Introduction

There are numerous medical evaluation tools for a variety of diseases, but in most cases, such tools are designed to reflect the viewpoint of health-care providers rather than that of patients. Evaluations focusing on hip-joint disease, such as the Harris Hip Score [1] and Merle d’ Aubigné and Postel score [2], are commonly used. In Japan, the criteria for hip-joint function proposed by the Japanese Orthopaedic Association (JOA Hip Score) [3] are also widely used. However, it has been reported that the JOA Hip Score is a reliable system only for patients with osteoarthritis of the hip that is treated conservatively [4]. Moreover, such evaluations by health-care providers can be biased and affected by intraobserver and interobserver differences, producing disease-state assessment results that differ significantly from patients’ perceived severity of their disease. In gauging the effect of medical therapies in addition to clinical results, it is necessary to assess patients’ quality of life (QOL). Thus, in recent years, evaluation criteria that can serve as patient-focused outcome indices have been attracting increasing attention. Health-related QOL criteria represent patient-based outcome index criteria. The Medical Outcomes Study Short-Form 36-Item Health Survey (SF-36) [5] offers comprehensive criteria, and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) [6] and Oxford Hip Score (OHS) [7] offer criteria specific to hip-joint disease. However, because these tools’ criteria do not take into account movements specific to the Asian lifestyle [8, 9], such as rising from the floor or squatting to use a Japanese-style toilet, they cannot be said to accurately evaluate the QOL of all patients [10–13].

In recent years, the Japanese Orthopaedic Association has been working on a plan to establish patient-based, multifaceted, and science-based evaluation criteria for a variety of diseases. As a part of that effort, the Japanese Hip Society has been asked to prepare criteria specific to hip-joint disease that also incorporate movements common in Japanese daily life. In response to this request, the Japanese Hip Society established the Subcommittee on Hip Disease Evaluation of the Clinical Outcome Committee of the Japanese Orthopaedic Association to draw up patient-based criteria specific to hip-joint disease with this consideration in mind. In this article, we describe the process of criteria creation, and consider the reliability and appropriateness of these final evaluation criteria. The complete hip-disease evaluation questionnaire and guide for mental-health-care providers are shown in the two appendixes to this article.

Materials and methods

To prepare these patient-based evaluation criteria, we first interviewed patients during office visits occurring between July and September 2006 about any difficulties related to their hips. This interview was conducted by physicians or nurses with open question methods at eight university hospitals and six municipal hospitals in the whole of Japan. We analyzed and pooled patients’ comments for use in preparing a questionnaire. Furthermore, we considered some preexisting QOL criteria and some evaluation criteria, such as those from the SF-36, and included some items from such sources in the questionnaire item pool. We then compiled a self-administered questionnaire for the purpose of preparing criteria and used it in a survey conducted at 12 university hospitals and 5 municipal hospitals throughout Japan from December 2007 to August 2008. Permission to conduct the survey was obtained from the ethics committee of each institution, and all patients consented to participate after being given complete information about this survey.

In order to select questionnaire question items and prepare evaluation criteria, the obtained data were subjected to factor analysis. In the factor analysis, we first identified the number of factors by principal component analysis [14], and then conducted rotation with obtained number using the Quartimin method [15]. To verify the reliability of the completed questionnaire, we calculated the Cronbach’s α [16] for each factor using the items applied. Statistical analysis was performed using SAS (version 9.1; SAS Institute Inc., Cary, NC, USA).

Results

A total of 464 comments were obtained from about 100 patients during oral questioning from the interview. Overlapping opinions and those with similar content were grouped together, and a pool of 84 items was finalized. In addition, we added items based on previously devised evaluation criteria.

We then created a questionnaire comprising 58 items for the survey (Table 1). Excluding the category of “pain,” which was an unnumbered item, five response categories were adopted (“strongly agree,” “agree,” “uncertain,” “disagree,” “strongly disagree”) for each item on the answer sheet. For some of the items, the questions were asked twice, one time for the left hip joint and another time for the right hip joint. To assess pain, a visual analog scale was also adopted. In this survey, we collected information for 501 cases. Regarding the replies to the questions concerning laterality of hip-joint involvement, we proposed a solution that was based on the following criteria:

Table 1 Questions used in a survey
Full size table

  • Criterion 1: The more problematic hip joint is counted.

  • Criterion 2: When problems are present bilaterally in the hips, the more painful hip joint is counted.

  • Criterion 3: In cases in which no decision can be made on the basis of criteria 1 and 2, the more severely affected side is counted for each item.

Patients who were diagnosed “no problem” with respect to the bilateral joint in criterion 1 were excluded from the analysis.

With the exception of the visual analog scale for pain, each item was given 0–4 points in increasing order, starting from “strongly agree.” With regard to the visual analog scale for pain, the length from the left side of the scale recorded by the respondent was divided into five stages and given 0–4 points for increasing levels of pain, so as to be consistent with the form of the replies to the other questions.

The scores obtained for these 58 items that could be rounded off were considered items for analysis, and the persons who replied to all of these items were considered targets for analysis. These amounted to 402 cases (Table 2).

Table 2 Summary demographic data for questionnaire respondents
Full size table

To investigate the number of categories, we performed principal component analysis. There were six principal components with eigenvalues exceeding 1, and the cumulative percentage of the six principal components was 72.2% (Table 3). A screeplot was prepared from these results, and the number of categories was decided to be three (Fig. 1).

Table 3 Results of principal component analysis
Full size table
Fig. 1
figure 1

Screeplot. Prin comp principal component

Full size image

Factor analysis was conducted using the Quartimin method with three categories. Although a few items with low commonality were seen in the prior communality estimates (minimum value, 0.404), we performed the analysis using all of the items, and Table 4 shows the factor pattern and the factor structure of main items that strongly associated with each item.

Table 4 Results of category analysis
Full size table

In view of the results of analysis of these categories, we selected question items, in consultation with clinicians, regarding each factor and considered the naming of the categories. In this way, as shown in Table 5, items were adopted for each category, and category names of “movement,” “mental,” and “pain” were selected. The final communality estimates of the applied items showed a maximum value of 0.825 and minimum value of 0.584. Using the items applied to the respective categories, Cronbach’s α values were calculated (Table 6). In addition, the correlation coefficient between the categories of “movement” and “mental” was calculated to be 0.66; that between “movement” and “pain,” 0.57; and that between “mental” and “pain,” 0.69.

Table 5 Items adopted as evaluation criteria
Full size table
Table 6 Reliability of each category
Full size table

Discussion

When preparing evaluation criteria, an important first step is the creation of an item pool that will form the basis of questions from which the criteria will be decided on. Because in the present criteria, a patient-based evaluation was the main element, we questioned patients face-to-face, focusing on their own hip joint and related difficulties in daily life and then creating an item pool from patients’ opinions.

Because all of the participants in our study are Japanese, we collected numerous opinions related to deep flexion and rotation of the hip joint associated with motions common in daily Japanese life, such as rising from the floor and using a Japanese-style toilet, and these were reflected in the final evaluation criteria. These items represent areas that could not be assessed in the WOMAC [10, 11, 13] and OHS [12], and thus have the important feature of including Asian lifestyle patterns. Notably, almost all of the questions finally adopted in the criteria were obtained by the initial oral questioning. From the viewpoint of patient-centered evaluation, the completed criteria can thus be considered to be fully appropriate.

The questionnaire used for the survey for the preparation of criteria consisted of 58 questions. Actually, we would have preferred to have used all of the items in the item pool as questions, but taking information bias into account, we decided that some of the items should not be adopted in the survey. When adopting items for this purpose, we placed special weight on the frequency with which items were raised during oral questioning, with items raised by multiple patients adopted whenever possible.

In the survey, we collected information for 501 cases, with the target participants amounting to 402 of these. Almost all of the 99 dropouts had inadequate replies; this occurred most frequently in persons of advancing age. In preparing the questionnaire, we used the large type character for easy reading and illustrated the sample replies, in addition to the number of questions, in order to take into consideration information bias. However, the burden of completing the survey might have been considerable in the elderly. The completed questionnaire was thus shortened to only 21 items and should be employed with care with elderly patients. The “seiza” is one of the common postures in Japan. However, this item was not included in the final 21 items. In the factor analysis results, seiza was not strongly associated as compared to other items. Squatting for a Japanese-style toilet requires more range of motion of the hip joint than “seiza” [8, 9, 13]. Therefore, the items of getting up from the floor and using a Japanese-style toilet will be available for including the seiza item.

In the factor analysis, the number of categories adopted was three, but we similarly investigated the scenario of adopting four or five categories. In each of these scenarios, the categories used here were expressed, whereas in the case of the remaining categories, we could not supply an appropriate interpretation and so decided against their adoption. In the selection of questions to make up the evaluation criteria, items 16 and 17 in the “movement” category were consolidated. For this reason, when calculating the Cronbach’s α coefficients and the correlation coefficients between each pair of categories, we adopted the lower of the scores for items 16 and 17.

Cronbach’s α coefficients reflect the reliability of the evaluation criteria according to the adopted questions, and a Cronbach’s α ≥ 0.70 was considered to indicate that a scale had internal-consistency reliability [17, 18]. In each case, the Cronbach’s α values were high, confirming sufficient reliability for these criteria. A self-administered patient-based questionnaire for hip-joint disease, the Japanese Hip-Disease Evaluation Questionnaire (JHEQ), was established through this process (Appendix 1). A guide for mental-health-care providers using the JHEQ was also developed (Appendix 2).

Because the JHEQ takes into account facets of the Asian lifestyle, it may help improve the assessment of QOL for Asian patients. At the same time, the JHEQ can also be useful in Western populations for evaluating patients who frequently engage in deep flexion of the hip joint. The JHEQ also makes possible preoperative and postoperative evaluation of factors that formerly were not be assessed. For example, after total hip arthroplasty it will now be possible to assess mental aspects such as anxiety associated with clinical events such as dislocation and reimplantation. Similarly, these criteria may facilitate investigations into differences in patient-based evaluations in those undergoing joint-preserving surgery with osteotomy and arthroscopy as compared with total hip arthroplasty. Issues still to be resolved include the fact that no comparison has yet been performed with evaluation criteria already in use. Additional studies are required to compare the JHEQ with the JOA Hip Score, the Harris Hip Score, the SF-36, and the WOMAC.

The Japanese Orthopedic Association Hip Disease Evaluation Questionnaire and guidelines, which are provided in the Appendix, were written originally in Japanese. After translation into English by qualified specialists, they were then back-translated into Japanese to confirm the accuracy of the English translation.