Routine management, healthcare resource use and patient and carer‐reported outcomes of patients with transfusion‐dependent β‐thalassaemia in the United Kingdom: A mixed methods observational study

Abstract Objectives We evaluated routine healthcare management, clinical status and patient‐ and carer‐reported outcomes in UK paediatric and adult patients with transfusion‐dependent β‐thalassaemia (TDT). Methods A multi‐centre, observational mixed‐methodology study evaluated 165 patients (50% male; median age 24.1 [interquartile range (IQR)] 11.8–37.2] years) from nine UK centres. Results Patients had a mean of 13.7 (standard deviation [SD] ±3.2) transfusion episodes/year (mean retrospective observation period 4.7 [±0.7] years). The median (IQR) for iron overload parameters at the last assessment during the observation period were: serum ferritin (n = 165) 1961.0 (1090.0–3003.0) μg/L (38% > 2500 μg/L); R2 liver iron (n = 119) 5.4 (2.9–11.6) mg/g (16% ≥15 mg/g); T2* cardiac iron (n = 132) 30.3 (22.0–37.1) ms (10% < 10 ms). All patients received ≥1 iron chelator during the observation period; 21% received combination therapy. Patients had a mean of 7.8 (±8.1) non‐transfusion‐related hospital attendances or admissions/year. Adult patients’ mean EQ‐5D utility score was 0.69 (±0.33; n = 94 [≥16 years]) and mean Transfusion‐dependent quality of life score was 58.6 (±18.4; n = 94 [≥18 years]). For Work Productivity and Activity impairment, mean activity impairment for patients ≥18 years (n = 88) was 48% (±32%) and for carers (n = 29) was 28% (±23%). Conclusions TDT presents significant burden on patients, carers and healthcare resources.


INTRODUCTION
Transfusion-dependent β-thalassaemia (TDT) is a rare, severe genetic disease affecting adult haemoglobin production, necessitating lifelong blood transfusions and iron chelation therapy (ICT) [1][2][3]. Blood transfusions lead to iron overload requiring regular monitoring and optimisation of ICT to mitigate complications; however, this intensive therapy and monitoring imposes a great burden on patients and their families [4][5][6][7][8]. Complications associated with iron overload include endocrine dysfunction and hepatic pathology; cardiac disease is the most serious, leading to arrhythmias and heart failure, and early mortality [1,3,9].
While life expectancy for patients with TDT has significantly improved in the past 50 years in the United Kingdom (UK) [10], the crude 10-year mortality rate for patients with TDT in England was recently reported as more than five times higher than the age/sex adjusted general population (6.2% vs. 1.2%; p < 0.001) [11], suggesting that despite care delivered in a high-income setting, mortality from TDT remains significantly higher than the general population.
Until recently, allogeneic haematopoietic stem cell transplantation (allo-HSCT) was the only potentially curative treatment. Allo-HSCT is generally reserved for paediatric patients with a human leukocyte antigen-matched related donor (25-30% of patients [12]), and not routinely offered for UK adult patients [3]. Recent advances in gene therapy have increased the avenues for potentially transformative treatments for TDT [13]. An understanding of the routine management of TDT and associated burden on patients, carers and healthcare resources is important in evaluating the impact of emerging treatments.
The present study evaluated current management pathways for TDT, clinical status, healthcare resource use, and the impact of TDT on quality of life (QoL) and work productivity of patients and carers in the United Kingdom.

Study design and setting
A multi-centre, observational mixed-methodology study, involving a retrospective chart review and cross-sectional survey of paediatric and adult patients with TDT and their carers, was conducted in nine UK National Health Service (NHS) centres. The study observation period for eligible patients with TDT diagnosed ≥5 years prior to data collection was the 5-year period prior to data collection or death; for patients with TDT diagnosed 2-5 years prior to data collection it was the period from TDT diagnosis to data collection or death ( Figure 1). Baseline was defined as the start of the patient's observation period. Centres were selected in order to give a good geographical spread (including centres in England, Scotland and Wales) and to be representative of practice across the range of centres managing patients with TDT in the UK, including adult haematology centres, paediatric only centres and centres managing both adult and paediatric patients. Potential centres were identified and selected based on a feasibility questionnaire and Plain language summary

What is the new aspect of your work?
We report the results of a mixed methods observational study that provides an integrated understanding of the

Endpoints and data collection
The primary endpoint was the number of blood transfusion  questionnaire [17,18]. TDT impact on work and daily activities was evaluated using the Work Productivity Activity Index (WPAI) questionnaire [19][20][21].
The EQ-5D-3L questionnaire was used (under licence) to assess the general health status of patients and their carers on the day of completion. Adults (≥16 years) were administered the EQ-5D-3L and children (8-15 years) were administered the EQ-5D-Y, a child-friendly EQ-5D version. A proxy version of the EQ-5D-Y was administered only to carers of children between the ages of 4-7 years, which required the proxy to rate the health of the child. No version of the EQ-5D-Y is available for children aged under 4 years [14 -16]. The EQ-5D-3L comprised two parts. The first evaluated five dimensions of health (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) on three levels (level 1: no problems; level 2: some problems; level 3: extreme problems) to give a 5-digit health state that was converted to a summary health utility score ranging from 1 (perfect health) to < 0 (0 is equivalent to death and negative values represent states worse than death) using the UK value set [22]. There is currently no value set for the selfcompleted EQ-5D-Y (8-15 years); therefore, utility scores could not be calculated for this age group [14]. The second part of the EQ-5D-3L used a visual analogue scale (VAS) to assess self-rated health from 0 (worst imaginable) to 100 (best imaginable). The EQ-5D-Y measured the same dimensions as the EQ-5D-3L worded in a manner suitable for completion by children.
TDT-related QoL for patients and carers was assessed using the TranQoL questionnaire (under a user agreement). Adults (≥18 years) were administered the adult version, carers were administered the parent version to complete about themselves, children (7-17 years) were administered the child version. A proxy version of TranQoL was administered only to carers of children <7 years of age [17,18].
TranQoL is a thalassaemia-specific QoL instrument that assesses the impact of disease over the 7 days prior to completion on physical health, emotional functioning, family functioning and school and career functioning, with the adult version also covering sexual activity. The individual domain scores and the TranQoL overall score range from 0 to 100, with higher scores representing better QoL [18].
The impact of TDT was also assessed using the WPAI questionnaire for adult patients (≥18 years) and carers only; the WPAI specific health problems (SHP) version for patients and the caregiver (WPAI:CG) version for carers were used as appropriate [19][20][21]. The WPAI assesses the impact of disease on work productivity over the 7 days prior to completion by evaluating absenteeism (the amount of work time missed), presenteeism (impaired working effectiveness [WPAI definition]) and overall work productivity loss (absenteeism and presenteeism), as well as assessing activity impairment (impact on non-work related activities). Domain scores are expressed as percent impairment, with higher scores indicating a greater degree of impairment [19,20].

Statistical analyses
Normally distributed quantitative variables and variables describing hospital resource utilisation are presented as mean (± standard deviation

Iron burden and iron chelation therapy
(28/119) had moderate LIC (>7 to < 15 mg/g) and 16% (19/119) had high LIC (≥15 mg/g; Figure 2B). The median cardiac T2* at the last assessment was 30. In post hoc analyses, 44% of patients had average intervals of <2 years between LIC assessments and 29% had intervals of <2 years between cardiac iron assessments (Supplementary Figure S2). Intervals of <2 years were observed for 43% of patients with high LIC and 54% of patients with severe cardiac iron at the last assessment during the observation period (Supplementary Figure S3). The proportion of patients with an interval of <2 years between assessments of LIC and cardiac iron stratified by age at baseline are summarised in Supplementary Figure S4.
The median age at initiation of ICT was 2.9 (IQR 1.

TDT-related hospital attendances and admissions
In addition to attendances for blood transfusions and cross-matching  Table S2). The mean length of stay for inpatient admissions was 5.6 (±6.2) days. Overall, patients had a mean of 7.8 (±8.1) non-transfusionrelated hospital attendances or admissions/year. Fewer than 5% of non-transfusion-related outpatient attendances, day case admissions and inpatient admissions were to cardiology or endocrinology (Supplementary Table S2).

Patient-and carer-reported outcomes
At least one patient-reported outcome questionnaire was completed  ) calculated using the UK value set [22]; there is currently no value set for the self-completed EQ-5D-Y (8-15 years), therefore utility scores could not be calculated for this age group; the UK reference value was calculated from published age-specific UK population norms weighted for the age-distribution of the adult patients assuming the population norm for age 16-17 years was equivalent to age 18-24 years [43];

DISCUSSION
This study provides a contemporary (2018-2019) and comprehensive picture of the substantial impact of TDT on the daily lives of patients and their carers, and the significant burden on secondary healthcare resources in the UK. Patient demographics were consistent with the wider population of patients with thalassaemia [11,23]. Based on recent data from England [24], we estimate that the study included approximately 15% of the UK population of patients with TDT. Comorbidities and TDT-related complications were commonplace, with the latter representative of the spectrum of well-documented complications [3,11,25,26].
Patients had a mean of 13.7 blood transfusion episodes/year, consistent with a recent UK Hospital Episodes Statistics database study [11]. In addition, patients had an average of 13.7 cross-match attendances and 7.8 other TDT-related hospital attendances and admissions annually, equating to almost three TDT-related appointments/month in total. The frequent hospital attendances and admissions have potential financial implications for patients and their families in terms of out-ofpocket expenses and impact on employment, as well as representing a considerable burden on healthcare resources [8,[27][28][29][30].
Despite management in specialist centres, significant iron overload affected a subset of patients; 10% had severe cardiac iron loading, 16% high LIC and 38% high serum ferritin at the last assessment and one patient died due to cardiogenic shock. Chronic iron overload is associated with significant morbidity and mortality and, despite advances in ICT and monitoring, iron-related cardiomyopathy remains the most  [3,34,35]. High iron burden may also partly reflect non-adherence to ICT and at least one non-adherence event was documented in 25% of patients; however, non-adherence to desferrioxamine was lower than expected based on previously published patient-reported and carer-reported adherence [36]. The reason for this is unclear, but may partly reflect under-reporting in routine clinical practice. Equally, patients on desferrioxamine may represent a motivated population who were managing complications and/or had good adherence with effective iron control and did not wish to switch to oral therapy. Non-adherence to ICT has been suggested to limit the survival advantage afforded by full adherence [37] and may be influenced by age, difficulties with administration, the life-long nature of treatment and the occurrence of side effects [36,38]. In this context, AEs associated with ICT were common, and approximately 50% of all

ACKNOWLEDGEMENTS
This study was funded by bluebird bio, who were involved in study design, interpretation of the data, and the drafting, revision and final approval of this manuscript for submission.
The authors would like to thank all the patients and their carers for participating in this study. The authors thank pH Associates Ltd, doing business as OPEN Health, a healthcare consultancy company specialising in real-world evidence, who were commissioned by bluebird bio to provide support for study design and implementation, data analysis and interpretation, and medical writing support (M Watt) in the drafting and submission of this manuscript.

DATA ACCESSIBILITY STATEMENT
How or where can the data be obtained?
Appropriately de-identified patient-level datasets and supporting documents may be shared following attainment of applicable marketing approvals and consistent with criteria established by bluebird bio and/or industry best practices to maintain the privacy of study participants. For more information please contact datashar-ing@bluebirdbio.com.

When will data availability begin?
Upon request following attainment of applicable marketing approvals.