Dawn AG, McGwin G, Lee PP. Patient Expectations Regarding Eye CareDevelopment and Results of the Eye Care Expectations Survey (ECES). Arch Ophthalmol. 2005;123(4):534-541. doi:10.1001/archopht.123.4.534
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
To develop an instrument (the Eye Care Expectations Survey [ECES]) that can identify and quantify the expectations of patients visiting eye care providers.
A total of 202 patients attending 4 ophthalmology practices affiliated with Duke University Eye Center enrolled in the study. All participants completed the pilot version of the ECES, which was developed through a review of the expectations literature and a multicondition focus group process. Factor analysis of patients’ responses was used to identify the performance characteristics of the ECES.
Factor analysis yielded factors that describe 4 distinct types of expectations: patient involvement in eye care, interpersonal manner, information about diagnosis and prognosis, and communication and clinical competence. These 4 factors explained 89% of the total variance. The factor analysis identified a subset of 27 questions (of the original 37) to represent the 4 factors. Estimates of internal consistency and test-retest reproducibility indicate that the ECES is reliable. In addition, tests of association with clinical variables support the construct validity of the survey. The psychometric properties of the ECES were influenced by the severity of the underlying eye condition as well as other demographic and clinical variables, including the patient’s rating of his or her own vision, best corrected visual acuity, sex, education, race, and household income.
Based on this cross-sectional study, the 27-item ECES appears to be a useful tool for better understanding patients’ expectations regarding eye care.
In recent years, patients’ perspectives have played an increasingly important role in health care.1 Patient-centered care is health care that is responsive to patients’ wants, needs, and preferences.2 The patient-centered care movement can be linked to related trends in medicine over the past decade. The shift toward continuous quality improvement, which gained momentum in the 1990s, places meeting patient expectations at the core of medicine’s mission.3 Furthermore, the growing integrative medicine movement insists on patients being active participants in their health care.4 Moreover, the rise of consumerism and dramatic increases in patients’ level of education have contributed to greater patient demand for information and involvement and rising expectations.5
There has been a growing body of literature on patient expectations over the past two decades. Previous research suggests that most patients have explicit expectations when they visit their physicians.6,7 Despite the importance of patients’ expectations, however, no standardized assessment instrument currently exists for measuring patients’ expectations.6,8 Disagreements over the most appropriate methods for measuring patient expectations have been a barrier to more refined understanding.9
Moreover, the vast majority of expectations research has been conducted in primary care settings. Little is known about patient expectations of ophthalmologists and eye care. The ophthalmology literature that does exist has focused primarily on expectations regarding surgical outcomes, such as patients’ expectations for cataract surgery.10 However, recent developments should make patient expectations increasingly important in ophthalmology. First, the dramatic rise in the number of refractive surgical procedures performed in the United States shows that increasing numbers of patients have been drawn to the field of ophthalmology for elective procedures for non–sight-threatening conditions.11 Second, increased competition in the eye care market has led to a greater appreciation for the need to understand patient desires.12,13
Motivated by the pivotal role of patient expectations in quality-of-care assessments and by the limited knowledge about patients’ expectations regarding eye care, we undertook a study to develop an instrument (the Eye Care Expectations Survey [ECES]) that can identify and quantify the expectations of patients visiting eye care providers. In addition, we explored which clinical and demographic characteristics are associated with patients’ expectations regarding eye care.
We first conducted a review of the literature on patient expectations regarding medical and surgical care between 1966 and 2002. The initial MEDLINE search terms were the title words expectations or desire or requests and the medical subject headings consumer satisfaction or patient satisfaction or physician-patient relations. We also examined review article reference lists for potentially relevant studies. We then reviewed and analyzed the existing literature.14
In general, value expectations, which refer to patients’ desires, hopes, or wishes concerning clinical events, are the dominant model throughout the expectations literature, which mostly involves studies conducted in primary care settings.15,16 The 10 most commonly addressed areas of patient expectations and requests are medical information, medication/prescription, counseling/psychosocial support, diagnostic testing, referral, physical examination, health advice, outcome of surgery or treatment, therapeutic listening, and waiting time.
In the initial qualitative phase of the study, we conducted focus groups with ophthalmology patients at Duke University Eye Center, Durham, NC, to help indicate the range of issues that might be explored in the subsequent quantitative research phase.17 We used the major expectation areas identified in the literature review as well as initial patient interviews as starting points to create the script for the focus groups. Following approval of the study by the Duke University Health System institutional review board, a total of 6 focus groups were conducted. A total of 38 patients participated in the focus groups, and groups ranged in size from 4 to 10 people. A variety of diagnoses were represented in the groups. At the end of each group, participants were presented with a copy of a patient concerns form18 and asked to identify items that they thought were important when they visited their ophthalmologist.
Two of the authors (A.G.D. and P.P.L.) reviewed the transcripts of the 6 focus groups and analyzed them for content and key concepts. Content analysis yielded 22 areas of expectations for eye care expressed by focus group participants.17 These expectation areas were based on consistent patterns of responses obtained from focus group participants and were based on findings that had the strongest, most broadly based support from participants in our groups as well as unique areas, even if mentioned by only one person.
The content of the pilot ECES was empirically derived from the multicondition focus group process described above. We constructed the ECES from the expectation areas identified in the focus groups and from items identified as important on the patient concerns forms.18 The original ECES instrument included the following components: 16 items to rate expectations for ongoing care on a 5-point scale ranging from not important to extremely important; 21 items to rate expectations for the specific visit on a 5-point scale ranging from not important to extremely important; and demographic questions regarding sex, race, age, education, occupational status, type of insurance, income, whether the individual had visited an eye doctor previously, and items adapted from the Visual Function Questionnaire19 rating overall health on a 5-point scale and rating vision on a 6-point scale.
To assess the clarity of the questions and instructions and the time required to complete the survey, we pilot-tested the ECES among patients at Duke University Eye Center and made appropriate changes to maximize response rates and patients’ comprehension of the survey.
A total of 202 patients with eye appointments who met the eligibility criteria were recruited from 4 ophthalmology practices affiliated with Duke University Eye Center between August and November 2002. Investigators at each practice site approached all potentially eligible subjects. Patients had to meet the following criteria: at least 18 years old; speak English (or be accompanied by an English-speaking individual willing to translate); have an eye appointment scheduled for that day; be able to complete a written questionnaire with minimal assistance; be willing to provide written informed consent; achieve a minimum score of 24 out of 30 on the Folstein Mini-Mental State Examination.
All subjects were asked to complete the ECES. The study protocol was approved by the Duke University Health System institutional review board, and all participants gave written informed consent prior to enrollment. To avoid a possible bias from information learned during the eye appointment, patients completed the ECES prior to seeing the ophthalmologist. Ophthalmologists were aware that the study was being conducted, but patients’ responses were not shared with physicians.
In addition, we collected clinical information from patients’ medical records, including their primary ophthalmic diagnoses and their best corrected visual acuity in the better and worse eyes. Based on the patients’ diagnoses, we classified them as having either a potentially irreversible blinding eye condition or a nonblinding eye condition. Blinding eye conditions included diagnoses such as glaucoma, age-related macular degeneration, and diabetic retinopathy, among others. Nonblinding eye conditions included patients with well eyes, refractive errors, and cataract, among others.
To assess the test-retest reliability of the ECES items, we had a group of 20 subjects complete the instrument on two occasions approximately 2 weeks apart. Due to the nature of the ECES questions, test-retest analysis was appropriate only for the 16 items used to rate expectations for ongoing eye care. The ECES questions regarding visit-specific expectations would be expected to change from visit to visit; thus, test-retest analysis was not performed for the 21 items used to rate expectations for the specific visit. We used the intraclass correlation coefficient (ICC) to determine the stability of subjects’ responses over time. The ICC (2-way mixed model) reflects the proportion of total variance in item scores associated with differences between the items themselves after the variance because of a specific source has been subtracted. In general, ICCs of 0.70 or greater indicate that items are reliable.
Exploratory factor analysis was used to identify the factor structure underlying the ECES. Factor analysis is a data-reduction technique used to determine the number and nature of factors that underlie a set of variables. Factors are generated that maximize intercorrelations among items within a factor while minimizing correlations between factors. Correlations between the 37 ECES variables were examined based on communality estimates consisting of squared multiple correlations. These communality estimates were derived using multiple iterations to achieve factor stability. The principal axis method was used to extract all factors that had eigenvalues greater than 1 and that therefore explained a significant amount of the total variance. Scree tests were used to identify the number of factors to retain. Because we had little reason to suspect that the factors would be independent of one another, the extracted factors were first rotated by a promax oblique rotation procedure with κ set to a power of 2. We also used a varimax solution to determine if an orthogonal rotation was more appropriate than an oblique rotation. We also performed a higher-order factor analysis on the factor intercorrelation matrix using the principal factor method to determine if it was possible to form a linear combination of the first-order factors.
Based on the results of the factor analysis, we examined the association between the retained factors and clinical and demographic variables. The relationships between each expectation factor and demographic and clinical variables, including age, sex, race, education, potentially blinding or nonblinding eye condition, best corrected visual acuity in the better and worse eyes, patients’ ratings of their overall health, and patients’ ratings of their vision, were tested using univariate linear regression models for each demographic or clinical variable. To identify the demographic and clinical variables that were independently associated with patients’ expectations, multivariate regression models were constructed for each of the expectation factors. We included the following variables in a single model for each dependent variable: age, sex, race, education, type of eye condition (blinding vs nonblinding), best corrected visual acuity in the better and worse eyes, patients’ ratings of their overall health, and patients’ ratings of their vision. A backward stepwise procedure was then followed so that only those variables demonstrating significant independent associations were retained. All tests of association were considered statistically significant at P ≤ .05.
Parameter estimates from these models can be used to assess the magnitude of the associations. For categorical independent variables, each parameter estimate represents the average difference with respect to the dependent variable between a given group and a common reference group (eg, excellent vs poor, very good vs poor, good vs poor, and fair vs poor). The choice of reference group is often arbitrary and is commonly based on convenience, common sense, or sample size. For independent continuous variables, the parameter estimate represents the slope of the linear relationship between the continuous dependent and independent variables and thus can often be used to evaluate the strength of the relationship.
To reach a final sample size of 202 patients, 240 patients were asked to participate while waiting for ophthalmology appointments at Duke University Eye Center and 3 satellite clinics. Of these, 20 (8.3%) were ineligible because they did not speak English, were under 18 years old, had inadequate vision to complete the surveys, failed to score 24 of 30 on the Mini-Mental Status Examination, or were illiterate. Of the 220 eligible patients, 202 (91.8%) agreed to participate. The most common reason given by patients who declined to participate was concern about review of their medical records. Patient clinical and demographic characteristics are summarized in Table 1. Among the 202 eligible study participants, the mean age was 53.3 years, 61.4% were female, and 73.8% were white. Slightly less than half (47.9%) of participants had completed a 4-year college degree. The vast majority (96.4%) of participants had visited an eye doctor previously. Slightly more than half of participants (56.8%) rated their vision as good or excellent. Over half (59.2%) of patients indicated that they had an annual household income greater than $40 000. A slight majority (58.9%) of participants had nonblinding eye conditions, while 41.1% had potentially blinding eye conditions. Study participants included patients of 24 different ophthalmologists.
Estimates of test-retest reproducibility indicate that the ECES is reliable. The ICCs were relatively high. For the majority of items, ICCs fell between 0.80 and 0.95. These findings indicate that subjects gave similar answers to the same questions at different points in time.
A factor analysis was performed on the 37 five-point expectation items to reduce dimensionality and gain conceptual clarity. Regardless of the technique used, we were able to consistently identify 4 factors. The results were therefore based on the promax oblique rotation procedure. The 4 factors that had an eigenvalue greater than 1 explained 89% of the total variance. Table 2 presents the factor names and the expectation items associated with each factor as well as the percentage distribution of patient responses to each of the retained expectation items. Table 3 presents the factor loadings of each of the retained expectation items, the eigenvalues, the percentage of variance accounted for by each factor, and Cronbach α values for each of the 4 factors.
Factor 1, patient involvement in eye care, accounted for 49% of the variance and was defined by items 17, 18, 22, and 31 through 37. Factor 2, interpersonal manner, accounted for 23% of the variance and was defined by items 4 through 9, 14, and 15. Factor 3, information about diagnosis and prognosis, accounted for 10% of the variance and was defined by items 20, 23 through 25, and 27. Finally, factor 4, communication and clinical competence, accounted for 7% of the variance and was defined by items 1 through 3 and 10.
Factors were scored using simple averages of the identified items for each of the 4 factors. The factor analysis identified a subset of 27 questions to represent the 4 factors. Thus, 10 of the original 37 items were not used in creating the factor-based scores.
We examined the association between each of the 4 retained factors and each of the clinical and demographic variables using univariate linear regression analyses (Table 4). The patients’ ratings of their overall health showed no significant association with any of the 4 factors. However, a number of clinical and demographic variables were significantly associated with one or more factors. We then constructed reduced multivariate regression models for each of the 4 expectation factors to identify the demographic and clinical variables that demonstrated significant independent associations with each factor.
Based on the univariate analyses, higher expectations for factor 1 were associated with the following patient characteristics: potentially blinding eye conditions (P<.001), patients’ lower ratings of their own vision (P<.001), poorer best corrected visual acuity in the better eye (P<.001), poorer best corrected visual acuity in the worse eye (P = .01), lower level of education (P = .002), race/ethnicity (P = .002), and lower household income (P = .03). The reduced multivariate model identified only condition type (blinding vs nonblinding) (P<.001) and education (P = .005) as significant independent predictor of patients’ expectations for factor 1.
Based on the univariate analyses, higher expectations for factor 2 were associated with potentially blinding eye conditions (P<.001), diagnostic category (P = .03), patients’ lower ratings of their own vision (P = .04), poorer best corrected visual acuity in the worse eye (P = .004), poorer best corrected visual acuity in the better eye (P = .03), being female (P = .04), lower level of education (P = .004), and race/ethnicity (P = .003). Glaucoma patients (P = .004) and patients with retinal conditions (P = .009) had significantly higher expectations for factor 2 than patients seen for routine eye care. Black patients had significantly higher expectations for factor 2 than white (P = .001) or Asian (P = .002) patients. The reduced multivariate model identified only condition type (blinding vs nonblinding) (P = .008), education (P = .03), and race (P = .04) as significant independent predictors of patients’ expectations for factor 2.
Univariate analyses identified only patients’ lower ratings of their own vision (P = .001) as being associated with higher expectations for factor 3. The reduced multivariate model confirmed patients’ ratings of their own vision (P = .001) as a significant predictor for factor 3.
No patient characteristics were significantly associated with expectations for factor 4 based on the univariate analyses. Univariate analyses found that being female (P = .05) and higher household income (P = .06) were associated with higher expectations for factor 4; however, these characteristics were at or slightly above the threshold for significance (P≥.05). The reduced multivariate model confirmed being female (P = .04) and higher household income (P = .01) as significant predictors for factor 4. In addition, the reduced multivariate model also identified patients’ ratings of their own vision (P = .07) and age (P = .05), although these variables were at or slightly above the threshold for significance.
The ECES provides an empirical assessment of the range and definition of patient expectations regarding eye care. Evaluation of patients’ expectations using the relatively short ECES provides valuable information regarding the types of care desired by ophthalmology patients. In addition, the ECES appeared to legitimize and facilitate ophthalmology patients’ identification of their own perceived desires and needs.
The ECES factor analysis revealed 4 categories of expectations: Factor 1, patient involvement in eye care, refers to issues related to medications, test results, paperwork, referral, information about advances in eye care, and lifestyle. Patient expectations are consistent with increasing trends toward collaborative decision making in medicine.20 Moreover, there is growing evidence that higher levels of patient involvement in care are linked to better clinical outcomes.21,22 Factor 2, interpersonal manner, refers to personal connection, empathy, courtesy, encouragement, professionalism, length of interaction, and appointment access. Evidence suggests that a physician’s affect toward patients is closely correlated with patient satisfaction.23,24 A previous literature review found that one of the most strongly supported relationships is the connection between “personal” care and high levels of satisfaction.25 Factor 3, information about diagnosis and prognosis, refers to the eye examination, diagnosis, etiology, and prognosis. A previous study of ophthalmology patients found that communication of medical information regarding diagnosis, prognosis, and treatment was a significant determinant of patient satisfaction.26 Factor 4, communication and clinical competence, refers to honesty, explanation in clear language, listening/addressing concerns, and clinical competence. Previous studies have demonstrating that patients have high expectations for medical information.1,26- 30
To our knowledge, no previous expectations research has produced an overall factor structure equivalent to ours. However, factors that we identified exhibit similarities to individual factors identified in previous expectations research, mostly conducted in primary care settings. “Patient involvement in eye care” is similar to the “mutual patient-physician responsibilities” factor identified by Greene et al31 and is related to the “medication effects” and “medication regimen and visits” factors identified by Joos et al.7 “Interpersonal manner” is related to the “ventilation and legitimization” factor identified by Good et al32 and the “professional attitudes and behaviors” and “personal confidant of patient” factors identified by Feletti et al.33 “Information about diagnosis and prognosis” is similar to the “investigation and treatment” factor identified by Valori et al34 and the “test results” factor identified by Good et al. “Communication and clinical competence” is similar to the “technical competence” and “communication, care, and reassurance” factors identified by Feletti et al. Although the 4 factors we identified share common themes with factors outlined in previous studies, significant differences in the overall factor structure are apparent. It is unclear whether this variation stems from differences in expectations between primary care and ophthalmology patients, differences in the research approach, or other sources.
The ECES also facilitated investigations of the relationships between expectations and important clinical and demographic variables. Prior to the study, we anticipated that patients with potentially blinding eye conditions would report higher levels of expectations regarding their eye care. The reduced multivariate models confirmed this hypothesis for factors 1 and 2 but not for factors 3 and 4.
Factors 1 (patient involvement in eye care) and 2 (interpersonal manner) emerged as the strongest factor subscales. Together these factors accounted for 72% of the variance. Patients with potentially blinding eye conditions had significantly higher expectations for both patient involvement in eye care and interpersonal manner than did patients with nonblinding eye conditions. It is unsurprising that patients with underlying eye disease of greater severity desire more involvement in their care and greater interpersonal support from their eye care provider.
The finding that patients who had completed fewer years of education had significantly higher expectations for both factors and that black patients reported significantly higher expectations than white patients for factor 2 may appear somewhat surprising. However, this result is consistent with the findings of a previous study by Kravitz et al,35 who found that members of minority groups with fewer years of education had higher expectations.
Factor 3 (information about diagnosis and prognosis) explained 10% of the variance. Patients’ lower ratings of their own vision was the only demographic or clinical variable that demonstrated a significant independent association with higher expectations for information about diagnosis and treatment. It is likely that patients distressed by visual symptoms may consult an ophthalmologist to discover a diagnostic label, an etiologic formulation, or a prognosis for their eye condition. However, it is unclear why patients with potentially blinding eye conditions did not have significantly higher expectations for factor 3, as they did for factors 1 and 2. One possible explanation is that the majority of patients classified as having potentially blinding eye conditions may have already received the diagnostic and prognostic information that they desired but that a smaller proportion of patients who reported poorer ratings of their own vision had received this information. Another possibility is that patients with worsening vision may report poorer ratings of their own vision, whereas the blinding vs nonblinding classification does not capture such trends.
Factor 4 (communication and clinical competence) accounted for 7% of the variance. The reduced multivariate model showed that sex and household income were significant predictors of expectations for communication and clinical competence. The observation that women place greater emphasis on communication may reflect more general sex-related differences. This finding is also consistent with our previous work with parents of pediatric ophthalmology patients, which suggested that mothers were more likely than fathers to identify expectations related to communication items.36 The finding that higher income was associated with higher expectations for factor 4 is somewhat surprising given that, in general, higher socioeconomic status was associated with lower expectations for factors 1 and 2. The reason for this discrepancy is unclear.
There are several possible limitations to this research investigating patient expectations regarding eye care. More female than male patients were enrolled in the study. However, the participation rate of male and female patients was essentially equivalent. These findings are consistent with those of a number of previous studies from the expectations literature that also had greater than 60% female participants.32,34 Moreover, in our study, sex differences were only significant in participants’ expectations for factor 2. It is important to consider the research setting as well. This research was conducted at sites affiliated with an academic eye center. Expectations among patients visiting university-affiliated sites may be different from those of patients in community ophthalmology practices not associated with a university. However, it is worth noting that 59% of the participants in this study were being seen for routine or refractive eye problems.
The focus group method used during the first phase of ECES development17 was selected to maximize the content validity of the questionnaire for a wide cross-section of ophthalmology patients. The use of a diverse patient population in the current study and tests of association with clinical variables support the construct validity of the survey. Nevertheless, the stability of the 4-factor structure of the ECES still awaits cross-validation with additional samples of ophthalmology patients.
Overall, the initial survey development and results are promising. A patient expectations survey may be a helpful clinical tool that has the potential to improve eye care. Further efforts to validate the conceptual model will be needed to determine the clinical and research utility of the ECES.
Correspondence: Paul P. Lee, MD, JD, Box 3802, Duke University Medical Center, Durham, NC 27710 (email@example.com).
Submitted for Publication: April 1, 2003; final revision received August 30, 2004; accepted August 30, 2004.
Financial Disclosure: None.
Funding/Support: This work was supported in part by a grant from Research to Prevent Blindness, New York, NY (Dr Lee is a recipient of the Lew Wasserman Merit Award), and by a gift from the Eberly family.
Acknowledgment: We gratefully acknowledge the assistance of the 24 physicians and their patients who enabled us to carry out this study. We also thank Cecilia Santiago-Turla, MD, for her valuable help with this study.