Swiss Orthopaedics Minimal Dataset: First Pilot Report of Reliability and Validity

Background. e Swiss Orthopaedics Minimal Dataset (SOMD) was launched seven years ago. It is a standardized, generic, and patient-reported outcome questionnaire, comprising ten items (location of disease, pain within the past four weeks, limitations at work/leisure/sleep/autonomy, subjective value of a body part, employment status, work disability (sick leave/pension), and household support). We conducted this study about the SOMD to report its reliability, validity, and clinical applicability.Methods. A retrospective observational cohort study was conducted. e test-retest study population (n� 60; lost to follow-up: n� 7 (12%)) was drawn from three retirement homes (in 2013), while the test study population (n� 14,180; excluded (e.g., duplicates): n� 1,990 (14%)) consisted of patients from a university hospital (in 2014–2017). In the test-retest study population, the same questionnaire was completed twice (at days 0 and 7). In the test study population, only the first questionnaire was included (to avoid duplicates). In a subgroup of the test study population (n� 302), only those patients who completed the SOMD andWestern Ontario andMcMaster Universities Osteoarthritis Index (WOMAC) of the hip within 14 days were considered (tominimize recall bias). Reliability (test-retest and internal consistency), criterion validity for the item of pain, and return rates were analyzed. Results. e test-retest study population (n� 53) showed very high test-retest reliability for all tested items of the SOMD (intraclass correlation coefficient� 0.96–1.00 (95% confidence interval 0.93–1.00), p< 0.001). e test study population (n� 12,190) revealed good internal consistency reliability for all ten items (Cronbach’s alpha� 0.80). e return rates of the SOMD were improvable (43% in 2016 and 31% in 2017). e subgroup of the test study population (n� 302) displayed a borderline acceptable criterion validity (correlation of the item of pain between SOMD and WOMAC hip: rho� 0.62, p< 0.001). Conclusion. is is the first report about the validation of the SOMD. A relatively high reliability (test-retest and internal consistency), borderline acceptable (criterion) validity for the item of pain, and improvable clinical implementation were observed. is analysis serves as the basis for a structured modification of the SOMD to improve its value.


Introduction
Health-related quality of life (HRQoL) describes the perceived well-being of individuals [1]. is can change over time (e.g., after treatment of disease). Accurate measurements (e.g., questionnaires) to detect this change are important not only for internal and external quality control but also for the evaluation of treatment success and costs. ere are many questionnaires to choose from, but the gold standard remains elusive. e Swiss Orthopaedics Minimal Dataset (SOMD) was introduced seven years ago, in 2013, as a measuring tool for HRQoL in order to assess the indications and results of all orthopaedic surgeries. It is a standardized, generic, and patient-reported outcome questionnaire. It consists of ten items: location of disease (only at one body part), pain within the past four weeks (score of 0-100), limitations at work/ leisure/sleep/autonomy (0-100 each), subjective value of a body part (0-100), employment status (training, employed, and retired), work disability ((A) sick leave and (B) pension) (0-100), and household support (0, <1, 1-2, 3-5, and >5 hours per day and a care assistant). e interpretation is based on a rating scale for each item without a single index value. It is free-of-charge, available in four languages (German, English, French, and Italian), accessible in electronic-and paper-based forms, and can be completed within five minutes. In our institution, it is used at first consultation and intermediate and final follow-ups. e psychometric properties (i.e., reliability, validity, and sensitivity) of the SOMD have not been adequately provided yet. So far, only content validity (i.e., expert opinion) has been established. e reliability, criterion validity, and clinical implementation (i.e., return rate) remain unknown. Reliability is defined as the consistency of an item. It can be assessed with a test-retest method providing an intraclass correlation coefficient (ICC) and internal consistency across different items within a test providing Cronbach's alpha (α). Criterion validity describes the correlation between an item of a new questionnaire with the same item of an established questionnaire providing Spearman's rho. A well-established, reliable, and valid questionnaire for comparison with the SOMD is the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), which was first described by Bellamy in a Master's thesis in 1982 [2][3][4][5][6][7]. It is one of the most commonly used questionnaires for the hip, available in around 100 languages, and consists of 24 questions about pain, stiffness, and daily activities. e item of pain refers to the last two days and different situations (walking, stairs, bed, sitting, and standing), which is particularly detail-oriented. Scoring is performed on a five-point (or level) Likert scale from none, over, and moderate to extreme.
We conducted the first study about the validation of the SOMD. e study hypothesis was that the SOMD is highly reliable, valid, and clinically applicable.

Methods
A retrospective observational cohort study was conducted. e cantonal ethics committee issued a waiver to allow this study with anonymous data without the need for informed consent (BASEC Request-Nr. 2018-00276). e test-retest study population (n � 60) was drawn from retirement homes in 2013. e loss to follow-up was acceptable (n � 7 (12%)). e retirement homes offered assisted but autonomous living. After thorough instruction of individuals, the German version of the SOMD was completed twice, initially at day 0 and again at day 7. Two items, employment status and work disability (sick leave/pension), were not evaluated in this elderly study population. e test study population (n � 14,180) consisted of all patients that filled out the questionnaire from a university hospital from April 2014 until December 2017. Only the first SOMD was considered. Duplicates and test questionnaires were excluded (n � 1,990 (14%)). Furthermore, a subgroup of the test study population was chosen, in which the WOMAC of the hip had also been completed within 14 days (to minimize recall bias) (n � 302).
Data were given as medians (interquartile range (IQR)). For the test-retest study population, the test-retest reliability (ICC (95% confidence interval (CI))) was calculated. For the test study population, the internal consistency (Cronbach's α of all ten items and item-rest correlations for each item), criterion validity (Spearman's correlation (rho) of the item of pain in the SOMD and WOMAC of the hip), and return rates were calculated. A scatterplot is provided for illustration. e significance level was set at p < 0.05. For a testretest reliability test, it was previously suggested that a sample size of ≥46 individuals would be needed to measure an ICC of 0.9 with a power of 0.8 at a significance level of 5% [8]. For internal consistency, it was also suggested that a sample size of ≥300 individuals would be sufficient [9]. Both of our study populations surpassed these numbers. Stata (IC13.1; StataCorp, College Station, Texas, United States of America) was used for analysis.

Advances in Orthopedics
In the subgroup of the test study population (n � 302), the criterion validity for pain was borderline acceptable (correlation of the item of pain between the SOMD and WOMAC of the hip: rho � 0.62) (Figure 1).

Discussion
is is the first study reporting on the validation of the SOMD. Our findings show that the SOMD is highly reliable (i.e., very high test-retest reliability and internal consistency) but has borderline acceptable validity (i.e., correlation of the item of pain in the SOMD and WOMAC of the hip) and improvable clinical implementation (return rates).
Overall, the SOMD appears as a valid measurement tool for the quantification of a joint-specific HRQoL. However, several modifications are necessary for revised future versions. In general, improvement of reliability can be achieved with item homogeneity (i.e., identical minimum and maximum of different items) and item selectivity (i.e., substantial differences in minimum and maximum of a single item). For improved internal and external quality control of surgical indications and outcomes, economic analyses, and research, a highly validated questionnaire, such as the EuroQol 5-Items (EQ-5D) [10], should be integrated into the SOMD to allow calculation of a single index value and quality-adjusted life years (QALYs). e EQ-5D is a standardized, generic, self-reported, fast, generalized, and validated questionnaire. It is available in 120 languages, highly reliable and valid, and commonly used to measure HRQoL. e first version with three levels was designed in 1990 and its current form with five levels (no, slight, moderate, severe, and extreme) has been used since 2009. It allows the calculation of 3,125 (5 5 ) different HRQoL states. Using crosswalk links, a single index value and QALYs (0 � dead to 1 � perfect health) can be calculated [11]. Using population value sets, this permits utility analyses. Furthermore, questions from the validated Core Outcome Measures Index (COMI) questionnaire [12] could be added to quantify the joint-specific HRQoL and level of pain. Additionally, survey reports can vary according to the educational level [13], which should be quantified by an additional question. We have proposed a revised version of the SOMD, which was drafted by a subcommittee of the Swiss Orthopaedics Panel of Quality and Methodology together with the general manager of the Swiss Implant Register (SIRIS) [14] and has been approved by Swiss Orthopaedics board of directors [15].

Conclusion
In conclusion, this is the first report about the validation of the SOMD. A relatively high reliability (test-retest and internal consistency), borderline acceptable (criterion) validity for the item of pain and improvable clinical implementation was observed. e analysis serves as the basis for a structured modification of an improved version of the SOMD.

Data Availability
All data analyzed during this study are included in this article.