Patient satisfaction with outpatient neurology services: A momentum for improvement

Results 212 patients with mean age of 40.1 and a 1:1 M: F ratio completed the survey. The variation of overall patient satisfaction (mean, 70.4; SD, 12.4) was independently predicted by patient clinical outcome expectations and satisfaction on waiting area, overall service of doctor and card room [R 2 = 0.305; F (8,195) = 10.685, p = 0.000]. Mean satisfaction scores for specific dimensions of the outpatient general neurology clinic ranged from 57.2 for waiting time at the clinic to 74.0 for overall service of the guards. Waiting time at the clinic stood first among the top five priority indices. Conclusion This survey demonstrates predictors of overall patient satisfaction with the outpatient neurology services, and delineates priority areas warranting further improvement. It is the first African study on patient satisfaction with neurology services, and provides a guide for neurological or other specialty clinics seeking to improve and expand medical services. Keywords Patient satisfaction Patient care experience Outpatient neurological service Satisfaction score Healthcare quality 1 Introduction There are no available reports on the level of patient satisfaction with the rapidly expanding neurological services in Ethiopia. The local parameters of patient satisfaction with neurological healthcare have not even been defined. This quantitative survey of the Addis Ababa, Ethiopia outpatient general neurology clinic (OGNC) will delineate the overall level of patient satisfaction, provide an assessment of current services, and form the foundation for improved expansion of neurological care. Patient satisfaction or dissatisfaction characterizes an expression of attitude on healthcare services [1,2] . It represents one factor in the overall tenuous measurement of healthcare quality [1–3] . Other considerations influencing patient satisfaction include sociodemographic characteristics, health status and service expectations [1,4,5] . A high level of patient satisfaction with healthcare is desirable because it promotes treatment adherence, service utilization and loyalty to the provider [2,4,5] . In developed countries healthcare evaluations have adapted to the consumerism trend by incorporating patient opinion of quality in order to address commercial applications [1,6] . In contrast, healthcare improvement in developing countries has traditionally focused on quantity rather than quality [7] . However, this dichotomy is rapidly eroding as valid quality assessment techniques are increasingly adopted, allowing the developing world to appreciate the potential economic benefits stemming from improved healthcare quality [7] . Quality improvement is, in fact, the most rapid, cost-effective method of improving the overall community health status in developing countries. [7] . Outcome measures of patient satisfaction and patient responsiveness are increasingly accepted as an integral component of the overall healthcare quality assessment in these regions [7–11] . In Ethiopia, neurological specialty care is limited to the capital city and partially integrated with other services. The dimensions of patient satisfaction at the Addis Ababa OGNC have not been assessed. Indeed, there are no published reports on patient satisfaction with neurology services anywhere in Africa. The global literature on outpatient neurology centers outside of Africa describes 60% or greater patient satisfaction with the overall service [12–17] . 2 Methods 2.1 Study setting This survey was performed in the Tikur Anbessa Specialized Hospital (TASH) OGNC in Addis Ababa, Ethiopia. It is a teaching hospital affiliated with Addis Ababa University Medical Faculty (AAU-MF). The OGNC encompasses three examination rooms with corridor waiting area, convenes twice weekly (Tuesday afternoon and Thursday morning), and is staffed by consultant neurologists, neurology residents, rotating internal medicine residents, and four permanent neurological nurses. 2.2 Study population The survey study population incorporates all patients attending the OGNC. These patients come from all regions of Ethiopia, and include (1) established patients (around 3500); (2) new patients referred from TASH; and (3) new patients referred from other national tertiary facilities. These latter patients (3–4 patients in each clinic session) are first screened by the TASH Internal Medicine Department prior to scheduling an initial neurological consultation. Appointment dates for all patients are memorialized in the registration book. An average of sixty patients are seen in each clinic session (estimated 7 month time frame for a visit to every patient). 2.3 Study design and sampling strategy The study utilized a cross-sectional analytical sample survey design and three-stage cluster sampling technique. Sample size, calculated using the following formula: n = (Z l − α/2 |ω) 2 . 〚π( l − π)〛(where Z l − α/2 = 1.96 at 95% confidence,ω = confidence interval width and π = population mean) and adjusted for the non-response rate (15%) and design effect (ad hoc value of 2) was 226. This allowed estimation of population proportion from a single sample at a desired precision of 10 percentage points around an expected overall patient satisfaction score of 0.50 with a two-sided 95% confidence level. We presumed maximum variability (Π = 0.50), and the calculated sample size was inflated to adjust for the complex study design (design effect) [18,19] . Two stages of simple random subsampling (two clusters at each level) were carried out to select the frame population (total of 14 clinics over 7 weeks). A simple random sampling (lottery method) from the registration book appointment roster of the frame population served to recruit the survey participants during each clinic session. On average 17 patients were recruited from each clinic session to achieve the required sample size. The potential survey participants included all OGNC patients 18 years of age and older. 2.4 Data collection tools and procedure Data were collected by structured interview and data abstraction form. The questionnaire measured patients' evaluations and affective responses to all possible specific dimensions of the OGNC healthcare, as well as an overall global service satisfaction. The construction of a 22-item English language based patient satisfaction scale was based on the authors' background knowledge of the survey target population, previous measures and review of the literature [12–17,20,21] . The questionnaire items were then repeatedly translated into local Amharic language through several revisions and back translations until all investigators agreed that the two versions were comparable. Prior to the study our version was cognitively pretested on 30 randomly selected OGNC patients to ensure appropriate wording and formatting of each question. On the final version of the questionnaire sampled patients were privately interviewed post-service by three trained non-health professional high-school graduates. The same data collector performed all of the patient interviews throughout the survey period. The questionnaire incorporated patient demographics (age, gender, residential address and service payment status) and attitudes regarding the OGNC healthcare. Patients reported their attitudes on overall as well as specific healthcare elements (overall service, waiting time for appointment visit, waiting time at the clinic, waiting area convenience, examination room convenience, doctor's attention to concerns and problems, and overall service of nurses, doctors, guards and card room), and clinical outcome expectation of the current visit on a four-point Likert-like scale (e.g. 1 = very dissatisfied, 2 = dissatisfied, 3 = satisfied and 4 = very satisfied). An open-ended item encouraged additional comments. The principal investigator and one of the data-collectors abstracted the medical charts and registration book to ascertain the appointment interval in days for initial (from referral to evaluation for new patients) and follow-up (from prior to current evaluation for established patients) visits, and the neurological diagnosis of each interviewed patient. Coding of patient satisfaction questionnaires assisted in linkage with abstracted data. 2.5 Statistical analysis Data were organized on a spreadsheet. Diagnoses were categorized in accordance with the International Statistical Classification of Diseases and Related Health Problems version 10 (ICD-10) [22] . The open-ended item responses were grouped into positive, neutral (no or mixed opinion) and negative comments regardless of their specific nature. Dichotomization were performed for service payment status (free- and pay-for-service), residential address (Addis Ababa and non-Addis Ababa) and expectations of clinical outcome [do not expect (1 = very unlikely and/or 2 = unlikely) and expect (3 = likely and/or 4 = very likely)]. The four-point scale responses were converted to 100-point maximum interval scale responses to provide a common yardstick for performance comparison of different OGNC healthcare characteristics [very unsatisfied (1) = 25, unsatisfied (2) = 50, satisfied (3) = 75 and very satisfied (4) = 100] [20,21] . The translated individual scores for each item were averaged to become the mean satisfaction score for a given specific service variable. The average of all respondents' overall satisfaction scores constituted the OGNC mean overall satisfaction score. Mean (standard deviation, SD) described the satisfaction scores, as well as appointment intervals for initial and follow-up visits. Mean (SD) and range values delineated age. Frequency and percentage defined patient clinical outcome expectations, neurological diagnostic categories, comment categories, categorical demographic variables and non-response rates. Normality of distribution of overall satisfaction scores were explored with statistical and graphical methods. Comparisons of the mean overall satisfaction score between two patient groups (gender, residential addresses, service payment status, neurological diagnostic categories, clinical outcome expectations, time of clinic visit and age groups) were carried out by Mann–Whitney U-test with two-tail significance level (p-value) ≤ 0.05. Linear associations of the overall satisfaction score with age, appointment interval for follow-up visits and specific satisfaction scores were assessed by Spearman's rho correlation. An absolute Spearman's rho correlation coefficient (r) of > 0.10 and one-tail p ≤ 0.05 were taken to signify an important association and statistical significance, respectively. The priority indices among specific OGNC healthcare elements were determined based on the magnitude of their mean satisfaction scores and their degree of correlation (r magnitude) with the overall patient satisfaction score. In the case of a tie, greater weight was given to the healthcare element having higher correlation with the overall satisfaction. The priority indices reflected ranking of top service issues identified by the patients as important, but where the OGNC performance was relatively poor [20,21] . A simultaneous multiple linear regression model assessed the combined and independent effect of predictor variables on the overall patient satisfaction by looking at the regression coefficient (R 2 ) and T statistics values with a cut-off point for significance at ≤ 0.05. Independent variables demonstrating a relationship with the overall patient satisfaction score on bivariate analyses were considered potential predictors for a regression model. All statistical tests and models were run on SPSS for Windows version 15.0 software. 2.6 Ethical considerations The AAU-MF Institutional Review Board provided scientific and ethical approval (AAUMF 01-008) for the survey protocol on August 12, 2009. Each sampled patient provided verbal consent prior to the survey. Patients refusing, or suffering cognitive or language impairment or acute illness precluding valid consent, were considered non-responders and excluded. 3 Results The 233 patients in this survey were recruited during a continuous seven week period in 2009. Eleven (4.7%) patients declined to participate in the study for personal reasons. Exit interview and data abstraction were performed on the remaining 222 (95.3%) consenting patients. Ten (4.5%) patients submitted incomplete questionnaires and were classified as non-responders. Two-hundred twelve patients (9% total non-response rate) completed the survey analysis. Participants ranged 18–81 years of age with a mean of 40.1(SD = 15.3) years. The vast majority of patients (181, 96.8%) presented to the interview during a follow-up visit. Demographic and clinical characteristics are denoted in Table 1 . The mean overall patient satisfaction score was 70.4 (SD = 12.4). Mean patient satisfaction scores on overall service quality of guards (74.0), nurses (73.5) and doctors (71.8), and doctor's attention to patient's concerns (72.9) were higher than the mean overall satisfaction score ( Table 2 ). 69.7% of patients (154 of 208) were expecting improvement of their illness from the OGNC neurological healthcare services. The mean appointment intervals at the OGNC were 12.5 (SD = 13.3) and 107.2 (SD = 73.0) days for initial (N = 6) and follow-up (N = 181) visits respectively. Fig. 1 depicts the distribution of overall patient satisfaction scores. The Mann–Whitney U-test demonstrated statistically significant higher mean overall patient satisfaction scores among patients expecting improvement of their illness by visiting the OGNC (z = − 6.178, p = 0.001). There were no significant differences in mean overall patient satisfaction scores between other patient groups ( Table 1 ). The mean overall patient satisfaction was not significantly influenced by other major diagnostic categories such as cerebral palsy–paralytic syndromes (73.3 vs. 70.2, p = 0.782), nerve root–cord compression disorders (70.5 vs. 70.3, p = 0.735), or extrapyramidal–movement disorders (69.0 vs. 70.5, p = 0.496). Spearman's rho correlation statistics demonstrate that overall patient satisfaction scores were not significantly correlated with patient satisfaction scores on overall service of guards [r (207) = 0.109, p = 0.059], appointment interval for follow-up visits [r (181) = 0.012, p = 0.436] or age [r (191) = − 0.006, p = 0.469]. Each of the remaining specific patient satisfaction scores demonstrated a statistically significant (p < 0.05) and variable magnitude positive linear correlation (r > 0.10) with overall patient satisfaction ( Table 2 ). Patient satisfaction scores on the overall service of doctors provided the strongest correlation [r (211) = 0.416, p = 0.001]. In a simultaneous multiple regression model eight variables (satisfaction on waiting time at the clinic; waiting time for appointment visit; waiting area convenience; examination room convenience; overall service of nurses, doctor and card room; and clinical outcome expectations) continued to have significant [F (8,195) = 10.685, p = 0.001] contribution to 30.5% (R 2 = 0.305) of the variation in the observed mean overall patient satisfaction score. However, the independent contribution to the variation of overall patient satisfaction was statistically significant only for patient clinical outcome expectations (p = 0.001), and satisfaction on waiting area convenience (p = 0.028), overall service of doctor (p = 0.004) and the card room (p = 0.001) ( Table 3 ). The identified priority indices for the OGNC included, in a decreasing order of importance: waiting time at the clinic (mean = 57.2, r = 0.223), waiting area convenience (mean = 59.1, r = 0.221), examination room convenience (mean = 58.7, r = 0.125), waiting time for appointment visit (mean = 61.9, r = 0.167) and overall card room service (mean = 66.9, r = 0.166) ( Table 2 ). Two-hundred ten (99.1%) patients provided open-ended comments on different components of the OGNC healthcare quality. There were 52.4% (110) positive, 6.2% (13) neutral and 41.4% (87) negative comments concerning the OGNC. The majority of negative comments centered on the lack of provider continuity, especially involving the neurologist. 4 Discussion Our study demonstrated a mean overall patient satisfaction score of 70.4 on TASH OGNC healthcare in Addis Ababa, Ethiopia. Overall patient satisfaction was significantly related to patient clinical outcome expectations, waiting area convenience, and overall service quality of the doctors and card room. The top five priority indices for the OGNC are waiting time at the clinic, waiting area convenience, examination room convenience, waiting time for appointment visit, and overall service of the card room. A 2008 survey of 2.3 million patients throughout the United States (USA) reported mean overall patient satisfaction scores of 91 or higher for a host of non-neurological outpatient services [20] . The available literature, however, describes a lower albeit variable level of overall patient satisfaction scores (60%–70.6%) for a diverse selection of outpatient neurology services [12–14] . Two studies of outpatient specialty neurological healthcare facilities revealed average patient satisfaction scores of 4 (1–5 scale) and 5.3 (1–7 scale) on the overall quality of headache and Parkinson's disease care, respectively [15,16] . A third study involving a pediatric epilepsy clinic reported an overall care giver satisfaction exceeding 80% [17] . Our OGNC mean overall patient satisfaction score is comparable to these three Western specialty neurology clinic scores. Moreover, our result exceeds the average overall satisfaction scores reported in non-neurological outpatient services from Ethiopia and other African countries (54.1–66%) [4,23–25] . Patient satisfaction scores on the overall service of our OGNC were not significantly influenced by age, time of clinic visit or patient service payment status. These findings do not comport with other studies. First, a number of reports describe older patients being more satisfied than younger patients with outpatient healthcare quality [4,20,26,27] . This age-related difference was attributed to higher expectations with resultant unmet healthcare needs among younger compared to older patients [20] . The majority of our sample patients were young (nearly 70% under 50) which may obviate the scenario for age coming out as important component of overall satisfaction in our survey. Second, in developed countries, patients receiving outpatient healthcare earlier in the day were more satisfied with the service than patients treated near the end of the workday [20,21] . The gradual decline of staff hospitality and increasing impatience of clients was suggested for relatively lower satisfaction with service provided late in the day [20] . Third, a developing country study found patients on a social security scheme rated certain outpatient care features lower than other patients [28] . However, there are no consistent reports providing unequivocal evidence of a relationship between satisfaction and service payment status. Our study revealed a significant positive correlation of patient satisfaction on the following variables with overall patient satisfaction: waiting time for appointment visit, waiting time at the clinic, overall nurse service, and examination room convenience. However, none of these variables demonstrated an independent effect on multivariate analysis. These unique findings warrant further explanation. First, the modest effects of these variables were probably nullified by the strongest overall patient satisfaction predictor: overall service of the doctor. This possibility is supported by an American study concluding that time spent with the physician was a stronger predictor of patient satisfaction than time spent in the waiting room, and the decrement in satisfaction associated with long waiting times was substantially reduced by increased time with the physician [29] . Additionally, various studies demonstrate that physician quality of service is the most important factor for patient satisfaction [4,27,30,31] . Second, these variables may have interacting effects with waiting area convenience and overall service of doctors, and therefore been extinguished during multivariate analysis. Third, presumably low patient expectations towards these service variables may minimize their influence on overall satisfaction. Patient expectations have an independent predictor effect on patient satisfaction. Studies suggest that patients expecting a better outcome from treatment intervention have higher overall satisfaction scores on healthcare [4,5] . Our findings comport with these studies. The survey patients expecting clinical improvement related to their outpatient neurological healthcare had a significantly higher overall mean satisfaction score ( Table 1 ). Moreover, patient positive clinical outcome expectation was an independent predictor of overall patient satisfaction with the OGNC ( Table 3 ). The priority index identifies service areas in greatest need of improvement. The five top service improvement priorities for the OGNC relate to its accessibility (waiting time at the clinic and waiting time for appointment), structure (convenience of waiting area and examination room), and the registration process (overall card room service). A difficult or confusing registration process combined with long waiting time and inconvenient care surroundings sets the stage for unsatisfying patient encounter [20,21] . The commencement of subspecialty clinics for common local neurological disorders will reduce the OGNC patient load, and thereby improve waiting times while simplifying the registration process. Additionally, improving the OGNC infrastructure may create a more comfortable clinic environment, increasing overall satisfaction, and thus improving patient care. This is the first African study soliciting patient views and expectations on healthcare quality of a general neurology clinic. It provides timely, relevant data allowing policy modification and service development of outpatient neurological care in Ethiopia. The survey establishes priorities mandating in-depth evaluation and improvement of the OGNC waiting times, waiting area, examination room, and card room registration process. This survey provides a solid foundation for future patient satisfaction research. The questionnaire we developed may be utilized by primary care physicians or any other specialty group. Questionnaire items were generated based on previous literature and our understanding of patient service complaints. Moreover the cognitive pretesting of the items and their skewed distribution of scores are concurrent with the recommendations of content validity of satisfaction measures [32] . The lack of demographic effects on overall satisfaction contributed to low sources of error, and demonstrated that our questionnaire holds promise for assessment of overall satisfaction with the quality of care in diverse populations and settings. It provided a fairly comprehensive assessment of satisfaction components of the OGNC healthcare quality and relevant determinants of satisfaction regarding patient variables. However, it is impossible for this survey, or any other survey, to exhaust all potential dimensions of patient satisfaction within the quality-of-care framework [1,2,7,33,34] . 5 Conclusion Our survey demonstrated an encouraging level of mean overall patient satisfaction in the OGNC, significantly influenced by the service quality of the waiting area, card room and doctors, as well as patient clinical outcome expectations. There may be enhanced patient satisfaction with improvement of the OGNC structural amenities and expansion of the neurology service. Future studies should provide a more comprehensive assessment encompassing the broad range of potential satisfaction determinants and components constituting a solid healthcare quality assessment framework. This unique survey is the first report out of Africa on patient satisfaction with neurology services, and provides a guide for neurological or other specialty clinics seeking to improve and expand medical care. 6 Competing interest and funding The investigators covered the cost of the study. None of the authors have any relevant disclosures. Acknowledgement We thank the patients for cooperating and providing information in our preparation of this report. References [1] J. Sitzia N. Wood Patient satisfaction: a review of issues and concepts Soc Sci Med 45 12 1997 1829 1843 [2] A. Donabedian The quality of care: how can it be assessed? JAMA 260 12 1988 1743 1748 [3] B.G. Druss R.A. Rosenheck M. Stolar Patient satisfaction and administrative measures as indicators of the quality of mental health care Psychiatr Serv 50 1999 1053 1058 [4] M.S. Westaway P. Rheeder D.G. Van Zyl J.R. Seager Interpersonal and organizational dimensions of patient satisfaction: the modeling effects of health status Int J Qual Health Care 15 4 2003 337 344 [5] O. Baron-Epel M. Dushenat N. Freidman Evaluation of the customer model: relationship between patients' expectations, perceptions and satisfaction with care Int J Qual Health Care 13 4 2001 317 323 [6] M. Draper P. Cohen H. Buchan Seeking customer view: what uses are hospital satisfaction surveys? Int J Qual Health Care 13 6 2001 463 468 [7] J.W. Peabody M.M. Taguiwalo D.A. Robalino J. Frenk Improving the quality of care in developing countries Disease control priorities in developing countries 2nd ed 2006 Oxford University Press New York (NY) 1293 1308 [8] Institute of Medicine Crossing the quality chasm: a new health system for the 21st century. Committee report 2001 National Academy Press Washington DC [9] M. Traverso M. Salamano C. Botta M. Colautti P. Palchik B. Pérez Questionnaire to assess patient satisfaction with pharmaceutical care in Spanish language Int J Qual Health Care 2007 1 8 [10] H.B. Michael I.G.P. Wiadnyana H. Wihardjo I. Pohan Patient satisfaction in developing countries Soc Sci Med 48 1999 989 996 [11] Darby C, Valentine N, Murray C JL, Amala de Silva. Strategy on measuring responsiveness. GPE Discussion Paper Series. Geneva: World Health Organization (WHO); (EIP/GPE/FAR World Health Organization; vol 23). [12] R.E.D. Bautista E.T. Glen N.K. Shetty Factors associated with satisfaction with care among patients with epilepsy Epilepsy Behav 11 2007 518 524 [13] S.I. Bekkelund R. Salvesen Patient satisfaction with a neurological specialist consultation for headache Scand J Prim Health Care 20 2002 157 160 [14] H.J. Westervelt L.B. Brown G. Tremont D.J. Javorsky R.A. Stern Patient and family perceptions of the neuropsychological evaluation: how are we doing? Clin Neuropsychol 21 2007 263 273 [15] L. Ridsdale L.C. Clark A.J. Dowson L.H. Goldstein L. Jenkins P. McCrone How do patients referred to neurologists for headache differ from those managed in primary care? Br J Gen Pract 57 538 2007 May 1 388 395 [16] K.A. Grosset D.G. Grosset Patient-perceived involvement and satisfaction in Parkinson's Disease: effect on therapy decisions and quality of life Mov Disord 20 5 2005 616 619 [17] D.W. Webb H. Coleman A. Fielder C.R. Kennedy An audit of paediatric epilepsy care Arch Dis Child 79 1998 145 148 [18] J.D. Elashoff S. Lemeshow Sample size determination in epidemiologic studies W. Ahrens I. Pigeot Handbook of epidemiology 2005 Springer-Verlag Berlin Heidelberg Bremen 561 592 [19] P.P. Biemer L.E. Lyberg Introduction to survey quality 2003 John Wiley and Sons, Inc. New Jersey [20] Press Ganey Associates Outpatient pulse report: patient perspectives on American Health Care. Survey report. South Bend (IN) 2009 Press Ganey Associates, Inc [21] Press Ganey Associates Physician's office and outpatient pulse report. Survey report. South Bend (IN) 2007 Press Ganey Associates, Inc [22] WHO ICD-10 updates [online]. 2007 [cited 2009 Dec 4]. Available from: ULR:http://www.who.int/classifications/icd/icd10updates/en. [23] B. Abdosh The quality of hospital services in eastern Ethiopia: Patient's perspective Ethiop J Health Dev 20 3 2006 199 200 [24] G. Mitike A. Mekonnen M. Osman Satisfaction on outpatient services in hospitals of the Amhara Region Ethiop Med J 40 4 Oct 2002 387 396 [25] L. Oljira S. Gebre-Selassie Satisfaction with outpatient health services at Jimma Hospital, South West Ethiopia Ethiop J Health Dev 15 3 2001 179 184 [26] S. Atkinson D. Haran Individual and district scale determinants of users' satisfaction with primary health care in developing countries Soc Sci Med 60 2005 501 513 [27] J.L. Jackson J. Chamberlin K. Kroenke Predictors of patient satisfaction Soc Sci Med 52 2001 609 620 [28] T. Viroj S. Bennett S. Khongswatt A. Supacutikul A. Mills Patient satisfaction in Bangkok: the impact of hospital ownership and patient payment status Int J Qual Health Care 11 4 1999 309 317 [29] Anderson RT, Camacho FT, Balkrishnan R. Willing to wait?: The influence of patient wait time on satisfaction with primary care. BMC Health Services Research [serial online] 2007 [cited 2009 Dec 8]. Available from: URL:http://www.biomedcentral.com/1472–6963/7/31. [30] J.M. Aldana H. Piechulek A. Al-Sabir Client satisfaction and quality of health care in rural Bangladesh Bull World Health Organ 79 2001 512 517 [31] R.M. Fitzpatrick A. Hopkins Patients' satisfaction with communication in neurological outpatient clinics J Psychosomotic Res 25 5 1981 329 334 [32] J. Sitzia How valid and reliable are patient satisfaction data? An analysis of 195 studies Int J Qual Health Care 11 4 1999 319 328 [33] O.A. Arah G.P. Westert J. Hurst N.S. Klazinga A conceptual framework for the OECD health care quality indicators project Int J Qual Health Care 2006 Sep:5–13 [34] H.J. Sixma J.J. Kerssens C.V. Campen L. Peters Quality of care from the patients' perspective: from theoretical concept to a new measuring instrument Health Expect 1 1998 82 95