Reliability and concurrent validity of the Dutch hip and knee replacement expectations surveys

Osteoarthrosis is the most common joint disorder in the world [1]. Patients with osteoarthrosis of the hip or knee joint experience pain, stiffness and loss of joint function. When conservative treatment does not result in less pain and better functioning, a total hip or knee replacement is the most common and successful surgical treatment. In 2005, the incidence of hip replacements in the Netherlands was 124 per 10⁵ inhabitants (20,281 operations); the incidence of knee replacements was 63 per 10⁵ inhabitants (10,329 operations) (Statistics Netherlands 2005).

In orthopaedics, increasing emphasis is placed on patient-reported outcome of the surgery, patient satisfaction and quality of life, and not solely on the technical success of the surgical procedure and the surgeon's rating of the outcome. Previous research has indicated that preoperative expectations are important determinants of patients' satisfaction and functional outcome of total joint replacement [2‐6]. Patients have multiple expectations of the outcome of total hip or knee replacement, mainly concerning relief of pain and improvement in physical function and psychosocial well-being [3]. Fulfilled expectations are linked to increased patient compliance with postoperative recommendations and return to follow-up care and monitoring [7]. Unrealistically high expectations can result in discouraged patients postoperatively and non-adherence with recommendations postoperatively, while unrealistically low expectations can result in less motivation to obtain full benefit from the surgery [7].

Additionally, research indicates that differences exist between the ratings of patients and surgeons on the outcome of the surgery [8]. Among the many explanations for this difference, it is hypothesised that a difference in expectations plays an important role [9]. It is therefore important to assess patients' expectations before surgery. Especially patients with a poor preoperative status often have high expectations which are potentially unrealistic [10].

To avoid unrealistic expectations, it has been recommended to query patients about their expectations before surgery [3, 11]. Physician-patient discussions about preoperative expectations should be an important part of clinical care [12]. Moreover, preoperative education classes have shown the ability to change the expectations and result in more equal expectations of patient and surgeon [13, 14]. Questionnaires can be used to guide discussions and evaluate interventions. However, no Dutch questionnaires are available to determine patients' expectations before total hip or knee arthroplasty. Therefore, the aims of the current research were (1) to translate the English-language Hospital for Special Surgery Hip Replacement Expectations Survey [3, 10, 13] and the Hospital for Special Surgery Knee Replacement Expectations Survey [13, 15] into Dutch according to international guidelines as described by Beaton et al. [16], and (2) to study test-retest reliability and concurrent validity of the two Dutch-language surveys.

As no questionnaires are available in the Dutch language to determine preoperative expectations of patients on a waiting list for total hip or knee arthroplasty, the first aim of this study was to translate the English-language Hospital for Special Surgery Hip Replacement Expectations Survey and the Hospital for Special Surgery Knee Replacement Expectations Survey [3, 10, 13, 15]. The surveys were translated according to the method described by Beaton et al., which is the official method according to the American Association of Orthopaedic Surgeons (AAOS) [16].

Second aim of the study was to determine the test-retest reliability and concurrent validity of the Dutch Hip Replacement Expectations Survey and the Dutch Knee Replacement Expectations Survey. The results of the reliability study show that both Dutch surveys have good test-retest reliability and internal consistency. The Bland and Altman analyses indicated no bias between the first and second measurements. The Intraclass Correlation Coefficients for the total scores were close to or above the criterion of 0.80 of Nunnally and Bernstein [21], and can therefore be considered high (0.79 and 0.87 for the Knee and Hip Survey, respectively). The item that scored the lowest ICC in the Dutch Hip Replacement Expectations Survey was the expectation of the ability to exercise or participate in sports. By contrast, this item scored high in the Dutch Knee Replacement Expectations Survey. When looking at the individual items of the Dutch Knee Replacement Expectations Survey, the lowest ICCs were found for the expectation regarding the ability to change position and the ability to climb stairs. This latter item also had a moderate ICC in the Dutch Hip Replacement Expectations Survey. It can only be speculated why patients rate their expectations differently when assessed twice. One reason might be that patients find it hard to estimate a certain expectation, resulting in a different answer at the second assessment. Overall, all ICCs were moderate to high (between 0.44 and 0.83) and the differences between the items are small, indicating good test-retest reliability of both surveys. Compared to the original English-language surveys, the internal consistency as determined with Cronbach's alpha was higher in the Dutch-language surveys (Hip 0.86 vs 0.77 in the original version; Knee 0.91 vs 0.79 in the original version) [13]. Both values satisfied the minimum criterion of 0.80 set by Nunnally and Bernstein [21]. No additional data is available concerning the reliability of the original surveys, therefore further comparison with the English-language version is not possible.

As there is no instrument available to determine patient expectations which can be considered the gold standard, the only available questionnaire described in the literature, the Expectation WOMAC, was chosen to determine concurrent validity of the Dutch Hip Replacement Expectations Survey and the Dutch Knee Replacement Expectations Survey [17]. Although the Dutch WOMAC is considered reliable and valid, and is only slightly adapted to result in the Expectation WOMAC, the psychometric properties of the Expectation WOMAC are unknown. To determine concurrent validity, the Pearson's correlations are calculated between the two Dutch Expectations Surveys and the Expectation WOMAC, which were moderate; a correlation between 0.4 and 0.6 is evidence that the same construct is being embraced [23]. However, the Bland and Altman analyses showed considerable bias between the two measures; the mean Expectation WOMAC score was systematically over 15 points lower than the mean score on the Dutch Hip/Knee Replacement Expectation Surveys. It is our hypothesis that the way the Expectation WOMAC is adapted from the original WOMAC results in answers whereby the patients also considers the current status. The bias therefore might reflect a poor validity of the Expectation WOMAC rather than of the Dutch Hip/Knee Replacement Expectations Surveys. An alternative way to determine validity would be using the expectations of orthopaedic surgeons as reference, however this is questionable considering the differences that exist between the expectations of patients and those of orthopaedic surgeons [9].

One of the strengths of the current study is that participants were patients from a university as well as a general hospital. The study also had some limitations. First, not all patients were willing to participate in the study. The response rate of the first mailing was 83% in the hip replacement group and 64% in the knee replacement group. When responders and non-responders represent different patients groups, results of the study might not be generalizable to all total hip and knee replacement patients. Second, some questionnaires had to be excluded from the analyses due to missing values. This is inherent to this older patient group, who is often unfamiliar with filling in questionnaires. One way to avoid missing values is to let patients complete the questionnaires in the hospital. Although the patients are more likely to give socially desirable answers, the questionnaire can be checked for missing values when turned in.

Until now, research on patients' expectations before total hip and knee replacement is scarce and has to our knowledge only been performed in the United States, United Kingdom, Australia and Canada, with only one study comparing three different (English-language) countries [24]. Now that the Hip Replacement Expectations Survey and the Knee Replacement Expectations Survey are available in Dutch, it is possible to determine patients' expectations of total hip or knee replacement in another culture and healthcare setting. The surveys can be used to guide preoperative discussions about expectations between patients and physicians in the outpatient clinic, and in preoperative education classes aiming to change unrealistic expectations. Moreover, cross-cultural comparison is possible and an important future research topic.

In conclusion, the Dutch Hip Replacement Expectations Survey and the Dutch Knee Replacement Expectations Survey are both reliable instruments to determine patient expectations before total hip or knee replacement. With respect to concurrent validity it can be concluded that the correlation between both Surveys and the Expectation WOMAC was moderate confirming that the same construct was determined. However, a systematic bias was found; patients scored systematically lower on the Expectation WOMAC compared to the Dutch Expectations Surveys.