Objective To develop a condition-specific patient-reported outcome measure, the Functional Assessment Scale for Acute Hamstring Injuries (FASH), de novo in three languages, following distinct and rigorous methodology for content generation, analysis and validation and to assess its psychometric properties.
Background To our knowledge, there is no patient-reported functional scale specific for acute hamstring injuries.
Methods The development of the scale followed specific guidelines, as well as de novo construction in three languages (Greek, English and German). Item generation was accomplished by selecting three different sources of items: literature review, focus group and key informant interviews. Content analysis was conducted by an expert committee. The 21 items selected as appropriate were tested through a structured content analytic method and item-content validity coefficient, and 10 were retained for the FASH. The validation and assessment of its psychometric properties followed theConsensus-based Standards for the selection of health status Measurement Instruments (COSMIN) recommendations to ensure quality, in a convenience sample of 140 participants.
Results The face validity was adequate and tested by expert committees, authors and participants. Content validity was characterised as well addressed and conducted independently by experts and through specific content validation procedures. The dimensionality analysis indicated a one-factor solution explaining the 95.8% of total variance. Known group validity was demonstrated by significant differences between patients and controls (p<0.001). The FASH exhibited very good test–retest reliability (intraclass correlation coefficient=0.9, p<0.001), internal consistency (α=0.98) and responsiveness (3.81 and 5.23 using baseline and pooled SD, respectively; standardised response mean (SRD)=4.68).
Conclusion This study provides initial evidence for psychometric properties of the first scale assessing hamstring injuries.
- Exercise testing
Statistics from Altmetric.com
Hamstring muscle injuries (HMIs) are among the most common soft tissue injuries in athletes.1–7 They are prevalent in soccer,2 ,8–16 Australian football,6 ,17–20 rugby union,1 ,21–23 American football,24 ,25 track and field athletics,3 ,4 ,26–31 and in other sports.5 ,7 ,32 ,33 These sports involve high-speed running, acceleration, deceleration, jumping and rapid change of direction.1–7 The main mechanisms of injury contain eccentric contraction at high velocity and slow stretching at outer range of motion.34
On the basis of the Munich consensus statement35 and other reports,4 ,36–39 acute HMIs are defined as having a non-contact-induced, stretch-induced or contraction-induced mechanism, with or without ‘macroscopic evidence’35 of structural muscle changes or functional issues. Moreover, HMIs are characterised by acute onset, well-defined localised posterior thigh pain during training or competition.4 ,35–39
HMIs result in significant time-loss from training and competition and have a high rate of recurrence.1 ,15 ,24 ,34 ,37 ,40–45 The first month after return to sport is the highest risk period for recurrence1 and the risk remains increased for about 1 year.20 ,42 The reinjury is usually more severe and results in a longer absence from play than the original injury.2
There are no validated instruments to measure symptoms and their effect on function and sports ability in patients with HMI. The two HMI-specific questionnaires are not pathology related,46 or not validated47 in patients with acute HMI.
An instrument is required to help evaluate the effectiveness of treatment, quantify a patient's clinical severity, compare different patient populations and facilitate comparative research.48 ,49 Therefore, the aim of this study was to develop and validate a patient-rated questionnaire50 that quantitatively measures the severity and the impact of symptoms on function and sports ability in patients with acute HMI.
Materials and methods
The development of a questionnaire followed the Consensus-based Standards for the selection of health status Measurement Instruments (COSMIN) recommendations.51
Development of the Functional Assessment Scale for Acute Hamstring Injuries questionnaire
The main objective in constructing the Functional Assessment Scale for Acute Hamstring Injuries (FASH) questionnaire was the de novo development for use in three languages: Greek, English and German. The basic structure of questionnaires developed in one language cannot be changed and requires adaptation for use in another culture and language.52 Furthermore, there are interpretation problems in the translation of terms due to cultural differences.52
The questionnaire was developed following set steps53 ,54 (see online supplementary figure S1). The study was approved by the University of Thessaly ethics committee and participants gave signed informed consent.
Step 1: Construct determination
The developers defined the proposed construct and content of the scale. As scales designed to measure one clinical presentation may also be sensitive to other clinical presentations,53 we decided to develop a scale that was not a diagnostic tool. The questionnaire proposed to assess the symptoms, severity and functional disability in patients with acute HMI.
Step 2: Item generation
A large pool of items, which would reflect the construct of interest and would be the basis of a unidimensional scale, was generated. Careful attention was given to the questions as no manipulation, statistical or otherwise, can compensate for poorly chosen questions.54
Three different sources of potential questions were chosen: literature review, focus group and key informant interviews.54 A literature review identified appropriate items from clinical and observational studies and related questionnaires. A focus group, 10 elite athletes (age 22–28 years) who had previously suffered from HMI, was interviewed about symptoms they perceived as important and their functional limitations. An expert committee consisting of the investigators, a sport physician, a physiotherapist and an athletic trainer, all with several years of experience with HMI, generated specific items from the themes of the focus group meeting. The focus group reviewed the items for their relevance, completeness and comprehension. Two physicians, two physical therapists and two athletic trainers were interviewed as key expert informants about their assessment and management of HMI.
Step 3: Content analysis
Three experts: a sport physician, a physiotherapist and an athletic trainer with years of experience in sport injuries, screened and evaluated the pool generated, independently of each other. They evaluated the items based on their content, edited items they perceived as redundant and/or incomprehensible and eliminated items considered irrelevant, potentially harmful or synonymous. The experts, upon completion of their independent work, convened to discuss the items until agreement was secured and face validity was ensured. From the 35 items that were initially gathered, consensus was reached for 21 items.
Step 4: Item-content relevance analysis
The 21 items were tested through a structured content analytic method.55 The items were distributed to 10 judges (2 academics, 2 sport physicians, 2 physiotherapists, 2 athletic trainers and 2 athletes; all holding higher degrees in relevant areas) who were not involved in the construction process of the scale. The judges matched each of the 21 items to a five-point rating scale (1=poor, 2=fair, 3=good, 4=very good, 5=excellent). This quantified item ratings allowing evaluation of judgments using quantitative statistical procedures.
The judges’ ratings were evaluated based on the content validation procedure of Aiken's56 item-content validity coefficient (V). The V statistic provides statistical significance of judges’ ratings about an item's content-match with its construct. The V values range from 0 to 1 (1=perfect agreement). The V values are then compared against a right-tailed binominal probability table provided by Aiken.56
From the 21 items analysed, 11 items were excluded with V values lower than 0.70. Ten items with V values ranging from 0.90 to 1.0 (p<0.01) were retained for the next step.
Step 5: Item scaling
A visual analogue/numeric scale is more accurate and sensitive than categorical verbal scales57 ,58 and has been used in the development of questionnaire.59 ,60 Seven questions used a 0–10 numerical rating, the remaining three questions that addressed sport activity and demanding functional tests used a categorical rating system on an incremental range of values.60–62 At this stage the questionnaire included three items from Victorian Institute of Sport Assessment (VISA) questionnaires.59 ,60 ,62 These items are psychometrically sound and development of new items to assess a clinical presentation is unnecessary.54
The highest score for the FASH is 100, representing normal/high level of physical function and ability. The theoretical minimum for the FASH is 0 points, representing complete disability. For each question, 10 represents high/normal level of physical ability and 0 represents complete disability or inability to perform a functional task.
Step 6: De novo development for use in three languages
The development of the questionnaire for use in Greek-speaking, English-speaking and German-speaking patients with acute HMI was accomplished in two steps. The composition of the three language versions of the questionnaire was conducted in a consensus meeting of the questionnaire developers and a trilingual consultant. Several steps were taken to ensure and verify the equivalence of the versions using the guidelines for cross-cultural adaptation63 (see online supplementary table S1). The purpose of this procedure was the adequate linguistic translation, as well as the cultural adaptation to maintain the content validity of the scale across different cultures. The source language was the Greek language and the target languages were English and German, respectively.
Step 7: Pretesting
The prefinal version of FASH was administered to 20 Greek-speaking, 20 English-speaking and 20 German-speaking physically active asymptomatic participants (age 20–28 years). Following the completion of the questionnaire, each individual was interviewed about their comprehension of the items and their chosen response. The testing revealed that a number of participants could not understand the sport and exercise terminology that described some functional tests. The expert committee added a description and/or image of the procedure for these items.
Step 8: Validation
The final FASH questionnaire was assessed for validity and reliability only in Greek-speaking patients.
The sample of 140 participants consisted of four groups; a healthy group (young active individuals), a group at risk of HMI (track and field athletes), a group with lower limb injuries other than HMI (ie, anterior cruciate ligament injury, meniscal tears, ankle strains) and a group with acute HMI (table 1). A qualified sports medicine physician examined all patients with HMI and ultrasonography was used to classify the muscle lesions.4
The active individuals were from two sport clubs in Athens and the track and field athletes were from clubs in Thessaloniki. Athletes in the injured groups were recruited from respective sports medicine centres. Individuals from different cities were used to minimise bias due to cultural or demographic factors. Participants required 10 min to complete the questionnaire and the response rate was 100%.
Inclusion and exclusion criteria
Participants were active in sport and older than 18 years. The exclusion criteria were pregnancy and referred spinal symptoms. For the healthy and at risk groups, further exclusion criteria were pain and functional deficits during physical activity. The HMI group gave a detailed history followed by a clinical examination and ultrasonography performed within 24–36 h from injury. Those who met the inclusion criteria were recruited (table 2).
The FASH was administered to all participants (N=140) and completed twice within 48–60 h. This interval between test–retest was chosen as the time period between repeated administrations should be short enough to ensure that clinical change has not occurred64 and clinical improvement in HMI is observed after 48 h.4 The minimal treatment (PRICE: protection, rest, ice, compression, elevation) was offered to participants until the completion of the second questionnaire. The measurement and psychometric properties of the FASH questionnaire are presented in table 3.
All statistical analyses were performed using SPSS V.17.0. The level of significance was set at p<0.05.
Concurrent validity was assessed with Spearman r between the scores obtained from FASH and VISA-H.59 Factorial validity was assessed using exploratory principal component analysis and eigenvalues over one were extracted.71 Known group validity and group differences were calculated using the Kruskal-Wallis test, with post-hoc comparisons using the Mann-Whitney test with Bonferroni correction for multiple testing (α level set at 0.0083) as resulted from the formula (k(k–1)/2) where k is the number of groups. Internal consistency of the total score was calculated using Cronbach's α.64
Test–retest reliability was defined by using two way random-effects intraclass correlation coefficient, type agreement (ICC2,1) because systematic differences are considered to be part of the measurement error64 and Spearman r. A Bland-Altman plot was constructed to indicate absolute agreement for test–retest measurements including a scatter plot of differences between applications, with 95% limits of agreement.64 In addition, the SE of measurement (SEM; SEM=SD×√(1−test–retest reliability coefficient)) and minimal detectable change (MDC95; MDC=1.96×√2×SEM) were calculated.64 ,71 ,72
The Wilcoxon test, using scores from first and third application of the FASH was conducted to assess responsiveness. Also, effect size (ES) by using both baseline and pooled SD and standardised response mean were calculated and interpreted according to published recommendations73 (values of 0.20, 0.50, and 0.80 or greater represent small, moderate and large responsiveness, respectively). Descriptive statistics were used for the characteristics of the participants and the calculation of ceiling and floor effects.
The questionnaire was completed with one of the investigators at all test occasions to standardise the procedure (table 4).
The face validity was tested by expert committees, authors and participants. Content validity was characterised and conducted independently by experts. Concurrent validity was demonstrated by a high level of correlation between the first FASH and VISA-H scores of participants (r=0.856; p<0.01).
In the dimensionality analysis, the Kaiser-Meyer-Olkin measure, with or without the other injuries group, verified the sampling adequacy (KMO=0.934 and KMO=0.955, respectively) and all KMO values for individual items were >0.895 and >0.921, respectively, which is well above the acceptable limit of 0.5.74 Bartlett's sphericity test (x2(45)=2750.561, x2(45)=3432.041, p<0.001, respectively) indicated that correlations between items were sufficiently large for principal component analysis. The analysis (without other injuries group) and the examination of the screeplot indicated a one-factor solution with eigenvalues over 1 explained 95.8% of the total variance (table 5). The communalities after factor extraction were 0.839–0.98.
Significant differences between groups were found using the Kruskal-Wallis test (p<0.001) in known group validity testing. The Mann-Whitney test did not reveal significant difference for the total score between the healthy group (Mdn=99.0) and the group at risk (Mdn=97.5; U=447.0, p=0.064, r=0.22), but showed significantly higher scores for the healthy group and the group at risk compared with patients with HMI (Mdn=19.0) group (U=0.0, p<0.001, r=0.854; U=0.0, p<0.001, r=0.863, respectively) and other injuries (Mdn=76.0) group (U=7.0, p<0.001, r=0.85; U=15.5, p<0.001, r=0.79, respectively). Furthermore, the other injuries scored significantly higher compared with hamstring injury group (U=15.0, p<0.001, r=0.83).
The responsiveness evaluation revealed statistically significant changes of the FASH scores from baseline (Mdn=19.0) to discharge (Mdn=92) for the patients with HMI (Z=5.5, p<0.001, table 6).
A Bland-Altman plot (figure 1) showed that the differences between two assessments were plotted around the zero line and were inside the limits of agreement. Additionally, the zero line was within the 95% CI of the mean difference between the second and first assessments, indicating no systematic bias.
No participant achieved the minimum or maximum possible score of FASH, while for the individual items of the scale, no item received a minimum or maximum score by more than 75% of the participants.
The FASH is a self-reported questionnaire developed for patients with acute HMI. The questionnaire was developed de novo in three languages and this reciprocal process allows for greater creativity and ensures fine-tuning and closer alignment of the produced versions.52
A high Cronbach α indicated excellent internal consistency. Furthermore, concurrent validity was demonstrated and known group validity was clarified by the expected differences among groups tested. Ceiling and floor effects were not found suggesting that FASH is an applicable tool for patients within the full severity range of HMI. The factor structure of the FASH presented a unique solution explaining the 95.8% of total variance and following the principles of parsimony and interpretability. The FASH displayed an excellent reliability for all groups. The Bland-Altman plot revealed that the vast majority of scores were contained within the 95% limits of agreement. The SEM values for all groups were low—a small portion of the 0–100 point scale.
The FASH had a large ES revealing that it is responsive to changes in clinical course and can potentially be used to monitor changes in symptoms over time and for specific interventions. The responsiveness analysis further corroborated the absence of ceiling and floor effects.
The sample configuration was conducted with recruitment from different cities of Greece to minimise bias due to cultural, semantic and demographic factors. This allows the findings to be generalised with caution, due to the small sample size and the inclusion of only young individuals.
The FASH is not a ‘diagnostic tool’ nor does it play a role in decision-making about treatment options. Clinical assessment and diagnosis are essential before administration of the FASH to patients with HMI.
The absence of a ‘gold standard’ for comparison should be considered as a limitation, however, the VISA-H,59 although not formally validated, was the only available instrument. The sample consisted of track and field athletes and did not included athletes from other sports (football/rugby) with high prevalence rates of HMI, a fact that limits its generalisability. We acknowledge that validation of the questionnaire in English-speaking and German-speaking patients, across a broader age range and in athletes of different sports are essential steps in the development of this tool. The FASH may allow comparable research data during assessment and management of patients with HMI in the international literature. Nevertheless, further validation in clinical settings is required to evaluate its contribution in making return to sport decisions.
What this study adds?
Provides an acute hamstring injuries questionnaire in three languages.
Our development and validation processes followed rigorous clearly described guidelines (eg, Consensus-based Standards for the selection of health status Measurement Instruments).
How might it impact on clinical practice?
The Functional Assessment Scale for Acute Hamstring Injuries (FASH) questionnaire can be used as an outcome measure to assess and manage patients with hamstring injuries.
If it proves successful, the FASH may aid specific rehabilitation programmes, direct the period for return to sport and evaluate the result (failure or success) of treatment after hamstring injuries.
The authors thank all the translators, the members of the expert committees and the participants for their valuable contributions to the present study.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
Contributors NM specified the idea and construct of the questionnaire and together with DC performed the data collection. VK planned the study design, wrote the manuscript and performed the statistical analyses. NM, VK, NP, DP, TN, PM and HL participated in expert committees and in the development process. GG contributed to the statistical analyses.
Competing interests None.
Patient consent Obtained.
Ethics approval The study was approved by the Research Ethics Committee of the University of Thessaly.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.