International Journal for Quality in Health Care 14:471-482 (2002)
© 2002 International Society for Quality in Health Care
Paper |
Reliability and validity of the Satisfaction with Hospital Care Questionnaire
1Department of Medical Psychology
2Department of Quality Improvement, Academic Medical Center/University of Amsterdam, The Netherlands
Objective. To establish the psychometric properties of the Satisfaction with Hospital Care Questionnaire (SHCQ) for measuring patient satisfaction and evaluations of hospital care quality.
Design and participants. Patients (n = 275) and staff members (n = 83) of four hospital wards completed the 57-item SHCQ addressing 13 aspects of care. Staff members completed the SHCQ from the patient’s perspective. The data were analyzed within the framework of generalizability theory.
Main outcome measures. Generalizability coefficients (GCs) and standard errors of measurement (SEs).
Results. GCs indicating differentiation among patients with different overall levels of satisfaction (SHCQ mean scores) were high (>0.90). GCs indicating differentiation among patients as to satisfaction with aspects of care (SHCQ scale scores) were generally satisfactory (>0.75) to high. Patients agreed well on overall level of hospital care quality (GCs >0.90) and differentiated reliably (GCs >0.80) among aspects of care. No differentiation among wards was found with respect to quality of care. Patients and staff agreed to a considerable extent (0.78) on ranking the SHCQ items on care quality, but staff ratings were lower. Reliability and validity of patients’ evaluations of quality of hospital care varied according to aspect of care.
Conclusions. The SHCQ reliably establishes both patient satisfaction and overall quality of hospital care. Whereas patients’ ratings may be too lenient, their ranking of the items on care quality appears to be valid, and is therefore suitable for monitoring and improving hospital care. Within scales, however, results should be interpreted more cautiously: for some items, patients cannot really tell the difference in quality of care.
Keywords: in-patients, questionnaire, reliability, satisfaction, validity
Research on patient satisfaction with medical care can be traced back to the late 1960s [1]. Over the past 30 years, an overwhelming number of publications on the topic has appeared [2,3]. At first, research focused on patient satisfaction as a condition to be satisfied in order to reach desirable clinical outcomes, such as appointment keeping or compliance with recommended treatment. Gradually, interest shifted to patient satisfaction as the dependent variable [4]. Patients’ views became an important tool in the processes of monitoring and improving quality of health care services [5]. Also, hospitals increasingly came to adopt a patient-centered attitude. Nowadays, hardly any hospital will fail to incorporate in-patient satisfaction ratings into their evaluations of care [6–11].
At the Academic Medical Center Amsterdam, The Netherlands, the Satisfaction with Hospital Care Questionnaire (SHCQ) is being used. The SHCQ can serve two measurement purposes: (1) measuring patient satisfaction, and (2) establishing hospital care quality. If patient satisfaction is the dependent variable of interest, like in studying the effects of doctors’ gender on patient satisfaction, patients are the subjects or objects of measurement. If hospital care quality is the dependent variable of interest, patients are judges rather than the objects of measurement. The difference is conceptual rather than having an impact on item wording or instructions. The goal of our study was to determine the reliability of the SHCQ for both measurement purposes, as a supplement to previous findings concerning internal consistency and test–retest reliability [12].
In addition, we examined the validity of the SHCQ for measuring quality of hospital care. As yet, relatively little is known about the validity of patients’ evaluations of hospital care [13]. Rubin [14] concludes from a literature review that three strategies have been used to assess the construct validity of patients’ ratings. One strategy is to determine the effect of experimentally manipulated features of care on patients’ satisfaction ratings [15,16]. A second strategy is to establish the amount of agreement between patient ratings and ratings from other sources [14]. A third strategy is to establish the relationship between satisfaction and theoretically related variables such as improved health [17], and willingness to return to the same hospital [18] or recommend the hospital [19]. Of these strategies, validating patients’ ratings to ratings from other sources is the one most rarely applied, most likely because eligible others for this purpose are not easily found.
To establish validity, one may think of a patient’s partner, other family members, or close friends who regularly came to visit the patient in hospital. But unless they rate a visible aspect of care like hygiene, these raters will have to rely largely on second hand information. Moreover, they probably get their information from the patient, implying that their ratings will be contaminated. As a results of this, we turned to hospital staff (nurses, doctors, and other disciplines). Because staff members’ and patients’ views of care quality are not necessarily the same [6,20–22], we asked staff members to rate hospital care quality from the patient’s perspective. Naturally, staff members also base their ratings inevitably on second hand information to some extent. However, information is then aggregated across a large number of patients and across a larger time span, reducing contamination with individual patient ratings.
To evaluate the psychometric properties of the SHCQ, five research questions were formulated:
- If the SHCQ is used to measure patient satisfaction, how reliably does the instrument differentiate among patients with different levels of satisfaction? In other words, if patient A is more satisfied than patient B on item 1, is he or she also more satisfied than patient B on other items (consistency in item responses)?
- If the SHCQ is used to evaluate hospital care quality, to what extent do patients agree and what would be the minimum number of patients needed to establish care quality with sufficient measurement precision?
- If the SHCQ is used to evaluate hospital care quality, do patients differentiate reliably between aspects of care, for instance, between medical care and ‘hotel’ aspects of care? If all aspects of care are rated to be at the same level of care quality, one would suspect a response bias, i.e. a halo-effect, in patient ratings.
- If the SHCQ is used to evaluate hospital care quality, can hospital wards be differentiated on care quality based on patients’ answers to the SHCQ?
- Do patients’ ratings on the SHCQ yield valid evaluations of hospital care quality?
Questions 1–4 address reliability: whether systematic variance is observed in patients’ ratings on the SHCQ that can be attributed to identifyable sources of variance: patients (question 1), items (question 2), scales of items measuring aspects of care (question 3), or wards (question 4). Question 5 concerns construct validity, i.e. agreement in ratings between patients and staff members served as the primary indicator. Patients and staff can be conceptualized as different measurement instruments for measuring quality of care from the patient’s point of view. In addition, reliability and validity are to some extent intricately related: if patients reliably discriminate between aspects of care, this finding also indicates construct validity.
All these questions can be answered by applying generalizability theory (GT), a method of data analysis which is explained further in Methods. In GT, two concepts play an important role: relative and absolute decisions about the objects of measurement (e.g. patients, items, wards). In relative decisions, the objects of measurement in a sample are compared with each other: their positions on a measurement scale (e.g. patient satisfaction, quality of care) are established and evaluated taking the sample’s mean score as the point of reference. In absolute decisions, each object’s position on a measurement scale is established and evaluated taking the scale’s natural zero-point as the reference, i.e. independently of the scores of the other objects in the same sample. Whereas in relative decisions only the object’s rank is of interest, in absolute decisions it concerns its precise score.
Methods
Participants, instrument, and procedure
Participants were 275 recently discharged patients and 83 staff members from four hospital wards of the Academic Medical Center (AMC), Amsterdam, The Netherlands. These four wards composed a stratified random selection out of a total of 18 eligible hospital wards divided across four divisions (Internal, Surgical, Neurological, and Obstetric Medicine). At each ward, 
100 consecutively discharged patients were invited to participate. Patients <18 years of age and patients who were judged upon discharge to be too ill to participate were excluded. Response rates varied from 59% to 70% per ward, with an overall response rate of 63% (n = 275). Respondents were 65% female and aged 18–93 years [mean 54 years, standard deviation (SD) 18 years]. Their level of education was slightly below the national average. Respondents stayed in hospital for an average of 12 days (median 7 days).
Patient satisfaction with hospital care was assessed by means of the SHCQ, a 57-item questionnaire addressing 13 aspects of care (two to seven items per aspect): outpatients’ clinic, admission procedures, nursing care, medical care, other disciplines (e.g. licensed social worker), information, patient autonomy, emotional support, ‘hotel’ aspects of care (e.g. meals), recreation facilities, miscellaneous aspects (e.g. rules and regulations), ease of access to the hospital, and discharge and aftercare. The SHCQ is based on a questionnaire developed by Visser [23,24], which was partially adapted and extended to meet the wishes of the AMC. The SHCQ has demonstrated good internal consistency and test–retest reliability [12,25]. Ratings were made on a 10-point scale, running from 1 (very poor) to 10 (excellent). Dutch patients are familiar with this scale, being the national approach to evaluate a student’s school performance.
Patients were handed the SHCQ by a nurse upon discharge from hospital. Completed questionnaires were to be returned to the researchers in a pre-paid return envelope. The accompanying letter emphasized that responses would be treated confidentially. If necessary, a reminder was sent after 2–3 weeks.
Staff members were asked to complete the SHCQ shortly after recruitment of patients on their wards had been finished and before patients’ ratings had been reported to them. The instruction was to indicate per item on the 10-point response scale how patients having been hospitalized on their ward over the past 3 months would, on average, evaluate the care described by that item. Seventeen to 29 staff members per ward (n = 83) completed the SHCQ, the majority of whom were nurses (71%).
Analysis
The data were analyzed within the framework of GT developed by Cronbach and co-workers [26]. Whereas classical test theory takes just one source of (error) variance into account, GT enables one to disentangle the relative contributions of different sources of (error) variance to total variance in item responses, generally called ‘facets in the measurement design’ [26–30]. Figure 1 shows the sources of variance that we took into account in our study of the reliability (a) and validity (b) of the SHCQ.
|
For each source, its contribution to total variance in the item responses can be estimated from the data (estimates of variance components; SPSS 9.0 for Windows, General Linear Model). These estimates of variance components are used to calculate generalizability coefficients (GCs) and standard errors (SEs) of measurement, as measures of homogeneity (reliability) and measurement precision, respectively. Based on these estimates of variance components, different research questions can be answered by calculating different GCs and SEs, using the formulae given by Cardinet et al. [27,28]. The formulae used in our study can be obtained from the first or second author. Table 1 gives an overview of our research questions and the interpretation of the corresponding GCs.
|
GCs range from zero to unity (perfect), and can be interpreted in several ways. Firstly, the GC indicates how well an observed score on a sample of measurements can be generalized to the universe score: the score when all possible measurements could be taken into account (e.g. patient’s score on an infinite number of satisfaction items rather than a representative selection of them). Secondly, the GC is a measure of how well one can differentiate between the levels of the facet of differentiation (the object of study, e.g. patients). Thirdly, the GC is a measure of the homogeneity of the facet(s) of instrumentation or generalization (any source of variance that affects the measures taken of the objects under study). These three interpretations are equivalent and independent of which facets are considered differentiation facets, and which facets are considered instrumentation facets [27,28].
Variance components of instrumentation facets contribute to error variance. If one is interested in the reliability of relative decisions (e.g. whether patient A is more satisfied than patient B), only variance components representing interactions of the instrumentation facets with the differentiation facet(s) contribute to error variance (see Table 1). If one is interested in the reliability of absolute decisions (e.g. whether patient A scores 6.0 rather than 7.0), all variance components associated with the instrumentation facet(s) contribute to error variance (e.g. 28). SEs are calculated by taking the root of the error variance.
To answer questions 1–4, we calculated (restricted maximum likelihood) estimates of variance components for the facets patients (P), items (I), scales (S), and wards (W), and their interactions (PI, PS, WI, and WS, respectively), as illustrated in Figure 1a. Note that P and I denote P:W and I:S, respectively, as patients are nested within wards and items are nested within scales. These variance components were estimated for the 275 patients.
To determine how reliably the SHCQ measures patient satisfaction (question 1), GCs for relative and absolute decisions were calculated taking patients as the facet of differentiation, and the SHCQ items and scales as the facets of instrumentation. These GCs express how well we can differentiate among patients with different satisfaction levels, and at the same time they express the homogeneity of the items and scales. GCs were calculated for the SHCQ mean score (overall level of patient satisfaction) and for scale scores (satisfaction with aspects of care), taking only the items as the instrumentation facet. The GC for relative decisions corresponds with the well known Cronbach’s alpha. Likewise, the SE for relative decisions corresponds to the standard error of measurement in classical test theory.
To determine how well patients agree on relative and absolute level of quality of care (question 2), we calculated GCs taking items and scales as the facets of differentiation, and patients as the facet of instrumentation. The GC for relative decisions corresponds to the coefficient of inter-rater reliability, and the GC for absolute decisions to the coefficient of inter-rater agreement. These were calculated for the SHCQ mean score (overall quality of care) and for scale scores (aspects’ quality of care).
With the same facets of differentiation and instrumentation we also estimated the minimum number of patients that is needed to achieve sufficient measurement precision. We choose SE 
0.25 as the criterion. With an SE of 0.25, the 95% confidence interval for a score indicating a certain level of care quality does not include values below or above that level. For instance, (6.0 ± 1.96) x 0.25 does not include values <5.5 or >6.5, which would be the cut-off points, respectively, for rounding to the integer value 6.0, meaning ‘sufficient’ (just acceptable, just passing grades) on the 10-point response scale.
Because the numbers of items per scale (two to seven) are relatively low, we chose SE 0.30 as the criterion for measurement precision at the level of scales (e.g. Medical Care). This means that we can be 90% confident that a scale score indicating a certain level of care quality does not include values below or above that level.
To assess whether patients differentiate reliably among aspects of care (question 3), GCs for relative and absolute decisions were calculated with scales as the facet of differentiation and patients as the facet of instrumentation, indicating the amount of differentiation among scales. These GCs can also be interpreted as coefficients of construct validity, as they indicate whether patients can or cannot really tell the difference in quality of care between different aspects of care.
By taking wards as the facet of differentiation, and items and scales as the facets of instrumentation, GCs for relative and absolute decisions were calculated that indicate how reliably the SHCQ differentiates between quality of hospital care of different wards (question 4).
To answer question 5, we calculated (restricted maximum likelihood) estimates of variance components for raters (R), type of rater (T, patients versus staff), items (I), and wards (W), and their interactions (RI, TI, TW, WI, and TIW), as illustrated in Figure 1b. Raters are nested within wards and type of rater. Including scales as a source of variance in the item responses did not alter the results appreciably; thus, this facet has been left out for reasons of simplicity. To determine how well patients and staff members agreed, we took items and wards as the facets of differentiation, and type of rater as the facet of instrumentation. GCs for relative and absolute decisions were calculated, indicating homogeneity of ratings. Like the GCs answering question 2, these GCs correspond to the coefficients of inter-rater reliability (relative level of care quality) and inter-rater agreement (absolute level of care quality), despite the fact that ‘rater’ now denotes type of rater (patients versus staff). These GCs were calculated for the SHCQ in total (overall quality of care) and separately for each scale (aspects’ quality of care).
Results
Estimates of variance components
Table 2 shows the overall (57 items) estimates of variance components and the percentages of variance in the item responses that the associated sources of variance accounted for. Likewise, estimates of variance components were calculated for separate scales (not shown). Concerning the assessment of reliability (questions 1–4), overall estimates were based on 11 635 out of 15 675 possible observations (275 patients x 57 items). The high percentage of missing observations (26%) is due to the fact that not all SHCQ items apply to all patients. As Table 2 shows, patients were the largest source of variance in the item responses (33%, and 43% for the patients x items interaction), whereas the contribution of wards was 0.
|
Concerning the assessment of validity (question 5), overall estimates were based on 16 191 out of 20 406 possible observations [(275 patients + 83 staff members) x 57 items]. Here, raters (22%), type of rater (25%), and the raters x items interaction (39%) contributed most to the variance in item responses.
Reliability of the SHCQ for measuring patient satisfaction
Both the overall GC for relative decisions and the one for absolute decisions were estimated at 0.96, indicating that the SHCQ mean score differentiates reliably among patients with different overall levels of satisfaction (Table 3). Measurement precision was also high, as is indicated by the low SEs of 0.15 and 0.17 for relative and absolute decisions, respectively. Hence, we can be 95% sure that the true overall level of satisfaction of a patient with an SHCQ mean score (averaged across all items) of, for example, 7.0 lies between 6.66 and 7.34.
|
The GCs for scale scores indicating differentiation among patients averaged 0.83 (range 0.64–0.93) for relative decisions and 0.82 (range 0.58–0.92) for absolute decisions (Table 3). Differentiation was excellent (>0.90) for ‘nursing care’, ‘medical care’, ‘emotional support’, and ‘discharge and aftercare’. Good differentiation (>0.80) was found for ‘other disciplines’, ‘information’, ‘patient autonomy’, and ‘ease of access to the hospital’. For the remaining scales, differentiation among patients was satisfactory (0.76–0.80), except for ‘recreation facilities’ (0.64).
SEs (Table 3) averaged 0.40 (range 0.30–0.52) for relative decisions and 0.42 (range 0.28–0.59) for absolute decisions, indicating insufficient measurement precision for most scales. In the ‘goal’ columns of Table 3 it can be seen that the minimum number of items to meet the requirement of sufficient measurement precision for ‘Outpatients’ Clinic’, for instance, is 15 items for relative decisions and 16 items for absolute decisions on care quality. In most cases, the number of items should be doubled or tripled to be 90% certain that a patient’s true scale score is not above or below his or her observed scale score. Positive exceptions, i.e. high reliability with sufficient measurement precision, were the five-item scales ‘nursing care’ and ‘medical care’.
Reliability of the SHCQ for measuring quality of care
The GCs indicating overall inter-rater reliability and inter-rater agreement (question 2) were both very high, >0.90 (Table 4). The number of patients to meet the requirement of sufficient measurement precision is nine patients for relative decisions and 19 patients for absolute decisions on care quality.
|
At the level of scales (Table 4), inter-rater reliability was excellent to good (>0.80) for ‘Outpatients’ Clinic’, ‘admission procedures’, ‘medical care’, ‘other disciplines’, information’, ‘hotel aspects of care (hotel care)’, ‘recreation facilities’, and ‘miscellaneous aspects’. For most of these scales even inter-rater agreement was excellent to good. Due to zero estimates of item variance components, reliability coefficients were zero for ‘patient autonomy’, ‘emotional support’, and ‘discharge and aftercare’, but measurement precision appeared high, as with all other scales.
The GCs indicating how well patients discriminate among different aspects of care (question 3) were good: 0.85 for relative decisions and 0.81 for absolute decisions. Their associated SEs were low (0.08 and 0.09, respectively), indicating excellent measurement precision.
It appeared impossible to differentiate among hospital wards with respect to overall care quality (question 4). The variance component for wards was estimated at zero (Table 1) and, therefore, the GCs for relative and absolute decisions were zero (Table 5). Still, measurement precision was high. For instance, only 11 patients are needed to obtain precise estimates of the absolute level of overall quality of care of hospital wards.
|
With respect to scales (Table 5), ‘medical care’, ‘other disciplines’, and ‘hotel care’ showed reliable differentiation among hospital wards with regard to ranking them on care quality. Only ‘medical care’ also differentiated well with respect to a ward’s absolute level of care quality. All scales showed low SEs, indicating high measurement precision.
Validity of the SHCQ for measuring quality of care
As Table 6 shows, overall inter-rater reliability was satisfactory (0.78), but overall inter-rater agreement low (0.37). For separate scales, inter-rater reliability varied from low to excellent (0.44–0.95), whereas inter-rater agreement was low for all scales (0.06–0.50). Inter-rater reliability was excellent to good (>0.80) for ‘medical care’, ‘hotel care’, ‘recreation facilities’, and ‘admission procedures’. Inter-rater reliability was sufficient for ‘Outpatients’ Clinic’ (0.76), and near sufficient for ‘nursing care’ (0.70) and ‘emotional support’ (0.71). For the remaining scales, even inter-rater reliability was low, indicating that patients and staff taking the patient’s perspective not only disagreed on absolute level but also on ranking the items within these scales on quality of care.
|
Item means based on patients’ and staff ratings, respectively, showed that staff ratings were systematically lower, except for ‘your encounters with the physiotherapist(s)’ (Table 7). Patients and staff members differed significantly in their ratings (grand mean = 7.4 versus 6.4, SD = 0.36 versus 0.68, t(56) = 16.6, P < 0.001, paired-samples t-test).
|
Discussion
Reliability of the SHCQ for measuring patient satisfaction
The SHCQ differentiates reliably and with high measurement precision among patients with different overall levels of satisfaction with hospital care. Differentiation at a more specific level of aspects of care (scales) was found to be reliable, but measurement precision was generally insufficient. In other words, we can be more sure for the SHCQ mean score than for scale scores that a patient’s ‘true’ level of satisfaction is in accordance with the observed level. Only a patient’s level of satisfaction with medical care and nursing care can be established with sufficient reliability and measurement precision. Measurement precision is of less importance, however, for most research purposes, which involve group comparisons, as observed scores and true scores are related in a linear manner.
Reliability of the SHCQ for evaluations of hospital care quality
Patients highly agreed on both the relative and absolute levels of care quality of the SHCQ items. Only few patients (nine for decisions on the relative level, 19 for decisions on the absolute level) are needed to establish overall quality of hospital care with sufficient measurement precision. Thus, the SHCQ would be a very efficient instrument to compare hospitals on overall quality of care.
For separate scales, inter-rater reliability and agreement were generally either high or zero. Lack of agreement among patients is due to a lack of variance in ratings among the items of a scale. For these items, apparently patients cannot really see differences in care quality. This observation applies to ‘patient autonomy’, ‘emotional support’, and ‘discharge and aftercare’. We might thus consider replacing one or more of the items in these scales. Leaving out items would mean less concrete information for ward staff. Also, a smaller number of items might threaten content validity when measuring patient satisfaction.
We found differentiation among scales to be good, with respect to both relative and absolute levels of care quality. In other words, patients can tell the difference in quality of care between different aspects of care. Measurement precision was high: less than a handful of patients are needed to establish which aspects are relatively high or low in care quality. These results not only suggest that patients’ evaluations of hospital care quality are valid, but also that hospitals do not need to perform large-scale patient satisfaction studies to get an indication of which aspects of care are most in need of quality improvement.
We found no differentiation among wards using all items of the SHCQ (overall quality of care). Because the variance component for wards was zero we cannot tell whether differences in overall quality of care among wards can be reliably established with the SHCQ. Some scales offered reliable differentiation, but almost exclusively with respect to ranking wards on care quality. The reason for the generally disappointing low GCs may be that many SHCQ scales concern hospital qualities rather than ward qualities. We expect that these GCs will be higher if the SHCQ is used to measure differences among wards of different hospitals.
Validity of the SHCQ for evaluations of hospital care quality
Patients and staff agreed to a considerable extent on ranking the SHCQ items on care quality (0.78). However, they disagreed on absolute level (0.37): like in other studies [6,20–22], staff ratings were systematically lower. We cannot infer whether patients are too lenient or staff too critical. Given that patients discriminate reliably among aspects of care, patients’ ratings seem valid, at least for the purpose of monitoring care quality and identifying those aspects of care that may need quality improvement.
Conclusions
The SHCQ is very reliable and shows high measurement precision for both measuring overall level of patient satisfaction and establishing overall level of hospital care quality. Furthermore, patients’ ratings on the SHCQ yield valid information for the hospital as to which items and aspects are relatively high or low in quality of care. Whether the SHCQ is informative at the level of scales, referring to aspects of care, depends on the purpose of measurement: whether one is interested in establishing patient satisfaction or hospital care quality. For measuring patient satisfaction, all scales except ‘recreation facilities’ appeared sufficiently to highly reliable. For evaluations of hospital care quality, reliability and validity varied by scale. Within some scales, even patients’ ranking of the items should be interpreted cautiously. Still, patients’ ratings may yield a valuable point of departure for hospital care quality improvement programs.
The authors gratefully acknowledge the comments of Willem K. B. Hofstee, Hanneke (J) C. J. M. De Haes, and Mirjam A. Sprangers on earlier versions of this paper. This research was funded by the Board of Directors of the Academic Medical Center/University of Amsterdam, The Netherlands.
Address reprint requests to A. A. J. Hendriks, Department of Medical Psychology (J4-401), Academic Medical Center/University of Amsterdam, PO Box 22660, 1100 DD Amsterdam, The Netherlands. E-mail: jolijn.hendriks{at}planet.nl 
Accepted for publication July 30, 2002.
References
- Hall JA, Dornan MC. Meta-analysis of satisfaction with medical care: description of research domain and analysis of overall satisfaction levels. Soc Sci Med 1988; 27: 637–644.
- Lewis JR. Patient views on quality care in general practice: literature review. Soc Sci Med 1994; 39: 655–670.
- Pascoe GC. Patient satisfaction in primary health care: a literature review and analysis. Eval Program Plann 1983; 6: 185–210.[Medline]
- Williams B. Patient satisfaction: a valid concept? Soc Sci Med 1994; 38: 509–516.
- Sitzia J, Wood N. Patient satisfaction: a review of issues and concepts. Soc Sci Med 1997; 45: 1829–1843.
- Arnetz JE, Arnetz BB. The development and application of a patient satisfaction measurement system for hospital-wide quality improvement. Int J Qual Health Care 1996; 8: 555–566.
- Carr-Hill RA. The measurement of patient satisfaction. J Public Health Med 1992; 14: 236–249.
[Abstract/Free Full Text] - Davis SL, Adams-Greenly M. Integrating patient satisfaction with a quality improvement program. JONA 1994; 24: 28–31.
- Drachman DA. Benchmarking patient satisfaction at academic health centers. J Qual Improv 1996; 22: 359–367.
- Hays RD, Larson C, Nelson EC, Batalden PB. Hospital Quality Trends: A short-form patient-based measure. Med Care 1991; 29: 661–668.[Web of Science][Medline]
- Meredith P. Patient satisfaction with communication in general surgery: Problems of measurement and improvement. Soc Sci Med 1993; 37: 591–602.
- Hendriks AAJ, Vrielink MR, Van Es SQ, Smets EMA. Patient Satisfaction Research: Final Report (in Dutch). Amsterdam: Academic Medical Center/University of Amsterdam, Department of Quality Improvement (J1-153), 2000.
- Sitzia J. How valid and reliable are patient satisfaction data? An analysis of 195 studies. Int J Qual Health Care 1999; 11: 319–328.
[Abstract/Free Full Text] - Rubin HR. Patients’ evaluations of hospital care: a review of the literature. Med Care 1990; 28: S3–S9.[Web of Science][Medline]
- Hinshaw AS, Gerber RM, Atwood JR, Allen JR. The use of predictive modeling to test nursing practice outcomes. Nurs Res 1983; 32: 35–42.[Web of Science][Medline]
- Ley P, Bradshaw PW, Kincey JA, Atherton ST. Increasing patients’ satisfaction with communications. Br J Soc Clin Psychol 1976; 15: 403–413.[Web of Science][Medline]
- Carmel S. Satisfaction with hospitalization: a comparative analysis of three types of services. Soc Sci Med 1985; 21: 1243–1249.
- Doering ER. Factors influencing inpatient satisfaction with care. Qual Rev Bull 1983; October: 291–299.
- Abramowitz S, Cote AA, Berry E. Analyzing patient satisfaction: a multianalytic approach. Qual Rev Bull 1987; April: 122–130.
- Kurata JH, Nogawa AN, Phillips DM, Hoffman S, Werblun MN. Patient and provider satisfaction with medical care. J Fam Pract 1992; 35: 176–179.[Web of Science][Medline]
- Merkel WT. Physician perception of patient satisfaction: do doctors know which patients are satisfied? Med Care 1984; 22: 453–459.[Web of Science][Medline]
- Rashid A, Forman W, Jagger C, Mann R. Consultations in general practice: a comparison of patients’ and doctors’ satisfaction. Br Med J 1989; 299: 1015–1016.
- Visser AP. The Experience of Staying in Hospital (in Dutch). Assen: Van Gorcum, 1984.
- Visser AP, ed. Research on Hospital Inpatients’ Satisfaction with Care (in Dutch). Lochum: De Tijdstroom, 1988.
- Hendriks AAJ, Vrielink MR, Smets EMA, Van Es SQ, De Haes JCJM. Improving the assessment of (in)patients’ satisfaction with hospital care. Med Care 2001; 39: 270–283.[Web of Science][Medline]
- Cronbach LJ, Gleser GC, Nanda H, Rajaratnam N. The Dependability of Behavioral Measurements: Theory of Generalizability for Scores and Profiles. New York: Wiley, 1972.
- Cardinet J, Tourneur Y, Allal L. The symmetry of generalizability theory: applications to educational measurement. J Educ Meas 1976; 13: 119–135.
- Cardinet J, Tourneur Y, Allal L. Extension of generalizability theory and its applications in educational measurement. J Educ Meas 1981; 18: 183–204.
- Shavelson RJ, Webb NM. Generalizability Theory: a Primer. Newbury Park: Sage, 1991.
- Shavelson RJ, Webb NM, Rowley GL. Generalizability theory. Am Psychol 1989; 44: 922–932.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Antoniotti, K. Baumstarck-Barrau, M.-C. Simeoni, C. Sapin, J. Labarere, L. Gerbaud, L. Boyer, C. Colin, P. Francois, and P. Auquier Validation of a French hospitalized patients' satisfaction questionnaire: the QSH-45 Int. J. Qual. Health Care, August 1, 2009; 21(4): 243 - 252. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Edward, L. C. Lemaire, B. Preckel, F. J. Oort, M. J. L. Bucx, M. W. Hollmann, and J. C. J. M. de Haes Patient Experiences with the Preoperative Assessment Clinic (PEPAC): validation of an instrument to measure patient experiences Br. J. Anaesth., November 1, 2007; 99(5): 666 - 672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Oltedal, A. Garratt, O. Bjertnaes, M. Bjornsdottir, M. Freil, and M. Sachs The NORPEQ patient experiences questionnaire: Data quality, internal consistency and validity following a Norwegian inpatient survey Scand J Public Health, October 1, 2007; 35(5): 540 - 547. [Abstract] [PDF]
|
|
|
|
|
|
A. A. J. Hendriks, E. M. A. Smets, M. R. Vrielink, S. Q. Van Es, and J. C. J. M. De Haes Is personality a determinant of patient satisfaction with hospital care? Int. J. Qual. Health Care, April 1, 2006; 18(2): 152 - 158. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||