Essentials Psychological benefit of physical fitness in people with hemophilia (PWH) is poorly understood. This study aimed to enable PWH to define, meet and maintain individual fitness goals. Themes identified were fear, self-confidence, ‘being normal’, pain, weight loss, ability, and getting fitter. Psychological wellbeing (motivation self-efficacy and self-esteem) improved during the study. 1 INTRODUCTION

Life expectancy for people with hemophilia (PWH) has increased because of advances in treatment. Alongside this, there is an emergence of age-related disorders such as cardiovascular disease, overweight, obesity, and cancer at the same rate as the general population.1-3 The UK government's public health strategy recommends that all adults are physically active for 150 min of moderate intensity activity each week4 and encourages participation in sport.5

Improved physical health has been shown to positively impact mental health.6 A large cross-sectional study demonstrated that physical exercise was significantly associated with improved self-reported mental health with a 22.3% and 20.1% reduction in mental health issues in team sports and aerobic gym activities, respectively.7

Contemporary hemophilia care supports physical activity “to promote physical fitness and normal neuromuscular development, with attention paid to physical functioning, healthy body weight, and self-esteem.”8 The physical benefits of exercise in PWH are well recognized,9 despite the psychological challenges they face.10 However, there are few data on the psychological determinants and outcomes of physical fitness in PWH.

In the SO-FIT study, boys (aged 8–17 years) self-reported good physical functioning overall but impairment in the domains of “friends” and “sports and school.”11 In a UK study of 84 children with hemophilia (mean age 11.5 years), 95% participated in sport and reported better physical performance and quality of life (QoL) than the boys who did not.12 Adults reported less sports participation, which was attributed to physical condition or fear of injury.13 Despite reporting motivation to “get fit” and “lose weight,” young adults also report that they are not reaching their full ability because of perceived fears about bleeding and injury.14 This concurs with Buxbaum et al., who report lower self-efficacy scores in adolescents with hemophilia than in controls and suggests that this is related to perceived barriers in physical activity.15 Taylor et al. reported that UK adults with hemophilia wished to be active but spent many hours in sedentary states because of pain or the fear of bleeding.16 These studies suggest that although, at this time, children with hemophilia engage in physical activities and sports, adolescents and adults report perceived challenges in participation because of their hemophilia. The psychological benefits of exercise and fitness are rarely reported by PWH because of previous personal history, a lack of self-confidence, and the perceived ongoing risk of injury.

The objectives of Project GYM were to identify strategies that would enable PWH to define, meet and maintain their individual fitness goals. The principal efficacy and safety results of Project GYM have been reported elsewhere17; the psychological findings of this study are reported in this paper.

2 METHODS

This feasibility study recruited young men, aged 18–25 years, with hemophilia A or B (all severities ± inhibitor), via three London hemophilia centers. Participants attended a Project GYM induction session and were given free access to a central London gym and a Fitbit activity tracker to wear. As a feasibility study, a pragmatic number of 20 participants was intended with 10 in each of two randomized subgroups: personal trainer versus “gym only” (i.e., self-trained). Following randomization using a sealed envelope system, half of the participants were given fully funded access to a personal trainer. The study received Health Research Authority and ethical approval (IRAS project ID: 241384; REC reference 18/WA/0179).

During the gym induction, participants completed baseline questionnaires to evaluate motivation to exercise (the Stages of Change Questionnaire),18, 19 physical activity (International Physical Activity Questionnaire [IPAQ]),20 self-efficacy (Self Efficacy to Manage Chronic Disease Scale),21 self-esteem (Rosenberg's Self-Esteem Scale),22 and QoL (EQ-5D).23 These were repeated 6 months later alongside a semistructured interview, either face to face or by telephone, to explore their reasons for joining the study, the facilitators or barriers to completing the program, and any other highlights or concerns. The interviews followed an initial topic guide that was adapted to reflect the views of earlier respondents. Interviews were recorded, transcribed verbatim, and analyzed using thematic analysis.

2.1 Analysis

This study was not designed to compare differences between the two groups but as a feasibility study to evaluate motivation to exercise and its impact on physical activity, self-efficacy, self-esteem, and QoL. The quantitative data were analyzed using SPSS (version 25; SPSS Chicago, IL). Descriptive data are shown as median ± standard deviation, median, and range (minimum–maximum). Bootstrap confidence intervals and standard deviations were computed using R (seed = 129,791, number of resamples = 100,000). Bootstrapping was used to estimate the summary statistics for this population from a small (yet representative) sample group.

Differences in efficacy endpoints (changes in scores start vs. end) were tested in R using Pearson's chi-squared test and Wilcoxon signed-rank test with continuity correction. A p < 0.05 was considered statistically significant.

The qualitative data from the interviews were analyzed using NVIVO 12 for Mac. Each interview was read independently by two authors (K.K., S.F.), and comments were identified and coded using thematic analysis.24 All coded comments were then reviewed and developed into overarching themes and agreed by all authors.

3 RESULTS

Nineteen participants median age of 22.1 (18.1–24.1) years from a wide ethnic mix were recruited; 14 had hemophilia A, 12 were severely affected, four had mild, and three had moderate hemophilia. The majority of participants all with severe hemophilia (13/19, 68.4%) received prophylactic factor replacement (three with extended half-life products), there were no current inhibitors; full patient demographic data are reported elsewhere.17 Nineteen participants commenced the study and 6-month follow-up questionnaires were completed by 17 participants; 10 participants agreed to the final recorded interview; five of these were face to face and five by telephone lasting a median 10 min (range 5.5–24.5). Of the remainder, six did not respond to interview invitations, and one left the United Kingdom for a gap year and was not able to be contacted. There were no discernible differences between those who were interviewed and those not in the overall data analysis.

3.1 Motivation to exercise

During the course of the study, there was a significant shift in motivation to exercise as shown by Stages of Change grouping (Table 1). At baseline, participants appeared to be motivated to change their exercise levels, with one in the contemplation phase, four in preparation, three in action phase, and 11 in the maintenance phase. By study end, 16 were in maintenance and one was in the action phase. Data collected using the IPAQ short form showed that activity levels increased in seven participants and decreased in two, with no change in eight; follow-up data were missing for two participants. The number of participants with a high level of physical activity (20 min vigorous activity on at least 3 days, or 7 or more days of walking for at least 30 min per day in a week [as defined by the IPAQ]) increased from eight to 13. Those in the low (no moderate or vigorous activity) and moderate (3 days of 20 min of activity or 5 days of walking for 30 min) categories decreased from four to zero and seven to four, respectively. Those participants with access to a personal trainer attended the gym more regularly; however, there was no statistically significant difference in IPAQ score between the two groups.

TABLE 1. Global changes in efficacy endpoints Measure Statistic Data Test p value

Prestudy

(n = 19)

Poststudy

(n = 17)

Stages of change Change in grouping (start vs. end)

Precontemplation (0)

Contemplation (1)

Preparation (4)

Action (3)

Maintenance (11)

Precontemplation (0)

Contemplation (0)

Preparation (0)

Action (1)

Maintenance (16)

Pearson's χ2 = 6.84, df = 2 0.03 IPAQ Change in score (start vs. end)

Low (4)

Moderate (7)

High (8)

Low (2)a

Moderate (4)

High (13)

Pearson's χ2 = 2.68, df = 2 0.27 Self-efficacy Change in score (start vs. end)

Median = 49

Range = 19–60

Median = 52.5

Range = 25–60

Wilcoxon signed-rank test with continuity correction

Z = 22.5

0.06 EQ-5D index score Change in score (start vs. end)

Median = 0.837

Range = 0.414–1.000

Median = 0.819

Range = 0.512–1.000

Wilcoxon signed-rank test with continuity correction

Z = 26

0.56 EQ-5D VAS Change in score (start vs. end)

Median = 75

Range = 15–100

Median = 80

Range = 45–98

Wilcoxon signed-rank test

Z = 24.5

0.05 Self-esteem Change in score (start vs. end) Total score median = 22, range = 12–30

Median = 26

Range = 13–30

Wilcoxon signed-rank test with continuity correction

Z = 11

0.02 a Assigned values because of noncompletion. 3.2 Self-efficacy

At baseline, the median self-efficacy score (scored from 0 to 10) for participants was 8.17 (range 3.17–10.0, n = 19); by study end, the scores had improved to 8.75 (range 4.17–10.0, n = 16). Although nonsignificant, the overall self-efficacy scores increased from baseline to study end (Table 2). Two subquestions (“confidence in keeping the physical discomfort or pain of disease from interfering with the things the patient wishes to do” and “confidence in keeping the emotional distress caused by disease from interfering with the things the patient wishes to do”) showed a significant increase (p < 0.05) in the poststudy survey results.

TABLE 2. Global changes in self-efficacy responses (n = 16) Question: How confident are you that you can Baseline Poststudy Z p value 1. Keep the fatigue caused by your disease from interfering with the things you want to do?

Median: 8; [min max]: 3 10

SD: 0.750; CI (high low): 9 7

Median: 8.5; [min max]: 4 10

SD: 0.790; CI (high low): 10 7

16 0.07 2. Keep the physical discomfort or pain of your disease from interfering with the things you want to do?

Median: 8; [min max]: 3 10

SD: 0.898; CI (high low): 9 5

Median: 8; [min max]: 4 10

SD: 0.592; CI (high low): 10 8

7 0.01 3. Keep the emotional distress caused by your disease from interfering with the things you want to do?

Median: 8; [min max]: 1 10

SD: 1.756; CI (high low): 10 5

Median: 9; [min max]: 3 10

SD: 0.763; CI (high low): 10 7

6.5 0.04 4. Keep any other symptoms or health problems you have from interfering with the things you want to do?

Median: 8.5; [min max]: 3 10

SD: 0.838; CI (high low): 9 6

Median: 9; [min max]: 4 10

SD: 0.741; CI (high low): 10 7

22.5 0.20 5. Do the different tasks and activities needed to manage your health condition so as to reduce your need to see a doctor?

Median: 9.5; [min max]: 3 10

SD: 0.719; CI (high low): 10 8

Median: 8.5; [min max]: 4 10

SD: 0.505; CI (high low): 9 8

44 0.72 6. Do things other than just taking medication to reduce how your illness affects your everyday life?

Median: 9; [min max]: 2 10

SD: 0.749; CI (high low): 10 7

Median: 9; [min max]: 4 10

SD: 0.653; CI (high low): 10 7

36.5 0.87 Total score

Median: 49; [min max]: 19 60

SD: 5.194; CI (high low): 56 39

Median: 952.5; [min max]: 25 60

SD: 3.092; CI (high low): 55 45

22.5 0.06 Note Bootstrap confidence intervals and standard deviations computed using R (seed = 129,791, number of resamples = 100,000). 3.3 Self-esteem

At baseline, the median self-esteem score using the Rosenberg's Self-Esteem Scale (range 0–30) was 22 (12–30, n = 19), with three participants recording scores below the reported normal range (15–25), suggesting low self-esteem. At study end, the median self-esteem score was 25 (range 13–30, n = 17). Twelve participants showed significant improvements (p = 0.02) and only one recorded a score below the normal range. Global changes in self-esteem responses are summarized in Table 3.

TABLE 3. Global changes in self-esteem responses Question Baseline Poststudy Z score p value 1. I feel that I am a person of worth, at least on an equal plane with others

n = 18

Median (min max): 3 (2 3);

SD 0.141; CI (high low) 3 2.5

n = 18

Median (min max): 3 (2 3);

SD 0.142; CI (high low) 3 2.5

5 0.99 2. I feel that I have a number of good qualities

n = 17a

Median (min max): 3 (2 3);

SD 0.289; CI (high low) 3 2

n = 17

Median (min max): 3 (2 3);

SD 0.179; CI (high low) 3 2

2 0.77 3. All in all, I am inclined to feel that I am a failure

n = 18

Median (min max): 2 (0 3);

SD 0.246; CI (high low) 3 2

n = 18

Median (min max): 3 (1 30);

SD 0.445; CI (high low) 3 2

0 0.09 4. I am able to do things as well as most other people

n = 18

Median (min max): 2 (1 3);

SD 0.357; CI (high low) 3 2

n = 18

Median (min max): 3 (2 3);

SD 0.246; CI (high low) 3 2

10 0.06 5. I feel I do not have much to be proud of

n = 17

Median (min max): 2 (1 3);

SD 0.292; CI (high low) 3 2

n = 17

Median (min max): 3 (1 3);

SD 0.180; CI (high low) 3 2

12 0.04 6. I take a positive attitude toward myself

n = 17

Median (min max): 2 (1 3);

SD 0.183; CI (high low) 3 2

n = 17

Median (min max): 3 (1 3);

SD 0.400; CI (high low) 3 2

9 0.1 7. On the whole, I am satisfied with myself

n = 17

Median (min max): 2 (1 3);

SD 0.458; CI (high low) 3 1

n = 17

Median (min max): 3 (1 3);

SD 0.470; CI (high low) 3 2

13.5 0.28 8. I wish I could have more respect for myself

n = 16

Median (min max): 2 (1 3);

SD 0.437; CI (high low) 2 1

n = 16

Median (min max): 2 (0 3);

SD 0.324; CI (high low) 3 1

3.5 0.13 9. I certainly feel useless at times

n = 17

Median (min max): 2 (0 3);

SD 0.440; CI (high low) 3 1

n = 17

Median (min max): 2 (1 3);

SD 0.482; CI (high low) 3 2

4 0.07 10. At times I think I am no good at all

n = 17

Median (min max): 2.5 (1 3);

SD 0.432; CI (high low) 3 2

n = 17

Median (min max): 3 (1 3);

SD 0.400; CI (high low) 3 2

9 0.82 Total score

n = 18

Median (min max): 22 (12 30);

SD 1.547; CI (high low) 25.5 19.75

n = 15

Median (min max): 26 (13 30);

SD 2.142; CI (high low) 28.5 20.5

11 0.01 Note Bootstrap confidence intervals and standard deviations computed using R (seed = 129,791, # resamples = 100,000). a Missing data where numbers are <18.

The baseline dataset indicated a significant (p < 0.01) correlation between self-esteem and self-efficacy that was maintained at study end (baseline Spearman's rank correlation coefficient (r) 0.65, p ≤ 0.01, n = 18; study end 0.90, p ≤ 0.01, n = 14).

3.4 Quality of life

QoL at baseline was reported as good (median EQ-5D index value 0.84, score range 0–1 where 1 corresponds to full health) and showed no significant change over the study duration (study end median value 0.82 [p < 0.56]), although the EQ-5D Visual Analog Scale (VAS) score (range 0–100) where 100 represents the “best” score increased significantly between study start and end (Wilcoxon signed-rank test, Z = 24.5, p = 0.05). Overall, scores remained the same for six participants, worsened for four, and improved for seven. Clinically meaningful change25 occurred in six participants with improved scores using the VAS. Table 4 shows the number and proportions of those reporting perfect health versus problems within the domains of the EQ-5D. One participant reported more problems with mobility at study end; however, reductions were seen in the number of participants reporting problems in both the pain/discomfort and anxiety/depression domains. For the pain domain, two with severe and two with moderate hemophilia reported “slight” pain on the EQ-5D; this did not correlate with their recent hemophilia joint health scores (median 1, range 0–7 [n = 7] in the moderate and 3, range 0–13 [n = 12] in the severe hemophilia groups). There was improvement with participants (n = 17) recording pain (no pain or slight pain) at follow-up (7 vs. 9 and 8 vs. 5, respectively). Two participants recorded moderate pain at both study start and end, and one reported slight pain becoming moderate, unrelated to bleeding. There was a significant correlation between EQ-5D index and self-esteem in both prestudy and poststudy data (p = 0.02 vs p = 0.01); using the VAS data, there was a significant correlation with self-esteem only in the poststudy data (p = 0.08 vs p = 0.01; Table 5).

TABLE 4. Numbers and proportions reporting perfect health versus problems within the domains of the EQ-5D Level Mobility Self-care Usual activities Pain/discomfort Anxiety/depression

Baseline

(n = 19)

Poststudy

(n = 17)

Baseline

(n = 19)

Poststudy

(n = 17)

Baseline

(n = 19)

Poststudy

(n = 17)

Baseline

(n = 19)

Poststudy

(n = 17)

Baseline

(n = 19)

Poststudy

(n = 17)

Number reporting no problem (level 1) 13 (68%) 10 (59%) 18 (94%) 16 (94%) 14 (74%) 12 [71%) 8 (42%) 9 (53%) 9 (47%) 11 (65%) Number reporting some problem (levels 2–5) 6 (32%) 7 (41%) 1 (6%) 1 (6%) 5 (26%) 5 (29%) 11 (58%) 8 (47%) 10 (53%) 6 (35%) Change in numbers reporting problems +1 0 0 −3 −4 TABLE 5. Regression analyses on EQ-5D and self-esteem score EQ-5D Index Self-esteem Spearman's rank correlation coefficient p value n Median Range Median Range Baseline 0.837 0.414–1.000 22 12–30 0.545 0.02 18 Poststudy 0.786 0.512–1.000 26 13–30 0.646 0.01 15 EQ-5D VAS Self-esteem Spearman's rank correlation coefficient p value n Median Range Median Range Baseline 75 15–100 22 12–30 0.425 0.08 18 Poststudy 80 45–98 26 13–30 0.633 0.01 15 Note Missing/incomplete responses: prestudy (1), poststudy (4). 3.5 Interview data

Key themes identified from the interviews were: fear and self-confidence, being normal, pain, weight loss, ability, getting fitter (Table 6).

TABLE 6. Themes and supporting quotes Theme Example quotes Fear and self-confidence “I was quite afraid, I really wanted to do a sport or gym to be fitter and healthier” (21 year old, severe hemophilia A) “This will sound a bit weird … but you are not the same as everyone else – don't try to do what others do, be conscious that you don't have to do this” (23 year old, severe hemophilia A) “My mum and dad refused to let me do sport [because of hemophilia], I like it and I want to be good at it” (22 year old, severe hemophilia A) “I feel more happier and more confident in general – it's made me really happy … it has really changed me, I'm going to be a different person because of this” (21 year old, severe hemophilia A) “I feel like it has gone up, the confidence and stuff” (20-year-old, moderate hemophilia B) Being normal “I'm certainly more aware that I can do as much as the next person can” (23 year old, severe hemophilia A) “I don't think people with hemophilia should be any different” (24 year old, severe hemophilia B) “I can train hard and I can be just as good as you, if not more” (22 year old, severe hemophilia A) Pain ‘I have sore ankles and now if I have done a certain amount of steps in the day I can think …that's activity not bleeds so I'm not taking pain killers just for the sake of it” (24 year old, severe hemophilia B) Weight loss