Translation and cross‑cultural adaptation
The translation and cultural adaptation process was performed according to the guidelines for the cross-cultural adaptation of self-report measures by Beaton et al. [24]. The final version was validated as stated in the Consensus-Based Standards for the Selection of Health Status Measurement Instruments (COSMIN) guidelines [25, 26].
The original English version was independently translated into Italian by one orthopedic surgeon with a good knowledge of English and an English teacher without a medical background; both were native Italian speakers. The initial Italian version of the questionnaire was chosen after a consensus meeting where discrepancies between the two initial Italian versions were debated and corrected. At the back translation stage, two native English speakers with fluent Italian, who were not informed of the study’s purpose and were blinded to the original English version, were asked to translate the chosen Italian version into English. An expert committee composed of the four translators, another orthopedic surgeon, and a psychometrist compared the original English versions of the KSTARS and ACL-RSI scale with the back-translated versions of the tests. A prefinal version was created and tested on 10 Italian-speaking volunteers to determine their comprehension of the Italian ACL-RSI version (ACL-RSI-It). None of the volunteers reported interpretation problems or redundancy. The authors then approved the final versions (see Additional files 1 and 2).
Population
Institutional Review Board approval was granted for this study. All participants and their parents gave valid consent to participate. A retrospective analysis of prospectively collected data for consecutive patients who underwent ACLR at the “Kilk Kilgour” Sports Traumatology Center, Sant’Andrea University Hospital of Rome, or the orthopedic surgery department of Santa Croce e Carle of Cuneo between January 1st, 2018 and January 1st, 2020 was conducted. The inclusion criteria were: (1) primary ACLR; (2) ≥ 18 years old; and (3) ≥ 6 months of follow-up. Participants were excluded if they:(1) underwent a bilateral ACLR, (2) had a history of previous knee injuries, (3) suffered from multi-ligament knee injuries or a concomitant fracture other than Segond’s, (4) had radiographic findings of knee osteoarthritis, or (5) did not have a good understanding of both the Italian and English languages.
In order to minimize the bias related to the multicenter design of this study, the patients all performed the same postoperative rehabilitation protocol. This included a brace locked in extension for the first 4 weeks, with weight bearing allowed with crutches from the first postoperative day; recovery of knee range of motion (ROM) starting at 2 weeks, with the brace temporarily removed during exercises; at 4 weeks, patients were allowed to walk without crutches and the brace; progressively increased ROM and muscle strengthening until the third month; from the third to the sixth month, patients underwent a progressive muscle strengthening program, and sport, specific exercises were performed. A return to sporting activities was allowed at 6 postoperative months.
After a preliminary telephone call explaining the study’s purpose, the patients were asked to fill an anonymized online form created for this purpose and administered via Google Forms (Google LLC, Mountain View, CA, USA). The form consisted of a first section asking for demographic data and a second section that included the KOOS, the Lysholm, the IKDC-SKF, and the ACL-RSI-It [27,28,29]. After 1 week, the attendees were asked to repeat the ACL-RSI-It to investigate the test–retest reliability.
Patient‑reported outcome measures
The original ACL-RSI scale is a self-administered questionnaire that consists of 12 items evaluated with a 10-cm visual analog scale (VAS) from 0 to 100 in 10-point increments. It is subdivided into three aspects that cover (1) emotions (five questions), (2) confidence in one’s performance (five questions), and (3) risk appraisal (two questions) [20]. Later, a modified ACL-RSI was developed by the same team; in this, the VAS scales with 11-point Likert scales in boxes with 10-point increments from 0 to 100 [18]. The total score is calculated by adding the values of the 12 items and then taking the percentage of the amount. High scores are indicative of a positive psychological response.
The IKDC-SKF [29], the KOOS [28], and the Lysholm score [27] are scales that are widely utilized in clinical practice; these were the references for the current study in their validated Italian versions.
K-STARTS score
The patient can perform the K-STARTS test if no deficits are present above 40% relative to the contralateral limb at a prior executed isokinetic test. The K-STARTS test provides a psychological status and physical performance assessment and is composed of four sections.
The first section of the K-STARTS appraises psychological readiness for RTS and is determined by the outcome of the ACL-RSI questionnaire. Three K-STARTS points are given for ACL-RSI scores of 76% or higher, 2 points for scores between 64 and 75%, 1 point for scores between 56 and 63%, and 0 points for scores less than 55%.
The second section of the K-STARTS evaluates neuromuscular control, as determined by the outcome of the Qualitative Analysis of Single-Leg Squat (QASLS). This tool provides an analysis of movement occurring in the arms, trunk, pelvis, thighs, knees, and feet during single-leg loading tasks [30]. The QASLS score ranges between 0 (best) and 10 (worst), with points increasing for inappropriate movement strategies. A patient with a QASLS of 0, 1, 2, or ≥ 3 scores, respectively, 3, 2, 1, and 0 points in K-STARTS. Furthermore, 3 points are deducted if a patient is judged to have a dynamic valgus of the limb during the single-leg loading task.
The third section of the K-STARTS is an estimation of the limb symmetry index. Four categories of hop test (single, triple, side, and crossover) are carried out, and the percentage deficit of the distance hopped on the involved leg compared to that hopped on the uninvolved contralateral leg is computed [23, 31, 32]. For each of the hop tests, a limb symmetry index of 90% or more corresponds to 3 K-STARTS points, between 80 and 89% corresponds to 2 points, and 79% or less to 1 point; if pain occurs during the test, no points are attributed.
The fourth section evaluates the ability to change direction using the Modified Illinois Change of Direction Test (MICODT) [33]. An average MICODT time of ≤ 12.5 s scores 3 K-STARTS points. If the time ranges from 12.51 to 13.5 s, it scores 2 points, and if the time is > 13.5 s, it scores 1 point. Moreover, if pain prevents the test, no points are given.
The total score is calculated as a percentage. If the final score is less than 50 points, a return to training is discouraged. If the score ranges between and 50 and 65, noncontact nonpivot training is allowed. For a score ranging between 65 and 80, noncontact pivot training is allowed. If the score is higher than 80, the athlete is allowed to return to contact training activity.
Statistical analysis
Statistical analysis was performed with IBM SPSS version 25.0 for macOS (IBM, Armonk, NY, USA). The significance threshold was set at p < 0.05.
Descriptive data were presented in the form of mean ± standard deviation and percentage.
Floor or ceiling effects were considered present if the proportion of patients (%) who had the minimum (0) or maximum (100) score on the scale was more than 15% [34].
Internal consistency was estimated using Cronbach’s α coefficient. Homogeneity between items within a questionnaire was considered excellent for α > 0.90 [34].
The intraclass correlation coefficient (ICC) with a 95% confidence interval was calculated to assess the instrument’s stability over time. To determine test–retest reliability, we used Spearman’s correlation coefficient and the ICC q; reproducibility was considered either excellent (q > 0.75), good (0.75 < q < 0.40), or poor (q < 0.40) [35]. A value of greater than 0.40 was considered acceptable.
The construct validity was assessed by comparing the ACL-RSI-It with the subjective IKDC, the different components of the KOOS, and the Lysholm score using Spearman’s correlation coefficient; the correlation was considered strong (r ≥ 0.5), medium (0.5 > r > 0.3), or small (r < 0.3).
Discriminant validity was tested by a Mann–Whitney U test between patients who returned to their previous level in sport and those who could not return to the same level.