The present report describes a longitudinal follow-up study of a sample of patients managed with total hip arthroplasty who had been enrolled in a previous study that was performed to compare preoperative function and the expectations for surgery12. Patients were enrolled in the preoperative study if they were able to speak English and were being evaluated for total hip arthroplasty. Patients were included in the current postoperative study if they had undergone total hip arthroplasty at the Hospital for Special Surgery and were available for review. During the preoperative evaluation, patients completed the self-administered American Academy of Orthopaedic Surgeons Lower Limb Core (Hip/Knee Module) questionnaire, a seventeen-item region-specific scale measuring symptoms and physical function13. Questions on this scale address the use of assistive devices, stiffness, the ability to put on shoes and socks, and pain in each hip and knee when walking on flat surfaces, going up or down stairs, and lying in bed. Response options are ranked according to severity, ranging from "not painful" to "could not do because of hip pain." The overall score can range from 0 (worst condition) to 100 (best condition). Patients also completed the Hospital for Special Surgery Hip Replacement Expectations Survey, a validated questionnaire measuring eighteen expectations for symptom relief and improvement in physical function and psychological well-being (as described below)12. Medical records were reviewed for clinical characteristics.
For the current postoperative study, patients were contacted and interviewed by telephone approximately four years after total hip arthroplasty. Five attempts were made, at different times of the day, to contact patients for the telephone follow-up. The patients were asked about their current condition with use of the Lower Limb Core and were also asked if they had had any complications (such as dislocations) as a result of total hip arthroplasty. They were told what expectations they had cited preoperatively and were asked to what extent each expectation was now fulfilled (completely, somewhat, or not at all). Patients were asked to rate their overall assessment of surgery with use of a single question that has been validated for general well-being14. The question was tailored to total hip arthroplasty as follows: "If you were to spend the rest of your life with your hip symptoms just the way they have been in the last twenty-four hours, how would you feel?" The seven response options ranged from "delighted" to "terrible." Patients also were asked whether they would have the surgery again if necessary and how satisfied they were with the results of surgery; the five response options for the latter question ranged from "very satisfied" to "very dissatisfied."
Data Analysis
Demographic and clinical characteristics were described with means and frequencies. The proportion of expectations fulfilled was calculated as the number of expectations fulfilled completely divided by the total number of expectations cited by that patient and was assessed with bivariate analysis with demographic and clinical characteristics. Given that the outcome was a count variable out of a known total and was skewed, the outcome was modeled with a binomial regression. Maximum likelihood was used to fit the model and to estimate the regression parameters. Both preoperative and postoperative demographic and clinical characteristics were independent variables.
The institutional review board at the Hospital for Special Surgery approved this study, and patients provided informed consent.
Source of Funding
The New York Chapter of the Arthritis Foundation supported this work and did not play a role in the design or interpretation of this study. Funding was used for salaries and supplies.
In total, 885 patients from the preoperative study had a total hip arthroplasty and 487 were contacted for the present longitudinal follow-up study. Compared with the patients who were contacted, those who were not contacted were more likely to be men (49% compared with 42%; p = 0.03) and to have been younger at the time of surgery (mean age, sixty-three compared with sixty-seven years; p < 0.0001), but they did not differ in terms of the mean number of expectations or Lower Limb Core scores (p > 0.05 for each variable). Of the 487 patients who were contacted, 405 were included in this analysis and eighty-two were excluded. Of the latter eighty-two patients, thirty-six did not want to participate in the full interview (thirty-three reported good outcomes and three reported poor outcomes), thirteen could not participate because of illness or hearing impairment, eleven did not speak English well enough for a telephone interview, and twenty-two were reported by family members to have died. The excluded patients were older (mean age, seventy-three compared with sixty-six years, p < 0.0001) but did not differ in terms of sex, the number of expectations, or Lower Limb Core scores (p > 0.05).
The mean time between total hip arthroplasty and follow-up was 4.4 years (range, three to six years), and the patients came from the practices of six orthopaedic surgeons (including two of the authors [E.A.S. and T.P.S.]). At the time of surgery, the mean age was sixty-six years (range, twenty-eight to eighty-eight years); 9% of the patients were fifty years old or less and 3% were forty years old or less (Table I). Fifty-eight percent of the patients were women, 96% were white, 71% were married, 66% were overweight or obese, and 8% had a body mass index of =35 kg/m2. Forty percent were working, 40% were retired, 12% were homemakers, and 8% were unemployed or disabled. Most patients underwent total hip arthroplasty for the treatment of osteoarthritis, and 7% had a bilateral procedure (either simultaneously [5%] or staged [2%]). The preoperative Lower Limb Core score reflected marked limitations.
Postoperatively, 29% of the patients continued to work, 3% were newly employed, 39% continued to be retired, 13% were newly retired, 12% continued to be homemakers, and 4% were unemployed or disabled. Twenty patients (5%) had complications resulting from total hip arthroplasty, mainly dislocation (eleven), infection (six), and loosening (two), and, of these, eight had a revision. Thirty-one percent of the patients had a residual limp. Scores for the Lower Limb Core improved, and the mean within-patient change (and standard deviation) was 34 ± 19 (p < 0.0001).
Patients had a spectrum of expectations related to physical and psychological health (Table II). Nineteen percent of the patients cited all eighteen possible expectations. The mean number of expectations per patient was sixteen, and the lowest number was four. The most common expectation was to improve walking (99% of patients), and the least common was to be employed for monetary reimbursement (42% of patients).
Forty-three percent of the patients reported that all of their expectations had been fulfilled completely, and an additional 32% reported that all of their expectations had been fulfilled somewhat. One patient reported that none of her expectations had been fulfilled to any extent. This patient had no remarkable preoperative characteristics, but, at the time of follow-up, she reported a hip infection, neurological deficits in the involved leg, and residual hip pain.
All subsequent results are presented for the outcome of having expectations fulfilled completely. The percentage of patients who reported that specific expectations had been fulfilled ranged from 63% (for the ability to cut toenails) to 92% (for psychological well-being and the ability to participate in recreational or social activities). The fulfillment of specific expectations did not vary according to most demographic characteristics; however, college graduates were more likely to have their expectations fulfilled with regard to the use of an assistive device, activities outside the home, and employment (p < 0.05); patients with a lower body mass index were more likely to have their expectations fulfilled with regard to climbing stairs and cutting toenails (p < 0.05); and employed patients were more likely to have their expectations fulfilled with regard to the use of an assistive device, sexual activity, social activities, and cutting toenails (p < 0.05). In addition, 57% of the patients who had been employed preoperatively expected to be working postoperatively and 100% reported that this expectation had been fulfilled. In comparison, 32% of patients who had not been working preoperatively expected to be working postoperatively and 77% reported that this expectation had been fulfilled (p < 0.0001). The preoperative variable most consistently associated with fulfilled expectations was functional status as measured with the Lower Limb Core. Patients with better preoperative Lower Limb Core scores were more likely to have fulfillment of thirteen of the eighteen specific expectations (with the exceptions of daytime pain, transfer, exercise, cut toenails, and psychological well-being).
The mean proportion of expectations fulfilled, calculated for each patient as the number of expectations that had been fulfilled completely divided by the total number of expectations cited, was 87% ± 22%, and the median was 94%.
The proportions of expectations fulfilled according to preoperative and postoperative variables are presented in Table III. For preoperative variables, patients who were forty years old or less, who were employed, who had a body mass index of <35 kg/m2, and who had better Lower Limb Core scores had a greater proportion of expectations fulfilled. Patients who had been managed with bilateral total hip arthroplasty had a greater proportion of expectations fulfilled (94%) in comparison with those who had been managed with unilateral total hip arthroplasty (89%), but this difference was not significant (p = 0.90). For postoperative variables, patients who were employed, who did not have complications, who did not have a limp, and who had better Lower Limb Core scores had a greater proportion of expectations fulfilled. In multivariate analysis including both preoperative and postoperative variable, having a postoperative limp and having better preoperative and postoperative Lower Limb Core scores remained significant (Table IV).
The proportion of expectations fulfilled also was compared according to the patients' overall assessment of total hip arthroplasty. When patients were asked how they would feel if the hip symptoms that they had had during the last twenty-four hours were to become permanent, 58% stated that they would be delighted, 15% stated that they would be pleased, 21% stated that they would be mostly satisfied, 2% stated that they would have mixed feelings, 2% stated that they would be mostly dissatisfied, 1% stated that they would be unhappy, and 1% stated that they would feel terrible. Patients who had a favorable response (i.e., those who stated that they would be delighted, pleased, or mostly satisfied) had a greater proportion of expectations fulfilled (90%) in comparison with those who did not have a favorable response (39%) (p < 0.0001). Overall, 94% of patients stated they were "very satisfied" and 96% stated that they would have the surgery again.
In this prospective cohort of patients, 87% of preoperatively-cited expectations were fulfilled completely when assessed approximately four years after total hip arthroplasty. We found that 43% of patients reported that all of their expectations had been fulfilled completely and an additional 32% reported that all of their expectations had been fulfilled somewhat. Several preoperative and postoperative demographic and clinical variables were associated with the fulfillment of expectations and, among these, functional status was the most closely associated.
In addition to providing a global picture of expectations, the present study revealed important information about specific expectations and patient characteristics that may be useful during preoperative discussions with patients. For example, among patients who expected to be employed, those who were not working preoperatively were less likely to have this expectation fulfilled in comparison with those who were working. Patients who had a body mass index of =35 kg/m2 were less likely to have their expectations fulfilled. The youngest patients, defined as those who were forty years of age or younger, were the most likely to have their expectations fulfilled. Although we did not consider patients' views on prosthetic longevity, this finding indicates that the youngest patients were most likely to obtain what they originally sought from total hip arthroplasty.
The results of the present study may have implications for when to perform total hip arthroplasty. While it was anticipated that a better postoperative Lower Limb Core score would be the most important variable associated with fulfilled expectations, it was interesting to find that a better preoperative score also was associated with more fulfilled expectations. In the multivariate model, accounting for other pertinent variables, the preoperative score remained independently associated with the proportion of fulfilled expectations. This finding indicates that although "where a patient ends up" is the most important variable, "where a patient starts out" also is important. In the previous cross-sectional analysis that preceded the present study, we reported that patients with worse preoperative Lower Limb Core scores had more expectations and were more likely to rate their expectations as very important12. At that time, we hypothesized that this may represent an unrealistic scenario and that all of these expectations might not be realized. Our current findings confirm this initial hypothesis. In this analysis, patients were more likely to have their expectations fulfilled if they were not the most severely impaired at the time of total hip arthroplasty. Choosing when to perform total hip arthroplasty, however, is a complex decision that must be balanced by surgical factors, other outcome goals, and patient expectations.
Another interesting finding was that having a postoperative limp was associated with more unfulfilled expectations. It seems intuitive that a limp would result in decreased function, and this would be accounted for in the postoperative Lower Limb Core score. However, the fact that having a limp remained independently associated with the outcome in the multivariate model implies an additional mechanism. During the original development of the total hip arthroplasty survey, patients often stated that having a limp was a stigma of disability and that getting rid of a limp was an important physical and psychological expectation of surgery5. It is possible that a residual postoperative limp also has an adverse psychological impact that subsequently affected how patients rated the fulfillment of all expectations.
To date, patient-centered assessments of total hip arthroplasty have focused mostly on satisfaction and symptom relief and have considered expectations in relation to these outcomes. For example, recent studies have shown that fulfilled expectations were associated with greater improvements in terms of symptoms and function after total hip arthroplasty and greater overall satisfaction after total knee arthroplasty2,3. In another study, fulfillment of only certain expectations, such as pain relief, was associated with satisfaction with both total hip arthroplasty and total knee arthroplasty1. In a previous study, we also found a high correlation between satisfaction and the expectations for total hip arthroplasty5. Thus, to date, the fulfillment of expectations has been considered mostly in the context of satisfaction. However, satisfaction is a global concept that is influenced by multiple factors, including personality and perceived quality of care. In the current study, in addition to satisfaction, we used a validated and more tangible measure of overall outcome, specifically, the attitude toward the symptoms during the last twenty-four hours becoming permanent. We found that the fulfillment of expectations was strongly correlated with this overall evaluation. The association between fulfilled expectations and this global outcome provides preliminary support for the validity of fulfilled expectations as an outcome unto itself.
The present study had several limitations. First, it was carried out in an orthopaedic referral center and may not be generalizable to patients in different settings. Second, we did not contact all eligible patients; however, our response rate is not atypical of long-term telephone follow-up studies15. In addition, those who were not contacted did not differ from those who were contacted in terms of baseline expectations and functional status. Third, we did not consider certain patient and clinical variables, such as race, diagnosis, and the type of prosthesis, which may affect the frequency and assessment of expectations4. Fourth, given that patients were drawn from multiple orthopaedic practices, we did not have uniform ratings for the severity of hip disease or surgical approach. These variables may have implications for certain analyses, particularly postoperative limp.
In the present study, the proportion of expectations that were fulfilled was high and was related to both preoperative and postoperative patient and clinical characteristics, particularly functional status. Our findings may be useful for counseling patients and for providing a framework for using fulfilled expectations as another patient-centered outcome of total hip arthroplasty. 