value... Outcome of treatment 3.70 (3.10 4.48),0.001 Kindness of the hospital s 2.78 (2.19 3.54),0.001 nurses Kindness of the hospital s 1.96 (1.51 2.56),0.001 physicians Organization of procedures 1.68 (1.39 2.00),0.001 and operations Quality of food 1.60 (1.40 1.92),0.001 Accommodation 1.50 (1.25 1.79),0.001 Individualized medical care 1.46 (1.22 1.75),0.001 Discharge procedures and 1.43 (1.23 1.66),0.001 instructions Physician s knowledge 1.38 (1.15 1.65),0.001 of patient anamnesis Efficiency of admitting 1.25 (1.10 1.50) 0.018 procedure Clear information about / undergoing operations Clear information about / anesthesia Clear information about / medication Clear reply of inquiries by / physicians Cleanliness / n ¼ 4381, Nagelkerke R 2 ¼ 0.72. /, non-significant variable. Post-visit expectations, recommendation of clinic and setting After their hospital stay, patients attached the highest priority to high quality care (5.77), followed by clear information about their undergoing treatment (5.62) as well as nursesapos; (5.43) and physicians kindness (5.41). Patients attached least importance to quality of food (4.80) and accommodation (4.74). Recommendation by relatives or friends on a certain hospital (3.39) was less important than recommendations by specialists (5.23) and GPs (4.88). Regarding setting, patients attached least importance to clinic size (3.26), followed by distance between clinic and place of residence (4.55), and location (4.73). Discussion The study identified key determinants of patient satisfaction in the investigated in-patient setting. Most prior research analyzed data gained from patients of a very limited number of hospitals. Analyzed data in this study were obtained from randomly selected patients of all 39 Dresden area hospitals using the same method; therefore, results are an outcome measure for the health services of the entire district. Main findings The results of the regression analysis showed that there are 10 determinants of global patient satisfaction. Thereby, both medical and service elements of the hospital stay were important dimensions. The outcome of treatment was overall the most salient predictor of global patient satisfaction followed by nursing and physicians kindness. These results are consonant with prior research which highlighted the importance of communication between patients and hospital staff [10, 13]. The major impact of nursing kindness and its greater meaning in comparison to physicians kindness corresponds with other study results [14, 15], which found communication with nurses and nursing care to be more important to patients. Possibly, patients experience more contact with nurses than doctors as nurses are the first responders to patients if they feel discomfort or have general questions [15]. Both organization of admission and discharge were associated with global patient satisfaction. Patients attached most importance to discharge procedures and instructions, which is partly inconsistent with study research of Elliott et al. [15], who found discharge information to be of least importance to patients. However, the strength of this association was 507

Schoenfelder et al. rather small compared with most of the other variables in the regression model. Items reflecting information receiving such as information about anesthesia, medication or undergoing operations did not have a major influence on patient satisfaction, although patients indicated these aspects would be highly relevant for future hospital stays. These results may suggest that patients lack of medical knowledge could have been influential, resulting in patients are not being able to judge if the received information is correct. This possibly also explains the heavy influence of the two service components, quality of food and accommodation on global satisfaction. Possibly, patients look for surrogate indicators of correct diagnosis and treatment to measure their own satisfaction [3]. However, cleanliness was not a predictor of global patient satisfaction. Regarding gender and age, the study results are consistent with most prior research that found similar satisfaction scores among men and women [9, 16, 17] and older patients being more satisfied with received services than younger individuals [9, 18]. One possible reason regarding the higher satisfaction rating of older study participants could be that older patients may be treated differently, e.g. more gently than younger ones [19]. However, the relation between global satisfaction and age faded when examining with multivariate techniques. Prior research indicated that a low health status leads to lower satisfaction scores [8, 20], which concurs with the results of the present study. Complications had a significant influence on different aspects of satisfaction in the bivariate analysis. Patients reporting about post-discharge complications were less likely to rate satisfaction regarding service and medical aspects of care as excellent or very good. Patients with experienced postdischarge complications significantly reported length of stay as too short. These patients judgments may have been affected by previous intensive political discussions, assuming that with the introduction of the diagnosis- related groups in Germany length of stay would decrease to a dimension that would deteriorate the quality of care. Methodology Patients of all 39 hospitals of one geographical area from six different departments were included in the data set and were randomly selected. Sample selection was based on the internal data pool of the participating statutory health insurances, covering 85% of the total population of the area in which the study was conducted in. Findings are representative for in-patients of the six departments in the investigated area. The 32% response in the present study may introduce a potential selection bias that could affect generalizability and validity of the results. However, Lasek et al. [21] found that the impact of non-response bias on satisfaction surveys of hospitalized patients might be relatively small. As the effect of non-response was not assessed in the present study, it is not possible to draw conclusions whether there were any differences in demographic or clinical characteristics. Regarding the high levels of satisfaction, there is a risk of an acquiescence and a social desirability bias that may have resulted in an over-reporting of satisfied patients [10, 22]. Another explanation could be the use of the single-item questions. Most patients tend to give positive answers if they are asked how satisfied they were [23, 24] even though if they have complaints about specific aspects of the received care [6]. It is unknown if proxy responses were involved in answering the questionnaire. This may represent another source of bias because patients receiving assistance when completing a questionnaire were found to be less satisfied than individuals who did so on their own [25]. In this study, a brief and feasible instrument was used. However, when revising the questionnaire, additional variables addressing patient demographics and visit characteristics should be included. For example, while the personal evaluation length of stay was recorded, the actual length of stay should also be included. Possibly, patients with a prolonged stay due to more serious conditions would have made different experiences on which to base their judgment compared with patients with a one-night stay. With exception of the factor recommendation, the internal consistency was satisfactory. To perform the regression analysis, the data set was divided into a development and a validation sample to avoid an overestimation of the parameters. Gained results were similar, indicating the variables serve as reliable determinants of global patient satisfaction. The logistic regression model explains 72% of the observed variation, indicating a very good data fit [26]. In summary, the present study identifies determinants of patient satisfaction and delivers information about aspects of the hospital stay that are not seen as relevant by patients. Findings are based on data gained from a large number of hospitals from one geographical area. Specifically, the degree to which patients were satisfied with the kindness of nurses and physicians showed strong predictive utility for overall patient satisfaction, while information regarding technical quality of care was not related to satisfaction ratings. Additionally, amenities which are not directly linked to the received care (such as accommodation) were of much importance to patients. Therefore, future research should put emphasis on collecting information about additional hotel aspects, regardless if these are important to patients or represent proxies for other aspects of the hospital stay. The results also suggest that variables measuring patients perceptions of care are more important determinants of global patient satisfaction in comparison to demographics and visit characteristics. Found variables are alterable and, therefore, provide implications for health providers aiming at improving the service quality and quality of care. Acknowledgements The authors thank the patients for their participation and also gratefully acknowledge Gregor Schoenfelder for his help with manuscript preparation. 508

Determinants of patient satisfaction Patient experience Funding This work was supported by the statutory health insurances Allgemeine Ortskrankenkasse Plus; Techniker Krankenkasse; Innungskrankenkasse classic and Barmer GEK. References 1. Pascoe GC. Patient satisfaction in primary health care: a literature review and analysis. Eval Program Plann 1983;6:185 210. 2. Ware JE, Snyder MK, Wright WR et al. Defining and measuring patient satisfaction with medical care. Eval Program Plann 1983;6:247 63. 3. Otani K, Kurz RS, Harris LE. Managing primary care using patient satisfaction measures. J Healthc Manag 2005;50:311 24. 4. Fitzpatrick R. Surveys of patient satisfaction: I important general considerations. BMJ 1991;302:887 9. 5. Savage R, Armstrong D. Effect of a general practitioner s communication style on patients satisfaction: a controlled study. BMJ 1991;301:968 70. 6. Epstein KR, Laine C, Farber NJ et al. Patients perceptions of Office Medical Practice: judging quality through the patients eyes. Am J Med Qual 1996;11:73 80. 7. Jackson JL, Chamberlin J, Kroenke K. Predictors of patient satisfaction. Soc Sci Med 2001;51:609 20. 8. Zhang Y, Rohrer J, Borders T et al. Patient satisfaction, selfrated health status, and health confidence: an assessment of the utility of single-item questions. Am J Med Qual 2007;22:42 9. 9. Hall JA, Dornan MC. Patient sociodemographic characteristics as predictors of satisfaction with medical care: a meta-analysis. Soc Sci Med 1990;30:811 8. 10. Sitzia J, Wood N. Patient satisfaction: a review of issues and concepts. Soc Sci Med 1997;45:1829 43. 11. GKV-Spitzenverband, Germany. https://www.g-qb.de/ (S(dub4m43pmwgzoy55mgol0d55))/gesetzliche.Grundlagen.aspx (12 November 2009: legal requirement of quality reports in Germany). 12. Kaiser HF, Rice J. Little Jiffy, Mark IV. Educ Psychol Meas 1974;34:111 7. 13. Sofaer S, Crofton C, Goldstein E et al. What do consumers want to know about the quality of care in hospitals? Health Serv Res 2005;40:2057 77. 14. Cleary PD, Keroy L, Karapanos G et al. Patient assessment of hospital care. QRB Qual Rev Bull 1989;15:172 9. 15. Elliott MN, Kanouse DE, Edwards CA et al. Components of care vary in importance for overall patient-reported experience by type of hospitalization. Med Care 2009;47:842 9. 16. Etter J, Perneger TV. Quantitative and qualitative assessment of patient satisfaction in a managed care plan. Eval Program Plann 1997;20:129 35. 17. Garcia-Aguilar J, Davey CS, Le CT et al. Patient satisfaction after surgical treatment for fistula-in-ano. Dis Colon Rectum 2000;43:1206 12. 18. Grogan S, Conner M, Norman P et al. Validation of a questionnaire measuring patient satisfaction with general practitioner services. Qual Health Care 2000;9:210 5. 19. Car-Hill R. The measurement of patient satisfaction. J Public Health Med 1992;14:236 49. 20. Jaipaul CK, Rosenthal GR. Are older patients more satisfied with hospital care than younger patients? J Gen Intern Med 2003;18:23 30. 21. Lasek RJ, Barkley W, Harper DL et al. Nonresponse bias on patient satisfaction surveys. Med Care 1997;35:646 52. 22. Nguyen TD, Attkisson CC, Stegner BL. Assessment of patient satisfaction: development and refinement of a service evaluation questionnaire. Eval Program Plann 1983;6:299 314. 23. van Campen C, Sixma HJ, Kerssens JJ et al. Assessing noninstitutionalized asthma and COPD patients priorities and perceptions of quality of health care: the development of the QUOTE-CNSLD instrument. J Asthma 1997;34:531 8. 24. Williams B. Patient satisfaction: a valid concept? Soc Sci Med 1994;38:509 16. 25. Quintana JM, González N, Bilbao A et al. Predictors of patient satisfaction with hospital care. BMC Health Serv Res 2006;6:102. 26. Backhaus K, Erichson P, Plinke W et al. Logistische regression. In Multivariate Analysemethoden: Eine anwendungsorientierte Einfuehrung (in German). Berlin: Springer, 2008. 509