Psychometric Validation of the Hemophilia Functional Ability Scoring Tool (Hemo-FAST)

Abstract

Background The Hemophilia Functional Ability Scoring Tool (Hemo-FAST), consisting of a patient-reported outcome (PRO) part and a clinician-reported outcome (ClinRO) part, was developed as a rapid and effective tool to assess functional mobility in clinical practice. This study (NCT04731701) aimed to validate the psychometric properties of Hemo-FAST for assessment of joint health in people with haemophilia (PwH).

Methods PwH A or B aged ≥18 years completed questionnaires including the PRO part of Hemo-FAST and the short-form 36 health survey (SF-36) during one study visit. Clinicians completed the Haemophilia Joint Health Score (HJHS) and the ClinRO part of Hemo-FAST at the same visit. Validation was performed using reliability, construct validity, and structure validity assessments.

Results The study enrolled 180 PwH A or B from 14 centres across France. Estimated time to complete the PRO part was mean (standard deviation) 4.6 (5.4) minutes. PRO items showed good test–retest reliability (intraclass correlation coefficient value ≥0.70). Inter-rater values were >0.70 for 7/9 ClinRO items, indicating good reliability. All items (15 PRO; 9 ClinRO) had high internal consistency (Cronbach's coefficient alpha: 0.97). Hemo-FAST demonstrated convergent construct validity with HJHS and the SF-36 physical component and discriminant construct validity with the SF-36 mental health component. Hemo-FAST scores distinguished between subgroups of people with expected differences in joint health status, including by haemophilia severity (p < 0.0001).

Conclusion This study successfully validated Hemo-FAST as a rapid and reliable tool for the functional assessment of joint health in adults with haemophilia, both in clinical practice and clinical research settings.

Keywords

haemophilia - clinical scoring tool - real-world evidence - joint health - assessment

Introduction

Haemophilia is a rare genetic disorder characterized by a deficiency in coagulation factor VIII (haemophilia A) or IX (haemophilia B), leading to spontaneous bleeds or excessive bleeding as a result of trauma, commonly in the muscles and joints.[1] Despite advances in treatment, joint bleeds may still occur in people with haemophilia (PwH) on prophylactic regimens; they can also be silent, and could result in arthropathy and reduced functional mobility.[1] [2] [3] [4] [5] Recurrent bleeding in PwH is not uncommon after even one instance of haemarthrosis and although prophylaxis decreases bleeding episodes, arthropathy can still occur during the lifetime of a PwH.[6] [7] Monitoring joint status to detect early functional damage is therefore critical for the clinical management of severe, moderate, and mild haemophilia alike; this allows for timely adjustments within prophylactic regimens in order to maintain joint health and minimize joint destruction.[1] According to the current World Federation of Hemophilia (WFH) guidelines, the musculoskeletal health of each patient should be assessed and documented at least annually.[1]

In order to gain insight into the impact haemophilia has on the daily lives of PwH, patient-reported outcomes (PROs) are being increasingly considered, alongside standard clinical outcomes such as factor levels and annual bleed rates. Furthermore, PROs are becoming critical in assessing efficacy of new treatments,[8] [9] and as a complementary measure to other objective measures, such as Haemophilia Joint Health Score (HJHS)[10] and also Haemophilia Early Arthropathy Detection with Ultrasound (HEAD-US).[11]

The HJHS is the current standard used for joint examination in PwH.[1] Administering the HJHS requires training and experience[12] and takes 45 to 60 minutes to complete, limiting its regular use in clinical practice.[13] However, time to completion may vary depending on the experience of the clinician. This has led to calls for new, less time-consuming tools to evaluate joint health.[13] The Hemophilia Functional Ability Scoring Tool (Hemo-FAST) is a simple scoring system designed to quickly assess functional mobility in PwH.[14] [15] By relying on routine physical examinations, Hemo-FAST is intended to be a fast and effective alternative to HJHS.

Hemo-FAST development was based on assessment of relevant items identified in an extensive literature review of existing tools for diseases that involve functionally damaged joints or affect mobility, such as rheumatoid arthritis and multiple sclerosis. The tool was subsequently developed with clinical input, patient testing, and refinement following cognitive interviews during the initial phase of the project.[14] [15] The Hemo-FAST tool comprises a PRO part and a clinician-reported outcome (ClinRO) part, with a score of 0 indicating the best possible joint health status and a score of 100 as the worst.[15]

The purpose of this study (NCT04731701) was to validate the psychometric properties of Hemo-FAST for assessment of joint health in adult PwH,[16] with the ultimate intention of using Hemo-FAST in routine clinical practice and research settings.[14]

Methods

Study Design

This was a low interventional study, with minimal risk, conducted in adults with haemophilia A or B of all severities across multiple centres in France. The study consisted of one visit, during which time PwH were asked to complete the PRO part of Hemo-FAST at both the beginning (test) and end (retest) of the visit. PwH also completed the short-form 36 health survey (SF-36) questionnaire and the patient global impression of severity (PGI-S) questionnaire. Clinicians were asked to complete both the ClinRO part of Hemo-FAST and version 2.1 of the HJHS questionnaire.[a] Hemo-FAST is available through registration at Mapi Research Trust (https://eprovide.mapi-trust.org/instruments/hemophilia-functional-ability-scoring-tool).

To test inter-rater reliability, two qualified clinicians completed the ClinRO part of Hemo-FAST independently for the same patient during the study visit, with both raters blinded to the other's findings. The first 50 PwH consecutively enrolled at sites who agreed to have two raters took part.

Population

Eligible people were aged ≥18 years, had a diagnosis of haemophilia A or B, and the ability to complete the PRO questionnaire in French and provide a signed informed consent form. The study exclusion criteria included: joint replacement within the last 6 months, a non-resolved joint or muscle bleeding events ≤7 days prior to the enrolment visit, comorbid illnesses or other illness that may have independently affected the HJHS and Hemo-FAST score and/or limited ability of the patient to participate in the study as determined by the investigator, and adults under guardianship or curatorship.

Data Collection

The age, sex, body weight, height, medical history (including joint surgery and relevant comorbidities), haemophilia history (including haemophilia type, severity, joint bleeds treated in the last 12 months, and treatment regimen), and history of inhibitors in PwH were collected at the study visit.

The components of the Hemo-FAST questionnaire are summarized in [Fig. 1]. Before validation, the PRO part of the Hemo-FAST tool comprised 17 questions. This was redefined as 15 questions during validation, with the removal of two questions. This was due to lack of variation in response to one question, whilst the second excluded question was considered redundant following reliability assessments of the initial Hemo-FAST tool. All of the results reported here are based on the validated tool that includes the 15-item PRO part (assessing mobility level, walking and bending, hand and finger function, arm function, self-care, household tasks, pain, and support from family and friends) and the 9-item ClinRO part (covering muscular atrophy, lameness, and range of motion) except the completion time, which was assessed for the pre-validation PRO part. The PRO questions were scored on a 4-point Likert scale or as a dichotomized score (apart from one question that was scored on a scale of 0–10). The ClinRO items in the Hemo-FAST tool were scored as a dichotomized score (except for one question scored on a 3-point Likert scale). Scores were transformed to represent the percentage of the total possible score for ease of interpretation, with a score of 0 indicating the best possible joint health whilst a score of 100 indicates the worst joint health. Hemo-FAST total scores were not calculated if more than 10% of questions were unanswered.

Fig. 1 Hemophilia Functional Ability Scoring Tool (Hemo-FAST) components. The pre-validation Hemo-FAST comprised 17 items in the patient-reported outcome (PRO) part whereas, following validation, the final Hemo-FAST PRO scores were calculated based on 15 questions. Time to complete the clinician-reported outcome (ClinRO) part of Hemo-FAST was not assessed in the study.

The HJHS version 2.1 comprised assessment of six joints and global gait, as described previously.[17] The total score, scaled 0 to 124, was calculated by adding together the score for each joint (out of 20) and the gait (out of 4). A high score indicates a worse outcome. When one or more items were missing, the total joint score was not calculated. Normalized HJHS total score, scaled 0 to 1, was computed by adding all non-missing item scores and dividing them by the maximum possible sum for those item scores.

The SF-36 version 1.0 comprises a 36-item, patient-reported survey across eight health domains: physical functioning (10 items), physical role functioning (four items), bodily pain (two items), general health perceptions (five items), vitality/energy/fatigue (four items), social functioning (two items), emotional role well-being (five items), and mental health (four items) over the past 4 weeks.[18] [19] SF-36 consists of a physical component and a mental component score; these can be further sub-divided into eight domains.

The PGI-S is a single self-reported item used to assess haemophilia-related symptom severity. In this study, PGI-S was worded, ‘When thinking about all of the symptoms and physical consequences related to haemophilia that you may have experienced during the past week, please indicate the one option that best described how your symptoms overall have been: no symptoms, mild symptoms, moderate symptoms, or severe symptoms.’[20]

Analyses

Test, Retest, and Completion Times

The data collated from the pre-validation PRO part of the Hemo-FAST questionnaire were utilized to calculate the average time interval between the test and retest, and the total time required to complete the pre-validation PRO part of Hemo-FAST.

Reliability Assessments

Test–retest reliability was assessed according to the intraclass correlation coefficient (ICC) for the Hemo-FAST PRO score between test and retest. ICC values <0.50 indicated poor reliability, values between 0.50 and <0.70 moderate reliability, and values ≥0.70 good reliability.[21] [22] [23]

Inter-rater reliability was assessed according to the ICC for the Hemo-FAST ClinRO score between two raters. A positive rating for reliability was given for an ICC of ≥0.70.

Internal consistency, which assessed intercorrelations across the items of Hemo-FAST, was analyzed according to Cronbach's alpha correlation. A value of 0.70 to 0.95 was considered acceptable,[22] [24] [25] whilst a value of >0.95 indicated high correlations among the items in the scale, suggesting redundancy of one or more items.[24] Inter-item correlations allow determination of whether the hypothesis that Hemo-FAST measured one overarching domain is valid, or if sub-domains should be considered. Inter-item correlation coefficients <0.20 indicates no correlation, whilst >0.70 indicates an item is redundant.[26] Item-to-factor correlation (also called item-to-total correlation) allows determination of whether the item contributed to the construct measurement, with values >0.30 indicating correlation.[23] [27]

Construct Validity Assessments

Construct validity was assessed according to the Pearson's correlation coefficient (r) between the Hemo-FAST score and those of HJHS and SF-36, to examine whether different measures of distinct but related concepts converged in the same direction. Convergent validity (the extent to which two measures relate) criteria were met if r was >0.50.[22] [28]

Known-groups validity assessed the extent to which Hemo-FAST scores were distinguishable between groups of PwH with different characteristics (e.g., haemophilia type, haemophilia severity, HJHS score, and PGI-S score) by t-tests or analysis of covariance. Additionally, three multivariate linear regression models were used to examine whether differences between groups were still present when adjusting for covariates.

Structural Validity Assessments

Floor and ceiling effect was assessed according to the proportion of PwH who achieved the lowest score/floor effect or the highest score/ceiling effect. Floor or ceiling effects were considered present if >15% of PwH achieved the lowest or highest possible scores, respectively.[29] Acceptability and data completeness were assessed according to the percentage of missing responses to analyze the extent to which the total score could be calculated.

Statistical Analysis

The full analysis set (FAS) consisted of all PwH enrolled in the study. A sensitivity analysis set was determined for the test–retest reliability assessment, excluding all PwH with an interval of <30 minutes between test and retest.

The psychometric evaluation was performed using R® environment (R Core Team, 2013; version 4.2.1). Other statistical evaluations were performed using the statistical analysis system (SAS) software package version 9.4 (SAS Institute Inc., Cary NC, USA).

It is generally recommended to include at least 10 patients for each scale item.[30] As the pre-validated Hemo-FAST PRO part was based on 17 questions, approximately 180 PwH were planned to be enrolled to allow for approximately 5% missing data and patient dropout.

Results

Population Characteristics and Demographics

The study enrolled 180 PwH A or B from 14 haemophilia treatment centres across France. Of the 180 enrolled PwH, 83% had haemophilia A and 17% had haemophilia B. No PwH withdrew from the study. The median (range) age of the population was 38.0 years (18.0–78.0) (n = 180) with a median (range) body mass index (BMI) of 24.7 kg/m² (15.4–41.8) (n = 161); one patient was female. Based on factor VIII or IX levels, haemophilia severity was rated as severe for 64%, moderate for 17%, and mild for 19% of PwH ([Table 1]). Of those people with severe haemophilia (n = 115), 96 (84%) and 19 (17%) were receiving prophylaxis and on-demand treatment, respectively.

Table 1 Population demographics and clinical characteristics at baseline (overall population)

Characteristic	All PwH, n = 180[a]
Male/female, n (%)	179 (99)/1 (1)
Age, median (range)	38.0 (18.0–78.0)
BMI (kg/m2), median (range)	24.7 (15.4–41.8)[b]
Haemophilia A/haemophilia B, n (%)	149 (83)/31 (17)
Severity of haemophilia, n (%)
Severe (<1% basal factor VIII or IX)	115 (64)
Moderate (1–5% basal factor VIII or IX)	30 (17)
Mild (>5%– < 40% basal factor VIII or IX)	35 (19)
Current treatment regimen, n (%)
Prophylaxis	111 (62)
On-demand	69 (38)
Type of prophylaxis, n (%)[c]
Primary	11 (10)
Secondary	68 (61)
Tertiary	30 (27)
Unknown	2 (2)
Number of joint bleeds in the 12 months prior to enrolment, n (%)
No bleeds	107 (59)
1–5 bleeds	52 (29)
≥6 bleeds	6 (3)
Unknown	15 (8)
PwH with at least one comorbidity (liver disease, renal disease, CVD, respiratory disease, or GID), n (%)	57 (32)
PwH with at least one surgery in ankles, elbows, knees, shoulders, or hips, n (%)	69 (38)
Inhibitors, n (%)
Inhibitors at the time of the study	9 (5)
A history of inhibitors	11 (6)
No inhibitors and no history of inhibitors	159 (88)
Unknown	1 (1)

Abbreviations: BMI, body mass index; CVD, cardiovascular disease; GID, gastrointestinal disease; PwH, people with haemophilia.

Notes: ^aOne PwH did not provide sufficient responses and was excluded from validation.

^b n = 161.

^c n = 111.

All PwH completed the Hemo-FAST PRO part (excluding one PwH who did not answer enough questions for his responses to be included). The Hemo-FAST PRO retest was subsequently completed by 175/180 PwH, whilst the SF-36 and PGI-S questionnaires were completed by 178 and 177 PwH, respectively. The Hemo-FAST ClinRO part was completed for all PwH and HJHS total score was evaluable for 158 PwH.

Clinical Characteristics

For the PwH who completed the SF-36 (n = 178), mean (standard deviation [SD]) physical and mental component summary scores were 70.3 (20.4) and 72.6 (19.1), respectively. Among the PwH who rated their symptoms via the PGI-S (n = 177), 31% had no haemophilia symptoms, while 41% had mild, 22% had moderate, and 7% had severe symptoms. The mean (SD) HJHS total score was 13.3 (15.8) and the median (range) was 8.0 (0.0–82.0) for 158 PwH; 22 PwH had non-evaluable HJHS scores (for 1 patient all HJHS questions were non-evaluable, with the remaining 21 PwH having one or more non-evaluable joints). To account for this a mean (SD) normalized HJHS total score of 0.13 (0.14; n = 179) was calculated, based on non-missing data.

Completion Time

The PRO part of the pre-validation Hemo-FAST questionnaire was completed by PwH (n = 146) at first test in a mean (SD) of 5.2 (6.1) minutes. The total mean time taken to complete the PRO part of the questionnaire was divided by 17 (as per 17 questions in the pre-validation PRO part), and then multiplied by 15 (as per 15 questions in the validated questionnaire) to give an estimated completion time of mean (SD) 4.6 (5.4) minutes for the validated PRO part of Hemo-FAST.

Reliability Assessments

Test–Retest Reliability

At first test (n = 179), mean (SD) Hemo-FAST total and Hemo-FAST PRO scores were 25.1 (23.0) and 24.1 (22.2), respectively ([Fig. 2A]). These scores were similar at retest (n = 175), when mean (SD) Hemo-FAST total and PRO scores were 24.8 (22.6) and 23.7 (21.5), respectively. The results for the final Hemo-FAST total score at test, by subgroup, are presented in [Table 2].

Fig. 2 Hemophilia Functional Ability Scoring Tool (Hemo-FAST) scores at test and retest in the overall population and subgroups. (A) Total score and patient-reported outcome (PRO) score in the overall population (n = 180). (B) Total score by severity of haemophilia. (C) Total score by treatment regimen. (D) Total score by type of haemophilia. Scores were not calculated if more than 10% of the questions were unanswered. A score of 0 indicated the best possible joint health status and 100 indicated the worst joint health. SD, standard deviation.

Table 2 Results for the Hemo-FAST total score at test, overall, and by subgroup

	Total final Hemo-FAST score at test
	n	Mean (SD)	Median	Min; Max
Overall
(n = 180)	179	25.1 (23.0)	18.8	0.0; 84.0
By age group
18–29 (n = 60)	60	10.3 (14.0)	4.2	0.0; 50.7
30–44 (n = 51)	50	28.1 (21.2)	20.6	0.0; 84.0
≥45 (n = 69)	69	35.8 (24.0)	38.9	0.0; 80.6
By severity of haemophilia
Severe (n = 115)	114	32.1 (23.6)	33.7	0.0; 84.0
Moderate (n = 30)	30	17.9 (20.0)	10.8	0.0; 77.8
Mild (n = 35)	35	8.6 (9.7)	4.9	0.0; 36.1
By type of haemophilia
Haemophilia A (n = 149)	148	26.0 (23.1)	19.8	0.0; 84.0
Haemophilia B (n = 31)	31	20.8 (22.4)	8.3	0.0; 72.2
By current regimen
On-demand (n = 69)	69	14.5 (17.8)	6.9	0.0; 65.3
Prophylactic (n = 111)	110	31.7 (23.5)	31.6	0.0; 84.0
Unknown (n = 0)	0	–	–	–
By occurrence of joint bleeding
No bleeding (n = 107)	106	22.6 (23.6)	14.9	0.0; 84.0
1 to 5 bleeding(s) (n = 52)	52	24.9 (20.3)	20.1	0.0; 77.8
≥6 bleedings (n = 6)	6	42.8 (18.6)	45.8	7.6; 61.6
Unknown (n = 15)	15	36.7 (24.6)	39.1	0.7; 70.8

Abbreviations: Hemo-FAST, Hemophilia Functional Ability Scoring Tool; Max, maximum; Min, minimum; n, number of people with haemophilia; SD, standard deviation.

The Hemo-FAST PRO questionnaire showed good test–retest reliability in both the FAS (ICC: 0.76–0.97) and sensitivity analysis (ICC: 0.78–0.97) sets ([Fig. 3]). The mean (SD) time between test and retest for the pre-validation Hemo-FAST was 44.7 (28.3) minutes (n = 144). The sensitivity analysis excluded PwH who had an interval of <30 minutes between test and retest (n = 95).

Fig. 3 Test–retest reliability of Hemophilia Functional Ability Scoring Tool (Hemo-FAST) patient-reported outcome (PRO) per question. *The sensitivity analysis set included people with haemophilia (PwH) who started the retest of Hemo-FAST PRO more than 30 minutes (and up to 169 minutes) after completing the first test. Q14 and Q17 are not shown because they were removed from the PRO part of the questionnaire following validation. Intraclass correlation coefficient (ICC) values <0.50 indicated poor reliability, values between 0.50 and <0.70 moderate reliability, and values ≥0.70 good reliability.

Inter-rater Reliability

Inter-rater values for the ClinRO questions, assessed according to the ICC, were over 0.7 for seven of the nine questions, indicating good reliability. There were two ClinRO questions with inter-rater values <0.7 at 0.67 and 0.44, indicating moderate-to-high and poor-to-moderate reliability, respectively. The ClinRO part was subsequently revised with additional instructions for all movement assessments, to provide further clarity for clinicians completing the questionnaire and to ensure a consistent approach to the physical examination.

Internal Consistency

Intercorrelation assessment across all items of Hemo-FAST resulted in a Cronbach's coefficient alpha of 0.97 for the entire questionnaire. This demonstrates a high level of internal consistency between the 24 questions, suggesting some redundancy of one or more items. However, inter-item correlations were similar for each question, ranging from 0.53 to 0.55, confirming good internal consistency and no redundancy for any of the specific 24 assessed questions. Item-to-factor correlation coefficients ranged from 0.37 to 0.86. This indicates that all questions are considered acceptable to reliably assess the overall construct evaluated by the Hemo-FAST questionnaire, albeit to different degrees.

Construct Validity Assessments

Construct Validity

The Hemo-FAST questionnaire demonstrated convergent construct validity with the HJHS total scores and with individual SF-36 domains assessing physical health ([Fig. 4]). The study also demonstrated discriminant construct validity of the Hemo-FAST total score regarding mental health with the SF-36 mental health component.

Fig. 4 Construct validity—correlation of Hemophilia Functional Ability Scoring Tool (Hemo-FAST) with Haemophilia Joint Health Score (HJHS) and short-form 36 health survey (SF-36)—presented using Pearson's correlation coefficient. Convergent validity criteria were met if r > 0.50.

Known-groups Validity

At both test and retest the total Hemo-FAST scores approximately doubled between ‘mild’ and ‘moderate’ and between ‘moderate’ and ‘severe’ haemophilia severity ratings ([Fig. 2B]). This was similarly seen with individual PRO and ClinRO scores (data not shown). Also, at test and retest, the Hemo-FAST total scores for PwH under prophylactic treatment were more than double to those measured for PwH treated on-demand, regardless of the severity of the disease ([Fig. 2C]). Similar Hemo-FAST total scores were obtained at test and retest for PwH, regardless of whether they had haemophilia A or B ([Fig. 2D]). The PRO score of the Hemo-FAST tool increased with age, haemophilia severity, and for PwH under prophylactic treatment.

Hemo-FAST scores also discriminated between subgroups of PwH with known differences in characteristics when stratified by HJHS scores and haemophilia severity, although there was no statistical difference between Hemo-FAST total scores for PwH A and B (p = 0.25, data not shown). Hemo-FAST total scores were significantly higher for PwH with abnormal HJHS scores (>3 for PwH ≤50 years old and >8 for PwH >50 years old) compared with PwH with normal HJHS scores (≤3 for PwH ≤50 years old and ≤8 for PwH >50 years old[17]) (p < 0.0001) ([Fig. 5A]). Hemo-FAST total scores were overall statistically different for the mild, moderate, and severe haemophilia subgroups (p < 0.0001) ([Fig. 5B]).

Fig. 5 Construct validity of the Hemophilia Functional Ability Scoring Tool (Hemo-FAST) scale. (A) Hemo-FAST score stratified by Haemophilia Joint Health Score (HJHS) scores. (B) Distribution of Hemo-FAST score according to haemophilia severity. Boxplots showing means (squares), medians (bars inside each box), interquartile ranges (box limits), and minimum and maximum (bars) excluding outliers (triangles). An outlier was defined as a value more than 1.5 times of the interquartile range below the first quartile or above the third quartile. (A) Changes in joint health according to HJHS score were defined as: normal HJHS ≤3 (abnormal >3) for people with haemophilia (PwH) ≤50 years old; normal HJHS ≤8 (abnormal >8) for PwH >50 years old.[17] (B) Mild haemophilia was defined as PwH with >5% to <40% basal factor VIII or IX; moderate haemophilia was defined as PwH with 1 to 5% basal factor VIII or IX; severe haemophilia was defined as PwH with <1% basal factor VIII or IX. F-statistic is the ratio of the variation between sample means and the variation within samples. F, F-statistic.

An adjusted multivariable linear regression model of the Hemo-FAST score demonstrated that Hemo-FAST scores increased with haemophilia severity for people with moderate (p = 0.005) or severe (p < 0.001) haemophilia, compared to people with mild haemophilia. Covariates that contributed significantly to this regression model included age (p < 0.001) and joint surgery status (PwH who had one or more joint surgeries compared with no surgeries) (p = 0.014). Covariates that did not have a significant impact on this regression model included BMI, at least one comorbidity, and number of joint bleeds within the past 12 months. In a separate adjusted regression model, the Hemo-FAST scores did not discriminate between patients with haemophilia A and B. Age had a significant impact also in this regression model (p = 0.003), as did joint surgery status (p < 0.001) and the presence of ≥6 joint bleeds within the past 12 months (p = 0.023).

Structure Validity Assessments

Floor and Ceiling Effects

There was no floor or ceiling effect. No Hemo-FAST total score reached the worst possible value of 100, indicating no ceiling effect. Fewer than 15% of PwH (12%) presented the best possible score of 0, indicating no floor effect.

Acceptability and Data Completeness

The acceptability and data completeness analysis showed around 1% or less of the Hemo-FAST PRO questions were unanswered (99% completion by all PwH). The Hemo-FAST total score could not be calculated for 1/180 participants.

Discussion

Monitoring joint status in PwH is critical to adjust treatment modalities for the maintenance of joint health.[1] Although the HJHS is currently the standard measure used to examine joint health in haemophilia in clinical studies,[1] [12] it is time consuming in daily routine and is entirely clinician-assessed.[13] [17] A simpler tool is therefore needed for use by both PwH and clinicians, with the use of a combined objective joint assessment and PRO to assess outcomes advocated in the literature.[10] Hemo-FAST was developed with the aim of providing a simple, convenient, and reliable tool for assessing joint health that captures clinician and patient perspectives.[14] [15] Hemo-FAST contains both clinician-reported assessments of muscular atrophy, lameness and range of motion, and PROs relating to mobility in their daily lives. Overall, the presented results successfully validate Hemo-FAST as a rapid and reliable tool to assess joint health in PwH.

The study cohort (n = 180) was expected to be comparable to the population assessed for the international validation of the HJHS in adult PwH (n = 192)[17]; in both studies, the ratio of PwH A to B was approximately 5:1, which is in agreement with the WFH-reported prevalence estimate of 80 to 85% and 15 to 20% of all haemophilia cases, respectively.[1] The proportions of severe, moderate, and mild haemophilia were comparable across the two studies. However, the proportion of those on prophylaxis treatment was slightly lower in the HJHS validation study when compared with the current validation study (56% vs. 62%).[17] In both studies, the proportion of PwH with severe versus moderate/mild haemophilia may be higher compared with the wider haemophilia population since those with less severe disease are less likely to regularly attend clinics or be recruited into studies.

Notably, median HJHS total score in our study was lower than in the HJHS validation study.[17] This may be due to slight differences in the study populations or due to the HJHS score not being calculated for 22 PwH in the current study, with all but 3 of them classified as severe. An additional factor may be that, in our study, PwH with a history of joint replacement within the last 6 months were excluded; this has the potential to favour those with low HJHS scores.

Hemo-FAST was designed to be quick to complete in order to overcome the limitation of existing, time-consuming tools: the PRO part of the validated Hemo-FAST questionnaire (15 questions) will take approximately 5 minutes or less to complete. This would allow for frequent evaluation of how joint health affects the daily lives of PwH. The completion time of the recently validated ACTIVLIM-Hemo measure, which assesses the perceived difficulty of PwH in completing certain activities, is also reported to be similar at within 5 minutes.[31] However, this measure was developed as an alternative to the Haemophilia Activity List, includes assessment of non-daily activities (such as skiing, sprinting, and playing a racquet sport), and contains no clinician part; so it can be considered complementary to Hemo-FAST in assessing the overall health of PwH. The Functional Independence Score in Haemophilia (FISH) is a performance-based instrument also developed to assess functional independence in PwH.[32] However, it utilizes a different rating system compared with Hemo-FAST, involves no full self-report by the patient, and was developed and validated in a group of patients who have significant arthropathy. Consequently, FISH is a useful tool for use in adolescents and adult patients who have not had access to prophylaxis, whilst Hemo-FAST was validated across a study population with a high percentage of patients on prophylaxis.[32]

The Hemo-FAST questionnaire demonstrated strong reproducibility: at the first test and retest, mean total score was 25.1 versus 24.8, and mean PRO score was 24.1 versus 23.7, respectively. The Hemo-FAST total score at test and retest was also similar between PwH A and B, and increased according to haemophilia severity. This is in agreement with the recent HJHS validation study for adults, which demonstrated that the HJHS also differentiates between haemophilia severity.[17]

Overall, the total score was higher for adult PwH receiving prophylactic treatment compared with on-demand treatment, as also confirmed by test and retest. This is likely due to a higher proportion of people with severe haemophilia being on prophylactic treatment compared with the overall study population (84% vs. 62%), rather than a reflection of the treatment regimen. Indeed, of this overall adult population receiving prophylaxis, more patients were receiving secondary or tertiary prophylaxis, which are initiated following two or more joint bleeds or the onset of joint disease, respectively,[33] rather than primary prophylaxis. Furthermore, in the 1980s, adoption of prophylaxis was delayed due to the risk of transmission of blood-borne diseases; this has likely impacted joint health in patients with haemophilia born around this time, despite now being on prophylaxis.[34] [35]

The Hemo-FAST score showed high internal consistency. The overall Cronbach's coefficient alpha of >0.95 suggested a degree of redundancy; however, removing one question at a time did not identify any individual question to exclude. Moreover, the inter-item correlation coefficients, which were all <0.70, did not indicate redundancy for any of the PRO or ClinRO Hemo-FAST items.

Convergent validity with total HJHS and the physical component of SF-36 demonstrated that Hemo-FAST captures the functional aspect of haemophilia. Conversely, Hemo-FAST showed discriminant construct validity with the mental component of SF-36, suggesting only functional and not mental impacts of haemophilia are captured. Although changes in joint health do have a direct impact on the mental health of PwH,[36] Hemo-FAST was developed to assess physical joint health, so this was to be expected.

The structural validity of Hemo-FAST was also demonstrated, with no floor or ceiling effect and almost all questions answered by all PwH (99%).

The study has a number of potential limitations. Although the time taken to complete the PRO part of Hemo-FAST was collected, the time taken to complete the ClinRO part of the questionnaire was not recorded; therefore, the time required to complete Hemo-FAST in its entirety is currently not available. However, since the ClinRO part of Hemo-FAST consists of nine questions, based around standard physical examinations performed during routine follow-up of a PwH, it was not found to be time consuming by the clinicians. The mean interval between test and retest was relatively short, which could potentially bias the results in favour of test–retest reliability. However, a sensitivity analysis excluding all PwH with a test–retest interval of <30 minutes was performed to address this and confirm test–retest reliability. A factor analysis was not undertaken, and no statistical analysis was performed to consider the value of prophylactic treatment (as a covariate) on the Hemo-FAST score; comparison of Hemo-FAST scores between the on-demand and prophylactically treated patient groups would require a higher number of PwH. Given the low number of people with severe haemophilia (n = 19) undergoing on-demand treatment, it was not possible to compare Hemo-FAST scores with those undergoing prophylaxis treatment. Furthermore, although Hemo-FAST provides a functional and physical examination of joint health, detection of early changes in joint status may only be possible using certain imaging techniques, such as ultrasound or MRI.[1] [37] The objective of Hemo-FAST was not to replace ultrasound to assess joint health, but rather to provide clinicians with a fast, reliable tool that would bring meaningful information on joint structure and function during an outpatient visit. In this complementary role, patient's functional capabilities and joint health assessment could trigger an additional structural evaluation with imaging technologies such as ultrasound examination. A prospective longitudinal study would be required to assess whether Hemo-FAST may also be an appropriate tool to detect early signs of joint deterioration.

Future investigations could assess the sensitivity of Hemo-FAST to treatment effects over time. The strong psychometric properties and the sensitivity of Hemo-FAST across clinically distinct categories, particularly in differentiating between the mild, moderate, and severe groups, suggest that Hemo-FAST may be able to capture clinical changes that occur over time, whether negative or positive.

Although Hemo-FAST was validated in a population of PwH in France, which is likely to be representative of middle- and high-income countries, it may not be as representative of lower-income countries. To help address this, official translations of the Hemo-FAST questionnaire to cover multiple languages will be released in the future. This will allow Hemo-FAST to be an accessible tool for PwH, their caregivers, and clinicians across different geographies.

Conclusions

The psychometric validation of Hemo-FAST supports its use as a rapid and reliable tool for the functional assessment of joint health in adults with haemophilia A and B, both in routine clinical practice and clinical research settings. To the best of our knowledge, this is the first score that evaluates joint health in daily practice both quickly and easily. The short completion time allows the test to be taken frequently, thus improving patients' experience by enabling them to have regular evaluations of their joint health. This in turn provides insight into how their joint health affects their daily life over time and allows for adaptation of treatment modalities. Overall, Hemo-FAST offers a practical and complete assessment of joint health and functionality as it captures both the patient and clinician perspectives through the use of a PRO and physical exam; its validation could establish Hemo-FAST as a new standard scoring system for use both in clinical practice and clinical research settings.

Conflict of Interest

V.B. declares support for the present manuscript from Sobi to the medical writer for medical writing; grants or contracts from CSL Behring paid to their institution; consulting fees from Sobi and LFB; and support for attending meetings and/or travel from Sobi. Y.R. declares payment or honoraria from Sobi, LFB, and Roche, and support for attending meetings and/or travel from Sobi, CSL Behring, Roche, and Novo Nordisk. D.D. and B.F. declare no conflicts of interest. N.D. declares support for the present manuscript from Sobi to their institution; consulting fees from Sobi and Octapharma; participation on a Data Safety Monitoring Board or Advisory Board for Sobi; and support for attending meetings and/or travel from Novo Nordisk, LFB, and Roche. M.-L.P.-J. declares payment or honoraria from Novo Nordisk, and support for attending meetings and/or travel from Amgen, Novo Nordisk, Octapharma, LFB, CSL Behring, and Sobi. S.Cas. declares consulting fees from Sobi; payment for expert testimony from Sobi, Novo Nordisk, LFB, and Takeda; and support for attending meetings and/or travel from Novo Nordisk, Sobi, and Roche. F.G.-V. declares payment or honoraria for oral presentations from Sobi, Roche, Takeda, and CSL Behring; support for attending meetings and/or travel from Sobi, Roche, LFB, CSL Behring, and Pfizer; and participation on a Data Safety Monitoring Board or Advisory Board for Sobi, Pfizer, and LFB. B.T. declares grants or contracts from Sobi, Novo Nordisk, Roche, Takeda, CSL Behring, and Octapharma paid to their institution; payment or honoraria from Novo Nordisk, Sobi, and CSL Behring; and support for attending meetings and/or travel from Novo Nordisk, Roche, and CSL Behring. A.H. declares payment or honoraria for oral presentations from Sobi, Roche, LFB, Takeda, Novo Nordisk, CSL Behring, and Octapharma; support for attending meetings and/or travel from Sobi, Roche, LFB, Takeda, Novo Nordisk, CSL Behring, and Octapharma; and participation on a Data Safety Monitoring Board or Advisory Board for Sobi, Roche, LFB, Takeda, Novo Nordisk, CSL Behring, and Octapharma. C.G. and N.K. are shareholders of stock or stock options of Sobi, are employees of Sobi, and declare support for the present manuscript from Sobi. M.Z. is an employee of Sobi, and declares support for the present manuscript from Sobi. S.Car. is a shareholder of stock or stock options of Sobi, is an employee of Sobi, and a deputy member of the Sobi board of directors (employee representative). E.B. is a shareholder of stock or stock options of Sobi and is an employee of Sobi. C.N. declares support for the present manuscript from Sobi; grants or contracts from Novo Nordisk and Sobi; consulting fees from Sobi; and payment or honoraria from Novo Nordisk and Sobi. A.L. declares support for the present manuscript from Sobi; grants or contracts from Sobi, Takeda, Bayer, CSL Behring, LFB, Novo Nordisk, Octopharma, Pfizer, and Roche; consulting fees from LFB, Pfizer, Roche, and Sobi; payment or honoraria from LFB, Pfizer, Roche, Sobi, and Takeda; and support for attending meetings and/or travel from CSL Behring, Novo Nordisk, Octapharma, Roche, and Sobi.

Acknowledgement

The authors thank the patients, the investigators, and their teams who took part in the study. We acknowledge the contribution of Prof. Hervé Chambost, Hôpital de la Timone, Marseille, France, and Dr. Roseline d'Oiron, Paris-Saclay University, Le Kremlin-Bicêtre, France, as additional investigators in this study. The authors also acknowledge: Daniela Bruni, Ph.D. Pharm.D. from Sobi for publication coordination; Hayley Macfarlane and Hayley Owen, Bioscript Group, Macclesfield, UK, for medical writing and editorial assistance; and Liz Southey, MSc, of The Salve Health, UK, for support with the graphical abstract. Sobi and Sanofi reviewed and provided feedback on the manuscript. The authors had full editorial control of the manuscript and provided their final approval of all content. Some data in this manuscript have previously been presented as posters at the 12th European Association for Haemophilia and Allied Disorders, Prague, Czech Republic, 6–8 February 2019; the 14th Annual Congress of the European Association for Haemophilia and Allied Disorders, virtual congress, 3–5 February 2021; the 12th Bari International Conference, Palermo, Italy, 8–10 September 2023; the 32nd Congress of the International Society on Thrombosis and Haemostasis, Bangkok, Thailand, 22–26 June 2024, and as an oral presentation at the World Federation of Hemophilia Congress, Madrid, Spain, 21–24 April 2024.

^a Hemophilia Joint Health Score 2.1 ©. The Hospital for Sick Children, Centre Hospitalier, Universitaire Sainte Justine, the Regents of the University of Colorado, Karolinska Hospital, University Medical Center Utrecht, 2009. Used under licence by The Hospital for Sick Children.

Data Availability Statement

Sobi is committed to responsible and ethical sharing of study data, while protecting individual participant integrity and compliance with applicable legislation. Data access will be granted in response to qualified research requests. All requests are evaluated by a cross-functional panel of experts within Sobi and a decision on sharing will be based on the scientific merit and feasibility of the research proposal, maintenance of personal integrity, and commitment to publication of the results. To request access to study data, a data sharing request form (available on www.sobi.com) should be sent to [e-mail]. Further information on Sobi's data sharing policy and process for requesting access can be found at: https://www.sobi.com/en/policies. Hemo-FAST© is protected by international copyright, with all rights reserved by Sobi. Do not use without permission. For information on, or permission to use Hemo-FAST©, please contact Mapi Research Trust at: https://eprovide.mapi-trust.org/instruments/hemophilia-functional-ability-scoring-tool.

Authors' Contribution

Conceptualization of the study: C.N., V.B., C.G., A.H., and A.L. Study design: C.N., A.H., and A.L. Data acquisition: C.N., D.D., F.G.-V., N.D., S.Car., V.B., Y.R., M.-L.P.-J., B.F., B.T., S.Cas., A.L., and C.G. Data analysis: C.N, N.K., E.B., and A.L. Data interpretation: C.N., N.K., S.Car., V.B., B.F., E.B., A.L., and C.G. All authors meet the International Committee of Medical Journal Editors criteria for authorship.

• References

• 1 Srivastava A, Santagostino E, Dougall A. et al; WFH Guidelines for the Management of Hemophilia panelists and co-authors. WFH Guidelines for the Management of Hemophilia, 3rd edition. Haemophilia 2020; 26 (Suppl. 06) 1-158
• 2 Lobet S, Hermans C, Lambert C. Optimal management of hemophilic arthropathy and hematomas. J Blood Med 2014; 5: 207-218
• 3 Olivieri M, Kurnik K, Pfluger T, Bidlingmaier C. Identification and long-term observation of early joint damage by magnetic resonance imaging in clinically asymptomatic joints in patients with haemophilia A or B despite prophylaxis. Haemophilia 2012; 18 (03) 369-374
• 4 Berntorp E, Dolan G, Hay C. et al. European retrospective study of real-life haemophilia treatment. Haemophilia 2017; 23 (01) 105-114
• 5 Wilkins RA, Stephensen D, Siddle H. et al. Twelve-month prevalence of haemarthrosis and joint disease using the Haemophilia Joint Health score: evaluation of the UK National Haemophilia Database and Haemtrack patient reported data: an observational study. BMJ Open 2022; 12 (01) e052358
• 6 Gualtierotti R, Solimeno LP, Peyvandi F. Hemophilic arthropathy: current knowledge and future perspectives. J Thromb Haemost 2021; 19 (09) 2112-2121
• 7 Skinner MW, Nugent D, Wilton P. et al. Achieving the unimaginable: health equity in haemophilia. Haemophilia 2020; 26 (01) 17-24
• 8 Manco-Johnson MJ, Warren BB, Buckner TW, Funk SM, Wang M. Outcome measures in haemophilia: beyond ABR (annualized bleeding rate). Haemophilia 2021; 27 (Suppl. 03) 87-95
• 9 Konkle BA, Skinner M, Iorio A. Hemophilia trials in the twenty-first century: defining patient important outcomes. Res Pract Thromb Haemost 2019; 3 (02) 184-192
• 10 Fischer K, Nijdam A, Holmström M. et al. Evaluating outcome of prophylaxis in haemophilia: objective and self-reported instruments should be combined. Haemophilia 2016; 22 (02) e80-e86
• 11 Martinoli C, Della Casa Alberighi O, Di Minno G. et al. Development and definition of a simplified scanning procedure and scoring method for Haemophilia Early Arthropathy Detection with Ultrasound (HEAD-US). Thromb Haemost 2013; 109 (06) 1170-1179
• 12 Fischer K, Poonnoose P, Dunn AL. et al; participants of the International Symposium on Outcome Measures in Hemophilic Arthropathy. Choosing outcome assessment tools in haemophilia care and research: a multidisciplinary perspective. Haemophilia 2017; 23 (01) 11-24
• 13 Kuijlaars IAR, van der Net J, Feldman BM. et al. Evaluating international Haemophilia Joint Health Score (HJHS) results combined with expert opinion: options for a shorter HJHS. Haemophilia 2020; 26 (06) 1072-1080
• 14 Barbay V, Negrier C, Harroche A. et al. Ongoing validation of a new disease-specific instrument assessing functional abilities in patients with haemophilia: The Haemophilia Functional Ability Scoring Tool (Hemo-FAST). Haemophilia 2021; 27: Abstract 149
• 15 Négrier C, Barbay V, Harroche A. et al. A simple functional mobility assessment tool for use in clinical practice or real-life studies with hemophilia patients: Step 1, development of the questionnaire (French version). Haemophilia 2019; 25: Abstract P068
• 16 ClinicalTrials.gov. Study identifier NCT04731701. Psychometric Validation of the Hemophilia Functional Ability Scoring Tool (Hemo-FAST). Accessed June 27, 2024 at: https://classic.clinicaltrials.gov/ct2/show/NCT04731701
• 17 St-Louis J, Abad A, Funk S. et al. The Hemophilia Joint Health Score version 2.1 Validation in Adult Patients Study: a multicenter international study. Res Pract Thromb Haemost 2022; 6 (02) e12690
• 18 Brazier JE, Harper R, Jones NM. et al. Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ 1992; 305 (6846) 160-164
• 19 Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30 (06) 473-483
• 20 Guy W. ECDEU Assessment Manual for Psychopharmacology. US Department of Health, Education, and Welfare, Public Health Service, Alcohol, Drug Abuse, and Mental Health Administration, National Institute of Mental Health, Psychopharmacology Research Branch, Division of Extramural Research Programs; 1976
• 21 Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 2016; 15 (02) 155-163
• 22 Terwee CB, Bot SD, de Boer MR. et al. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 2007; 60 (01) 34-42
• 23 Nunnally J, Bernstein I. The assessment of reliability. In: Psychometric Theory. 3rd ed.. McGraw-Hill; 1994: 248-292
• 24 Tavakol M, Dennick R. Making sense of Cronbach's alpha. Int J Med Educ 2011; 2: 53-55
• 25 Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika 1951; 16: 297-334
• 26 Ferketich S. Focus on psychometrics. Aspects of item analysis. Res Nurs Health 1991; 14 (02) 165-168
• 27 de Vet HCW, Terwee CB, Mokkink LB. et al. Measurement in Medicine—A Practical Guide. New York, USA: Cambridge University Press; 2011
• 28 Messick J. Validity of psychological assessment: validation of inferences from persons' responses and performances as scientific inquiry into score meaning. Am Psychol 1994; 50: 741-749
• 29 McHorney CA, Tarlov AR. Individual-patient monitoring in clinical practice: are available health status surveys adequate?. Qual Life Res 1995; 4 (04) 293-307
• 30 Nunnally JC. Psychometric Theory. 2nd ed.. New York: McGraw-Hill; 1978
• 31 Lobet S, Lambert C, Foubert A. et al. ACTIVLIM-Hemo: a new self-reported, unidimensional and linear measure of activity limitations in persons with haemophilia. Haemophilia 2023; 29 (01) 317-328
• 32 Poonnoose PM, Thomas R, Keshava SN. et al. Psychometric analysis of the Functional Independence Score in Haemophilia (FISH). Haemophilia 2007; 13 (05) 620-626
• 33 Blanchette VS, Key NS, Ljung LR, Manco-Johnson MJ, van den Berg HM, Srivastava A. Subcommittee on Factor VIII, Factor IX and Rare Coagulation Disorders of the Scientific and Standardization Committee of the International Society on Thrombosis and Hemostasis. Definitions in hemophilia: communication from the SSC of the ISTH. J Thromb Haemost 2014; 12 (11) 1935-1939
• 34 Meunier S, Trossaërt M, Berger C. et al; Groupe Prophylaxie CoMETH. [French guidelines. Long-term prophylaxis for severe haemophilia A and B children to prevent haemophiliac arthropathy]. Arch Pediatr 2009; 16 (12) 1571-1578
• 35 O'Mahony B. Haemophilia care in Europe: past progress and future promise. Haemophilia 2020; 26 (05) 752-758
• 36 Al-Huniti A, Reyes Hernandez M, Ten Eyck P, Staber JM. Mental health disorders in haemophilia: systematic literature review and meta-analysis. Haemophilia 2020; 26 (03) 431-442
• 37 De la Corte-Rodriguez H, Rodriguez-Merchan EC, Alvarez-Roman MT, Martin-Salces M, Jimenez-Yuste V. Clinical assessment and point of care ultrasonography: how to diagnose haemophilic synovitis. Haemophilia 2022; 28 (01) 138-144

Publication History

Received: 23 August 2024

Accepted: 10 December 2024

Article published online:
21 January 2025