Pager CK. Expectations and Outcomes in Cataract SurgeryA Prospective Test of 2 Models of Satisfaction. Arch Ophthalmol. 2004;122(12):1788-1792. doi:10.1001/archopht.122.12.1788
To document patients’ preoperative expectations for postoperativeoutcomes. To measure the relative contribution of patient understanding, expectations,outcome, and expectation-outcome discrepancy in determining patient satisfaction.
One hundred twenty-one patients were surveyed just before and 1 monthafter cataract surgery regarding their understanding of the procedure, satisfactionwith their vision, and both current and expected visual function for eachof the items on the Visual Function Index (VF-14).
Sixty percent of patients expected to achieve a perfect VF-14 score.The average expected VF-14 score was 96.1, compared with an achieved VF-14score of just 89.8. The most unrealistic expectations involved driving atnight, reading small print, and doing fine handiwork. Surprisingly, improvementin visual function was not correlated with satisfaction in vision. While patientunderstanding, expectations, and achieved VF-14 score did correlate with satisfaction,when controlling for other factors, only achievement-expectation discrepancywas independently predictive.
This study provides support for the expectation-outcome discrepancymodel of patient satisfaction. Further, it highlights the highly unrealisticexpectations harbored by patients with cataract and emphasizes the importancefor physicians to control their patients’ expectations. Controllingpatient expectations may be more effective than improving patients’postoperative outcome in terms of maximizing patient satisfaction.
Patient satisfaction is an increasingly important objective for healthservices.1,2 This trend reflectsthe reality that the choice and success of many treatments are based on subjectivepatient-defined criteria2- 4;that patient satisfaction is an element of health status itself,5,6 withsatisfied patients reporting greater compliance, well-being, and treatmentoutcomes7- 9; thathealth care is increasingly embracing principles of consumerism and autonomy1,10; that health care is becoming increasinglyprivatized and economically competitive,11,12 withsatisfied patients both remaining with and recommending their provider7,13- 15; andthat a satisfied patient is the best defense against malpractice lawsuits.11
The dominant theoretical model of satisfaction involves meeting patientexpectations, that is, minimizing the expectation-outcome discrepancy. Thisconception of satisfaction has been variously reported as its actual definition11,16,17 or its primary causalfactor.4,10,11,18 Suggestedmechanisms for this effect include cognitive dissonance, patient conditioning,memory and symptom reporting, and anxiety.19,20
However, the importance of expectation-outcome congruence has not alwaysbeen borne out empirically,19,21,22 withthe actual (perceived) outcome18,23 orexpectations alone via the placebo effect10,20,24 accountingfor most of a patient’s satisfaction in some studies. Additionally,the extent to which patients feel adequately informed about their illnesshas been frequently demonstrated as an important and independent determinantof patient satisfaction.1,25
Cataract surgery has increased 400% during the last 10 years26 to become the most common surgical procedure in thedeveloped world,27 exceeding 1.6 million operationsin the United States alone28 because of anaging population29,30 and dramaticexpansion of indications following improvements in technology.31,32 Visualacuity is no longer considered a useful measure of need for surgery,33,34 with patient concern, lifestyle,and subjective reports of function being the primary factors for referral.35,36 Furthermore, cataract surgery ispredominantly delivered through the private health care sector,31 wherepatient satisfaction is of particular consequence.
However, patient satisfaction has rarely been considered in cataractsurgery,34,37 and no studies haveaddressed the role of the hypothesized determinants of patient satisfaction.The objectives of this study are first, to document patients’ preoperativeexpectations for postoperative outcomes and second, to measure the relativecontribution of patient understanding, expectations, outcome, and expectation-outcomediscrepancy in determining patient satisfaction.
One hundred sixty consecutive patients undergoing day-stay cataractsurgery at Sydney Private Hospital, Sydney, Australia, were invited to participatein this study when first registering at the hospital for their surgery. SydneyPrivate Hospital is the largest private provider of cataract surgery in NewSouth Wales and serves a large cross-section of patients from throughout metropolitanSydney. Institutional review board and hospital approval were obtained.
After obtaining informed consent, the patient’s age, sex, visualacuity, and current visual function were recorded by a research interviewer,along with the amount of information the patient had already received regardingthis surgery. Visual function was measured using the Visual Function Index(VF-14),38,39 a widely used scalebased on trouble conducting common binocular activities, with final scoreranging from 0 (no visual ability) to 100 (no visual disability). Patientswere then asked to rate their expected postoperative functional outcome foreach of the 14 items on the VF-14 scale. As part of a related investigation,patients were shown a short video describing either the anatomy of cataractor the procedures experienced during the cataract surgery itself. Neithervideo concerned the postoperative outcomes or expectations that are the subjectof this study. All the data collection occurred 2 to 4 hours prior to surgerybut before the instillation of eye drops, changing of clothes, or any clinicalinduction beyond the registration forms.
On hospital discharge, patients were given an addressed, postage-paidenvelope containing a questionnaire to be completed 1 month after surgery.This questionnaire included the VF-14 and the opportunity for patients torate their degree of understanding and overall satisfaction on a 12-cm visualanalog scale. Patients who did not return their questionnaires were remindedby a follow-up telephone call, and an additional questionnaire and reply envelopewere provided if required.
The visual analog scale markings were converted into a scale of 0 to10, and all data were double entered and verified using an Excel spreadsheet,then converted into SPSS for Windows version 11 (SPSS Inc, Chicago, Ill) forfurther analysis.
Spearman rank correlations were used to measure the relationship betweenoverall satisfaction and aspects of expected or achieved visual function (eachitem of the VF-14 is recorded on a 5-point ordinal scale). A multiple regressionwas used to measure the isolated contribution of postoperative function, improvement,and expectation-outcome discrepancy toward predicting satisfaction, whilecontrolling for the other factors. Finally, both Mann-Whitney U and 2-tailed, independent-sample t testswere used to ensure that the preoperative video content had no influence onany of the measured outcomes.
One hundred sixty patients were approached, and 141 (88%) agreed toparticipate. Of these, 121 (85%) returned their 1-month questionnaire. Therewere no differences in preoperative characteristics between these 121 patientsand those who were lost to follow-up (Mann-Whitney U and t tests not shown). The basic preoperative and postoperativecharacteristics of these patients are presented in Table 1. The mean (SD) age was 73.8 (9.2) years, 91 (64.5%) patientswere female, 55 (39%) patients had previous cataract surgery, and the medianvisual acuity was 6/18 in the operative and 6/10 in the nonoperative eye,with a mean (SD) preoperative VF-14 score of 84.7 (14.8) and postoperativeVF-14 score of 89.8 (16.4). There were no preoperative or postoperative differenceson any measure based on video content (Mann-Whitney U and t tests not shown).
Patients had very high expectations for postoperative function, expectingto achieve a mean (SD) VF-14 score of 96.1 (10.3), representing near-perfectvision and an 11-point expected gain. Eighty-four patients (60%) expectedto achieve a perfect VF-14 score of 100. Only 18 patients did not expect atleast some VF-14 improvement, 12 because their preoperative VF-14 score wasalready 100; the other 6 all had preoperative VF-14 scores higher than 90and expected the same VF-14 score postoperatively.
Table 2 presents for each itemof the VF-14 scale the expected amount of improvement and the expectation-outcomediscrepancy. Driving at night, reading small print, doing fine handiwork,and reading a newspaper or book are the items for which patients expectedthe greatest degree of improvement and for which, not unexpectedly, patients’actual outcomes were the furthest from what they expected. The average expectation-outcomediscrepancy was 6.3 points, with 80 patients (66%) failing to equal or exceedtheir expectations at 1-month follow-up.
Table 3 presents the correlationsbetween patient satisfaction and either actual outcome or expectation-outcomediscrepancy for each item on the VF-14 scale. Actual outcome and satisfactionwere significantly correlated with 11 VF-14 items and more strongly than the6 VF-14 items for which expectation-outcome discrepancy and satisfaction werecorrelated.
Table 4 presents the correlationsbetween patient satisfaction and overall expected improvement, actual improvement,actual outcome, and expectation-outcome discrepancy. Actual outcome was marginallybetter correlated than achievement-outcome discrepancy (0.32 vs 0.28), andthe degree of improvement was not significantly related to patient satisfactionat all.
Table 5 presents a multiple regressionof the significant predictors of patient satisfaction. When controlling foreach other, only the expectation-outcome discrepancy was a significant independentdeterminant of patient satisfaction (standardized β = −0.30; P = .04). Multiple regressions with other configurationsof predictor variables consistently returned expectation-outcome discrepancyas the only independently predictive factor of patient satisfaction.
Before the operation, the majority of patients (115 [81%]) felt at thetime of operation that they “learned as much as I wanted to know”and only 19 (13%) felt that they’d like to have been told more. At 1month, the first group’s mean (SD) overall satisfaction was 8.0 (2.0),which was not significantly different from the second group’s satisfactionof 7.9 (2.0) by Mann-Whitney U test, nor were preoperativeinformation and 1-month postoperative satisfaction significantly correlated.However, the patients’ responses to this preoperative question aboutinformation were narrowly distributed and temporally distant from the 1-monthpostoperative question about satisfaction.
More significantly, patients’ response to the question “howwell did you understand what was happening to you” at 1 month aftersurgery (mean [SD] 8.0 [2.0]) was strongly correlated with patients’postoperative satisfaction (ρ = 0.56; P<.001).
The results of this study lend support to the applicability of the expectation-outcomediscrepancy model of patient satisfaction in the context of day-stay cataractsurgery. Importantly, the improvement in visual function experienced by apatient did not at all significantly correlate with overall satisfaction,a finding that runs counter to common expectations. However, expected improvement,actual outcome, and expectation-outcome discrepancy were all moderately correlatedwith satisfaction. Clearly, patient satisfaction is a complex and multidimensionalconstruct that cannot be explained by a single variable.7,10,24 However,it is worth noting that when controlling for other factors, neither actualimprovement nor actual outcome were predictive of patient satisfaction, whilethe expectation-outcome discrepancy was.
These findings further reinforce the only previous study to formallydocument patient expectations,40 demonstratingthat patient expectations are unreasonably high and few patients realisticallyachieve them. In short, 70-year-old patients expect cataract surgery to allowthem to see like 20-year-olds, although at this age there are inevitably otherfactors beside cataract that will impair their vision to some degree. It isnot surprising that ocular comorbidity is a major predictor of patient dissatisfactionwith cataract surgery.41,42
As with the previous study, reading small print, doing fine handiwork,and reading a newspaper were the most likely areas of high expectations forimprovement (as well as disappointment with outcome). For our sample, drivingat night was also a source of unrealistic and unfulfilled expectations forthe 34 patients (24%) who did so.
Patients’ perceived level of understanding at 1 month was stronglycorrelated with satisfaction. Some part of this correlation may be owing toan overarching response bias toward visual analog scales presented at thesame sitting, despite the fact that the questions were presented as clearlyindependent and visual analog scales have been shown to be both reliable andvalid measures in similar contexts.43,44 Additionally,the correlation between perceived understanding and satisfaction suppliesno evidence as to the direction of causation. Nonetheless, it is likely thatperceived understanding, as with expectation-outcome discrepancy, is an importantcomponent of patient satisfaction.
Although the strength of correlations was modest, explaining just 4%to 17% of the variance (r2), they werestrongly significant and consistent in magnitude with other studies of satisfaction,the determination of which is highly multifactorial. We must also acknowledgethat the difference between interviewer-assisted preoperative and self-administeredpostoperative questionnaires may affect the scores obtained, although thisdifference was consistent across all patients and should not affect the findingsof this study. Furthermore, the VF-14 in particular has been extensively validatedand found to be both stable and reliable across a wide range of contexts.38,39 This study did not investigate thechanging trajectory of patient expectations across time prior to surgery,nor the impact on satisfaction on the tiny minority who experienced operativecomplications. Further research into these subsidiary questions could providea more nuanced view of the relationship between expectations, outcome, andsatisfaction.
To provide the highest level of satisfaction, health care professionalsmust control their patients’ expectations and understanding of treatment,11 and the close relationship between accurate patientexpectations and adequate informed consent cannot be ignored.40 Infact, to improve patient satisfaction, health care professionals would beadvised to pay more attention to patient understanding and expectations, evenat the expense of improving patient outcome. In this study, the degree ofimprovement in visual function was not significantly correlated with patientsatisfaction at all.
This study highlights the particular areas where greater attention toinforming patients’ preoperative expectations would be warranted, aswell as important rationales for doing so. Finally, this study strongly supportsthe proposition that perceived patient understanding and expectation-outcomediscrepancy are important factors in overall patient satisfaction.
Correspondence: Chet K. Pager, BMed (Hons),DipEd, MA, 67 Missenden Rd PO, Royal Prince Albert Hospital, Camperdown NSW2050, Australia (firstname.lastname@example.org).
Financial Disclosure: None.
Submitted for Publication: October 1, 2003;final revision received March 29, 2004; accepted June 29, 2004.
Funding/Support: This study was supported inpart by the National Health and Medical Research Council of Australia, Canberra.