Assessment of Clinical Characteristics and Socio-Contextual Factors on Medication Adherence in Children and Adolescents with Hemophilia

Abstract

Introduction:

Socio-contextual factors, including Social Determinants of Health (SDOH) affect health outcomes and clinical progression in chronic blood diseases such as hemophilia. Continuous prophylaxis with intravenous clotting factor concentrates or subcutaneous factor VIII mimetic agents reduces bleeding and improves quality of life. However, adherence to prophylaxis is challenging.

Aim:

To investigate the impact of clinical, and socio-contextual factors, including SDOH, on adherence to prophylaxis in children and adolescents with hemophilia.

Methods:

In this cross-sectional study, parent and adolescent participants were enrolled from 2018 to 2023. Adherence to prophylaxis was measured by Hemophilia Regimen Treatment Adherence Scale-Prophylaxis as overall adherence as well as ‘time,’ ‘dose,’ ‘plan,’ ‘remember,’ ‘skip’ and ‘communicate’ subscales. Socio-contextual factors were assessed by self-reported validated surveys. Linear regression analysis was performed.

Results:

Of 41 parent participants, 56% were Hispanic. There was no difference in adherence scores between Hispanic and non-Hispanic parents (p=0.97). Older age of child (p=0.002), prophylaxis via peripheral intravenous route (vs central line, p=0.004) and use of Spanish language for medical communication (vs English, p=0.002) were associated with lower adherence. Higher trust in provider was associated with higher adherence (p=0.051).

Conclusion:

Our findings may help guide tailored interventions to improve adherence, including targeting older children/adolescents as they become independent in their care, language-concordant education, and measures to improve social support and enhance trust in healthcare providers.

Introduction

Prophylactic medication for prevention of bleeding is considered standard of care in persons with severe hemophilia and for those with moderate disease with severe bleeding phenotype.¹ Adherence rates to prophylaxis vary based on patient and treatment characteristics, though most published studies involving children and adolescents have showed relatively high adherence^2–13 (Supplemental Table 1). Barriers to adherence to prophylaxis prevent persons with hemophilia from achieving the full range of prophylaxis benefits. Barriers are multifactorial and have been categorized into five categories: patient-related, condition-related, treatment-related, health-care system-related, and related to socioeconomic factors.¹⁴ Although adherence to prophylaxis requires adequate health literacy, access to high-cost medication, and regular specialized medical appointments, the role of socio-contextual factors, particularly social determinants of health (SDOH), in adherence to prophylaxis has not been thoroughly explored. Evaluation of associations between racial or ethnic differences and hemophilia-related outcomes has been limited in the Hispanic population.¹⁵ A recent study identified no differences in hemophilia related distress scores between Hispanic and non-Hispanic White participants, though the study was limited by English speaking participants only and a small proportion of Hispanics.¹⁶ This highlights the need for understanding sociodemographic differences beyond race and ethnicity and the urgency to investigate the role of contextual barriers such as healthcare costs, insurance coverage, transportation, sick leave, childcare as well as language discordance.¹⁷ The objective of this study was to investigate the impact of clinical characteristics and socio-contextual factors, including SDOH, on adherence to prophylaxis in children and adolescents with hemophilia in a diverse population, with a representative sample of Hispanics, a group underrepresented in hemophilia research. We hypothesized that Hispanic ethnicity, lower acculturation, use of Spanish language for medical communication, and limited health literacy are associated with lower adherence to prophylaxis.

Methods

Study design and participants

This cross-sectional study was conducted from September 1, 2017 to October 31, 2023 at Rady Children’s Hospital San Diego (RCHSD), a quaternary pediatric hospital-based hemophilia treatment center (HTC) for ~120 children/adolescents with hemophilia. The study was approved by the University of California San Diego/RCHSD Institutional Review Board. Children (0–12.99 years) and adolescents (13–17.99 years) with hemophilia who were receiving continuous prophylaxis with either clotting factor concentrate (CFC) or emicizumab (a Factor VIII mimetic) for at least six months prior to study enrollment were identified through the HTC database and appointment schedules. Parents of the identified children/adolescents, and adolescents with hemophilia were eligible if they had understanding of spoken and written English and/or Spanish language. Participants were consented/assented in their preferred language at scheduled outpatient visits at the RCHSD HTC and did not receive any compensation for participation.

Study measures

Adherence

The primary outcome was adherence to prophylaxis which was measured using the English and Spanish versions of the Validated Hemophilia Regimen Treatment Adherence Scale-Prophylaxis (VERITAS- Pro) for those on CFC prophylaxis.¹⁸ VERITAS- Pro is a 24-item scale which includes six subscales with Likert scale questions. The total adherence score ranges from 24 (most adherent) to 120 (least adherent); adherent is defined as a total score <57. The six subscale adherence scores for Time, Dose, Plan, Remember, Skip and Communicate range from 4 (most adherent) to 20 (least adherent); similar to total score, a lower score indicates higher adherence. The Spanish version of VERITAS-Pro was used for participants with preferred language as Spanish, and permission was obtained for use of both English and Spanish versions. Since a similar validated tool for measurement of adherence to emicizumab prophylaxis was not available at the time of this study, we designed an adapted version of the VERITAS-Pro scale to measure adherence to emicizumab (Supporting Information 2). The adapted version differs from the recently validated, emicizumab specific, VERITASNexGen tool¹⁹ since it did not capture ‘injection’ and ‘bleeds’ subscales, however, had subscales such as ‘remembering’, ‘skipping’ and ‘communicating’ like the original VERITAS-Pro tool.¹⁸

Covariates of interest

Socio-demographic information was collected via a self-reported questionnaire: parental age, sex, education level, race and ethnicity, preferred language for medical communication, marital status, employment, and child’s/adolescent’s health insurance type. Median income was assessed based on zip code data of home address and compared to median income of the respective residence county as determined by the United States Census Bureau.²⁰ Clinical data were abstracted from the electronic health record: child’s/adolescent’s age, type of hemophilia [hemophilia A vs hemophilia B], severity of hemophilia [severe (factor activity < 1%) vs moderate (factor activity 1–5%) vs mild (factor activity >5%)], presence of active inhibitor, type of prophylaxis (standard half-life CFC vs extended half-life CFC vs emicizumab), frequency of prophylaxis (more frequent- daily, every other day or three times a week vs less frequent- twice a week, once a week, once every 10 days, once every two weeks or once every four weeks) and route of administration of prophylaxis [peripheral intravenous (IV) vs central venous catheter (CVC) vs subcutaneous]. Socio-contextual factors, including SDOH, such as health literacy²¹ and household material hardship,^22,23 and trust in provider,²⁴ functional social support,²⁵ bleeding disorder knowledge,²⁶ and acculturation²⁷ (for Hispanics only) were obtained via administering respective surveys (Supplemental Table 2). All surveys were available in English and Spanish and were administered by bilingual research staff. For instruments that were not validated in Spanish, we utilized the Brislin²⁸ method, comparing back-translated documents with the source language (English), which has been used extensively in cross-cultural research.

Statistical analysis

A sample size of 40 consecutively enrolled parent participants of children/adolescents with hemophilia and 15 adolescent participants with hemophilia was planned. Descriptive statistics were calculated for demographic, clinical and socio-contextual characteristics for all participants. The primary analysis was conducted for parent participants. Mean [standard deviation (SD)] and frequency (percentage) were used for continuous and categorical variables respectively. Associations between adherence outcomes (total and six individual subscale scores) were analyzed using three different predictor subsets (socio-demographics, clinical characteristics, and socio-contextual factors) in a linear regression model for parent participants. A univariate analysis was performed and covariates with p-value <0.2 underwent backward elimination to only include variables with p-value <0.1 for further analysis in the multivariable model. Coefficient estimates (β) and 95% confidence intervals (CI) were reported. The p-values were two-sided and those <0.05 were considered statistically significant. When a parent of more than one child/adolescent with hemophilia (n=4) participated, data were collected for each child/adolescent. When both parents of the same child/adolescent with hemophilia (n=6) enrolled, data were collected from each parent. Given the small number of clusters, a linear regression model was utilized. A sensitivity analysis was performed using univariate linear regression without data of all clusters (n=10 excluded).

Due to the small sample size of adolescent participants (n =15), descriptive statistics were calculated only for adolescent data and correlations were assessed by Spearman’s rho as appropriate. All statistical analyses were performed using the R software, version 4.0.2 (R Project for Statistical Computing, 2020).

Results

Parent participants

Forty-one parents of 38 children/adolescents with hemophilia participated in the study. The majority of parent participants were female (88%). Most common self-identified race and ethnicity of parents was White (76%) and Hispanic (56%), respectively. Mean (SD) age of children with hemophilia was 8.5 (5.9) years. Most common type of hemophilia was hemophilia A (83%), most had severe hemophilia (83%), and most did not have an active inhibitor (85%). The majority were on CFC (85%) prophylaxis. Table 1 summarizes socio-demographic, clinical, and socio-contextual characteristics of parents and children/adolescents.

TABLE 1.

Baseline sociodemographic, clinical, and socio-contextual characteristics

	Ethnicity		Total n=41
	Hispanics n=23 (56%)	Non-Hispanics n=18 (44%)
Sociodemographic Characteristics of Parent(s) and Children/Adolescents
Parental Age, n (%)
18–44 years	21 (91%)	14 (78%)	35 (85%)
≥45 years	2 (9%)	4 (22%)	6 (15%)
Parental Sex, n (%)
Female	21 (91%)	15 (83%)	36 (88%)
Male	2 (9%)	3 (17%)	5 (12%)
Parental Level of Education, n (%)
High School or less	16 (70%)	2 (11%)	18 (44%)
Some College and above	7 (30%)	16 (89%)	23 (56%)
Parental Race, n (%)
White/Caucasian	16 (70%)	15 (83%)	31 (76%)
Asian	0 (0%)	1 (6%)	1 (2%)
Native Hawaiian/Pacific Highlander	0 (0%)	1 (6%)	1 (2%)
Mixed/Multiple	7 (30%)	1 (6%)	8 (20%)
Parental Preferred Language, n (%)
English	9 (39%)	18 (100%)	27 (66%)
Spanish	14 (61%)	0 (0%)	14 (34%)
Parental Marital Status, n (%)
Married	14 (61%)	11 (61%)	25 (61%)
Unmarried	9 (39%)	7 (39%)	16 (39%)
Parental Employment Status, n (%)
Employed	10 (44%)	13 (72%)	23 (56%)
Unemployed	13 (56%)	5 (28%)	18 (44%)
Child/Adolescent’s Insurance Type
Public	18 (78%)	5 (28%)	23 (56%)
Private	5 (22%)	13 (72%)	18 (44%)
Child/Adolescent’s Age (years), mean (SD)	9.1 (6.0)	7.7 (5.8)	8.5 (5.9)
Household Income, n (%)	n=21	n=12	n=33
Below Median County Income	16 (76%)	10 (83%)	26 (79%)
At or Above Median County Income	5 (24%)	2 (17%)	7 (21%)
Clinical Characteristics of Child/Adolescent
Hemophilia Type, n (%)
Hemophilia A	20 (87%)	14 (78%)	34 (83%)
Hemophilia B	3 (13%)	4 (22%)	7 (17%)
Hemophilia Severity, n (%)
Mild/Moderate	4 (17%)	3 (17%)	7 (17%)
Severe	19 (83%)	15 (83%)	34 (83%)
Presence of Active Inhibitor, n (%)
Yes	3 (13%)	3 (17%)	6 (15%)
No	20 (87%)	15 (83%)	35 (85%)
Type of Prophylaxis, n (%)
Standard Half-life CFC	11 (48%)	5 (28%)	16 (39%)
Extended Half-life CFC	11 (48%)	8 (44%)	19 (46%)
Emicizumab	1 (4%)	5 (25%)	6 (15%)
Frequency of Prophylaxis, n (%)
More Frequent	16 (70%)	7 (39%)	23 (56%)
Less Frequent	7 (30%)	11 (61%)	18 (44%)
Prophylaxis Route, n (%)
CVC	15 (65%)	7 (39%)	22 (54%)
Peripheral IV	7 (30%)	6 (33%)	13 (32%)
Subcutaneous	1 (4%)	5 (28%)	6 (15%)
Socio-contextual Characteristics Per Parent Report
Health Literacy, n (%)
Adequate	8 (35%)	16 (89%)	24 (59%)
Inadequate	15 (65%)	2 (11%)	17 (41%)
Household Material Hardship, n (%)	n=11	n=8	n=19
Hardship	6 (55%)	4 (50%)	10 (53%)
No hardship	5 (45%)	4 (50%)	9 (47%)
Trust in Provider
PEDI-Tips score, mean (SD)	47.5 (4.9)	49.4 (4.1)	48.3 (4.6)
Functional Social Support	n=11	n=8	n=19
Duke UNC Functional Social Support score, mean (SD)	54 (10)	56 (11)	54 (10)
Bleeding Disorder Disease Knowledge	n=17	n=13	n=30
Knowledge score, mean (SD)	15.2 (3.1)	17.4 (2.1)	16.1 (2.9)
Acculturation (for Hispanics only), n (%)
High Acculturation	16 (70%)	NA	16 (70%)
Low Acculturation	7 (30%)	NA	7 (30%)

SD: standard deviation, CFC: clotting factor concentrate, CVC: central venous catheter, IV: intravenous, PEDI-Tips- Pediatric Trust in Physician, NA: not applicable

Adherence to prophylaxis (parent-report)

Forty parent participants were included in the primary analysis of adherence. One parent who did not complete the outcome measure survey for adherence was not included in the analysis. The mean (SD) total VERITAS-Pro score for parental report was 33 (8), and there was no significant difference in total or subscale adherence scores between Hispanics and non-Hispanics (Table 2).

TABLE 2.

Parent-report of total and subscale adherence scores of children/adolescents with hemophilia

	Hispanics n=22	Non-Hispanics n=18	All n=40
Total score (range: 24 to 120*), mean (SD)	33 (8)	32 (9)	33 (8)
Subscale scores (range: 4 to 20*), mean (SD)
Time	5.6 (3.7)	7.4 (4.6)	6.4 (4.1)
Dose	5.1 (2.2)	4.6 (1.9)	4.9 (2.1)
Plan	5.6 (2.2)	4.5 (0.6)	5.1 (1.7)
Remember	5.6 (1.8)	5.2 (2.1)	5.4 (1.9)
Skip	4.6 (1.5)	5.1 (2.3)	4.8 (1.9)
Communicate	6.0 (2.3)	5.7 (1.8)	5.9 (2.1)

^*Lower score = higher adherence, Higher score = lower adherence. SD: standard deviation

In the multivariable regression analysis of sociodemographic factors and adherence scores, older age of child/adolescent was significantly associated with lower total adherence [β estimate, 0.67 [95% CI, 0.26 to 1.1]; p=0.002). Older age of child/adolescent was also significantly associated with lower adherence to the ‘time’ subscale (β estimate, 0.26 [95% CI, 0.04 to 0.47]; p=0.02). Male sex of parent (β estimate, 2.4 [95% CI, 0.54 to 4.2]; p=0.01) and older age of child/adolescent (β estimate, 0.11 [95% CI, 0.01 to 0.22]; p=0.04) were significantly associated with lower adherence to the ‘dose’ subscale. Spanish as the language used for medical communication was significantly associated with lower adherence to the ‘plan’ subscale (β estimate,1.8 [95% CI, 0.69 to 2.8]; p=0.002). Older age of child/adolescent was significantly associated with lower adherence to the ‘remember’ subscale (β estimate, 0.13 [95% CI, 0.04 to 0.23]; p= 0.007). No sociodemographic variables were significantly associated with ‘skip’ or ‘communicate’ subscales (Supplemental Table 3).

In the multivariable regression analysis of clinical characteristics and adherence scores, administration via peripheral IV had lower total adherence than administration via CVC (β estimate, 8.00 [95% CI, 2.7 to 13]; p =0.004) and there was no significant difference between CVC and subcutaneous routes (p=0.80). Similarly, administration via peripheral IV route was significantly associated with lower adherence to the ‘time’ subscale (β estimate, 3.7 [95% CI, 1.1 to 6.3]; p=0.007), ‘remember’ subscale (β estimate, 1.8 [95% CI, 0.55 to 3.1]; p=0.006) and ‘skip’ subscale (β estimate, 1.5 [95% CI, 0.17 to 2.8]; p=0.03) as compared to CVC route. No clinical characteristic variables were significantly associated with ‘dose’ or ‘plan’ or ‘communicate’ subscales (Supplemental Table 4 ).

In the multivariable regression analysis of socio-contextual factors and adherence scores, lower trust scores were associated with lower total adherence, though significance was borderline (β estimate, −0.56 [95% CI, −1.1 to 0.01]; p=0.05). Lower trust scores were also significantly associated with lower adherence for the ‘plan’ subscale (β estimate, −0.18 [95% CI, −0.28 to −0.07]; p=0.002) and ‘remember’ subscale (β estimate, −0.17 [95% CI, −0.30 to −0.05]; p=0.008). None of the other socio-contextual factors were significantly associated with ‘time’ or ‘dose’ or ‘skip’ or ‘communicate’ subscales (Supplemental Table 5).

Sensitivity analysis of parent data

By excluding 10 parents with more than one child/adolescent or both parents for the same child/adolescent from the analysis, hemophilia type became significantly associated with the ‘time’ subscale, and the presence of active inhibitor became significantly associated with the ‘plan’ and ‘skip’ subscales. Prophylaxis route was no longer significantly associated with ‘remember’ and ‘skip’ subscale scores.

Results of adolescent participants

Mean (SD) age of adolescent participants was 14.9 (1.4) years and the self-reported race and ethnicity of adolescents were White (53%) and Hispanic (67%), respectively. Baseline socio-demographic, clinical and socio-contextual characteristics of adolescent participants are summarized by ethnicity in Supplemental Table 6 . The mean (SD) total and subscale adherence scores reported by adolescents are shown by ethnicity in Table 3. The mean (SD) total adherence scores of adolescents based on sociodemographic, clinical, and socio-contextual characteristics for categorical variables are summarized in Supplemental Table 7. For continuous variables, we found that younger adolescents (spearman rho, −0.02 [95% CI −0.62 to 0.47]), lower trust (spearman rho, −0.48 [95% CI, −0.82 to 0.16]), lesser social support (spearman rho, −0.62 [95% CI, −1.0 to 0.07]) and lower bleeding disorder disease knowledge (spearman rho, −0.16 [95% CI, −0.73 to 0.34]) were associated with lower total adherence although not reaching statistical significance.

TABLE 3.

Adolescent-report of total and subscale adherence scores

	Hispanics n=10	Non-Hispanics n=5	All n=15
Total score (range: 24 to 120*), mean (SD)	45 (16)	52 (16)	48 (16)
Subscale scores (range: 4 to 20*), mean (SD)
Time	6.4 (2.2)	9.6 (5.3)	7.5 (3.7)
Dose	6.8 (4.0)	5.2 (2.2)	6.3 (3.5)
Plan	8.6 (4.5)	10.0 (3.1)	9.1 (4.0)
Remember	7.8 (3.3)	9.0 (5.2)	8.2 (3.8)
Skip	7.2 (2.3)	7.8 (3.6)	7.4 (2.6)
Communicate	8.6 (3.6)	10.4 (4.3)	9.2 (3.8)

^*Lower score = higher adherence, Higher score = lower adherence. SD: standard deviation

Discussion

In this cross-sectional study at a large pediatric hospital-based HTC with a representative proportion of Hispanics we found that, although parents and adolescents reported high adherence to prophylaxis as measured by VERITAS-Pro, certain adverse SDOH, clinical characteristics and socio-contextual factors were associated with lower adherence to prophylaxis in children and adolescents with hemophilia.

Parents of older children/adolescents reported inferior total and ‘time,’ ‘dose’ and ‘remember’ subscale adherence. Previous studies have shown that adherence worsens as age advances from childhood to adolescence and young adult age.^4,8,13,14,29 Reasons for poorer adherence in adolescents and young adults include transfer of hemophilia related care from caregiver to self, transition of care from pediatric to adult models of care, challenges navigating the healthcare system, difficulty maintaining health insurance, and taking on new responsibilities.¹⁴ Adolescents and young adults who have few symptoms during childhood may lack understanding of the importance of continued prophylaxis with advancing age.^29,30 A multidisciplinary approach supporting self-management skills of young patients is key to a successful transition for young persons with hemophilia.^30,31

We found that adherence to CFC prophylaxis via peripheral IV was lower than with CVC. Difficulty with venous access has often been cited as reason for non-adherence.^14,29,32 These challenges may be overcome with use of CVC for administration of CFC since they offer advantages of easier access, less risk of vein scarring, prompt treatment of bleeding episodes, and decreased healthcare costs due to less emergency visits.³³ We did not identify a difference in adherence between CFC via CVC and emicizumab prophylaxis via subcutaneous route. This is in contrast to prior studies which demonstrate high acceptability of emicizumab,³⁴ and high adherence compared to CFC or bypassing agent prophylaxis.³⁵ Also, we did not identify a difference in adherence between participants on standard half-life versus extended half-life CFC. This finding is similar to reports by providers of persons with hemophilia,³⁶ though patients switching from standard half-life to extended half-life CFC have reported to have maintained or improved adherence rates with extended half-life product use.³⁷ Lack of difference in adherence between types of prophylaxis is likely due to small sample size, particularly the low numbers of patients prescribed emicizumab during the study period. Larger studies could evaluate differences in adherence between prophylaxis regimens and impact on clinical outcomes.

We found significant associations between various socio-contextual factors and adherence. Data related to socio-contextual factors collected using validated tools (Supplemental Table 2) was not only limited to SDOH such as health literacy and household material hardship but also included important characteristics that affect the care of persons with bleeding disorders and can impact adherence to medication use such as social support, knowledge of disease, trust in provider, and acculturation.^9,38–41 Half of our study participants self-identified as Hispanic, contributing to the scant literature on outcomes in this marginalized population. Even though Hispanic parents had lower levels of education, employment, health literacy and bleeding disorder knowledge than non-Hispanics, no difference in adherence rates were identified by ethnicity. However, parents of children with hemophilia who used Spanish for medical communication had lower adherence in the ‘plan’ subscale. ‘Plan’ subscale refers to ensuring adequate supplies for medication administration, including keeping track of doses before they run out. Parents who use Spanish may not understand directions for refilling medications and/or have difficulty communicating with pharmacies for ensuring timely delivery of medications impacting adherence. Thus, addressing SDOH such as limited English proficiency with language concordant education including verbal and/or written instructions is of utmost importance to ensure adherence. Culturally aware case managers and parent liaisons can advocate for children with hemophilia and their parents. Diverse racial and ethnic representation in HTC staff also facilitates building trustworthy relations which can address patient concerns, honor autonomy and engage families in behavioral change and improvement initiatives related to hemophilia treatment and education.¹⁷ This is particularly relevant since Hispanic persons are the largest growing ethnic minority in the United States.⁴² While no studies have explicitly looked at differences in adherence between persons with hemophilia of different races, Merz et al. raise an important point that race, being a social construct, should not be used as the only proxy for disease process or disparity.¹⁷ Future studies including more diverse samples of patients are needed to disentangle factors associated with disparate outcomes.

Male parents were noted to have lower adherence scores in the ‘dose’ subscale. ‘Dose’ subscale refers to the use of recommended or prescribed dose. Mothers tend to assume most responsibility related to treatment. Fathers may not regularly infuse factor.⁴³ Differences in parental gender roles in the care of children and adolescents with hemophilia could be attributed to its X-linked model of inheritance. In individuals with family history of hemophilia, mothers who have witnessed family members affected by the same condition, may be more cognizant of the care needed.

Parents (in regression analysis) and adolescents (in descriptive analysis) with higher trust in provider reported higher adherence to prophylaxis in several subscales. The role of trust in provider in adherence to prophylaxis in children and adolescents with hemophilia, though of great relevance, is understudied. Providers are pivotal in shared decision-making for persons with hemophilia when choosing the best treatment option based on the patient’s values and preferences.⁴⁴ Hence, having a trusting patient-provider relationship considerably facilitates adherence to prophylaxis. Congruently with our results, in a study of adult men with hemophilia, higher physician trust scores were associated with higher adherence scores.⁴⁰ In addition, De Moerloose et al. showed that close relationships with hematologists and nurses was associated with greater adherence to therapy.³² Thus, efforts must be undertaken to establish trusting relations between persons with hemophilia, their caregivers, and their HTC providers. Trusting relationships are also essential when pediatric patients transfer to adult HTCs, since lack of trust in the new medical team may contribute to treatment non-adherence.³¹ Social support through family, peers, and psychosocial teams from HTC and the community can contribute to improved disease perception and overall better adherence.^9,14,41

This single-center study has some limitations. The small sample size, attributed, in part, to difficulty in recruiting participants due to competing research studies and the need for completing multiple questionnaires, precluded comparison between parent and adolescent dyads and was insufficient to support robust statistical analysis of the adolescent data. Larger multi-center studies can investigate differences in reported adherence between caregivers and adolescents. In addition, the study was pursued at a large tertiary HTC where access to care for persons with hemophilia differs from those in resource-constrained settings, including low- and middle-income countries, limiting generalizability of results. Moreover, although the sample comprised a large proportion of Hispanics, Hispanic participants were primarily from Mexico and may limit generalizability to other Hispanic subgroups.

Several measures for monitoring adherence to prophylaxis with CFC in children, adolescents and young adults with hemophilia are available. Supplemental Table 1 summarizes listed measures and reported levels of adherence in pediatric studies of persons with hemophilia on CFC prophylaxis. In general, the reported adherence rates were high, however it is important to note that there are discrepancies between subjective and objective measures of adherence⁴⁵ and no single tool/measure may truly reflect overall adherence. Less data are available about adherence to emicizumab in childhood.^34,46 In our study, measurement of adherence to prophylaxis was self-reported using the VERITAS-Pro tool and, although our study was initially designed to capture pharmacy records as an objective means of measuring adherence, we could not use this measure due to the multiple pharmacies dispensing CFC or emicizumab for our patients and inability to consistently obtain records. While the VERITAS-Pro tool may have overestimated adherence in this study, it still offers the advantage of identifying modifiable areas of non-adherence such that targeted interventions by providers and the HTC team can be undertaken.¹⁸ Evaluation of adherence to emicizumab was limited, given relatively few participants who were prescribed emicizumab during the study period and a lack of validated instrument at the time. Future studies should include a larger number of patients prescribed emicizumab, use of the VERITASNexGen tool,¹⁹ validated to evaluate adherence to emicizumab, and inclusion of patients using new prophylaxis therapies as they are approved.

Conclusion

Understanding the role of barriers to adherence in diverse racial and ethnic populations, including the role of socio-contextual factors and SDOH such as acculturation and English-limited proficiency, can lead to the development of culture and language concordant interventions in the care of persons with hemophilia. Utilization of validated adherence tools to identify patient specific areas of non-adherence and addressing socio-contextual factors that negatively impact adherence should be incorporated in clinical practice. Implementing strategies to improve adherence during transition from childhood to adolescence, personalizing prophylaxis to minimize administration burden, building trustworthy relations with providers to aid in shared decision-making, and ensuring robust psychosocial and community networks support may help mitigate non-adherence and improve long-term outcomes for persons with hemophilia.

Abbreviation Key

SDOH

Social Determinants of Health

RCHSD

Rady Children’s Hospital San Diego

HTC

Hemophilia Treatment Center

CFC

Clotting factor concentrate

VERITAS-Pro

Validated Hemophilia Regimen Treatment Adherence Scale-Prophylaxis

IV

Intravenous

CVC

Central venous catheter

SD

Standard deviation

CI

Confidence interval