Healthcare resource use associated with the diagnosis of transthyretin amyloidosis cardiomyopathy

Abstract Objectives Our primary aim was to evaluate the healthcare resource use associated with the diagnosis of transthyretin amyloidosis cardiomyopathy. Second, we aim to assess the effect of the number of diagnostic tests and clinical contact points on the total time and costs between symptom onset and diagnosis defining a quantitative hypothetical optimized diagnostic pathway. Setting Clinical and cost data were collected from patients presenting between 2010 and 2018 in a tertiary referral institution in South London involving two participating hospitals. Participants Thirty‐eight adult patients with a definite diagnosis of transthyretin amyloidosis cardiomyopathy were included, mostly male (n = 28, 74%) and of African‐Caribbean descent (n = 23, 64%). We excluded patients without a confirmed transthyretin amyloidosis cardiomyopathy or those on inotersen, patisiran, or diflunisal at point of referral. Primary and secondary outcome measures The average time between first presentation and final diagnosis, and the cost per patient per month. By comparing to a more optimal clinical pathway towards diagnosis, we considered what could be the theoretical gain in terms of time to diagnosis and financial savings. Results The average time between first presentation and final diagnosis was 2.74 years. The average cost per patient per month was higher with progressive heart failure symptoms. A hypothetical optimal pathway reduces time to diagnosis of 1.65 to 1.74 years per patient. The potential financial savings are estimated within the range of £3000 to £4800 per patient. Conclusions Patients diagnosed with transthyretin amyloidosis cardiomyopathy have substantial healthcare resource utilization and costs starting from symptom onset. Higher costs were observed with progression in symptoms and appear linked to a delayed diagnosis. The number of additional diagnostic tests and clinical contact points may contribute to this and could represent a path to explore further for important health and cost savings, with more efficient pathways for these patients to be managed.

Higher costs were observed with progression in symptoms and appear linked to a delayed diagnosis. The number of additional diagnostic tests and clinical contact points may contribute to this and could represent a path to explore further for important health and cost savings, with more efficient pathways for these patients to be managed.

| INTRODUCTION
Amyloidosis is a family of diseases, characterized by the pathological deposition of misfolded, insoluble fibril proteins. 1 Transthyretin amyloidosis (ATTR) exists as two subtypes -age related, also known as wild type TTR (ATTRwt), and an inherited form, also known variant ATTR (ATTRv), which differ in age of onset and disease penetrance. 1,2 Deposition of the misfolded proteins in both cases can be multisystemic with extracardiac involvement in the central or peripheral nerves, renal and gastrointestinal systems, with symptoms in these affected areas often predating the cardiac manifestations. 3 This considerable variability in clinical presentations related to the disease necessitates a multidisciplinary approach particularly with close contact between cardiology and neurology specialties to reach an early diagnosis and initiate appropriate management. 4 Myocardial involvement with development of an infiltrative cardiomyopathy is associated with the worse outcomes, and as such, delays in this diagnosis can be detrimental. [5][6][7] Confounding this further, is evidence suggesting the ATTRwt form of cardiomyopathy is not uncommon, being overlooked or misdiagnosed as other conditions, such as hypertensive heart disease and heart failure with preserved ejection fraction. 5,8,9 These patients are at additional risk, as they are likely to receive traditional medical therapy that is either ineffective or are accompanied by adverse outcomes in the presence of this overarching diagnosis. 10 Emerging, targeted therapies hold promise for improving outcomes in those individuals in whom this condition is correctly and timely identified. 11,12 The NHS National Amyloidosis Centre (NAC) is commissioned in the UK for the comprehensive evaluation and management advice for the national caseload of patients with amyloidosis. 13,14 They have provided a useful overview that describes the natural history and outcomes of ATTR cardiomyopathy in the UK. 15,16 Although there is increased awareness for the condition, due to the increased availability of advanced non-invasive imaging techniques, the number of new diagnoses is rapidly expanding and it will be challenging for the NAC alone to manage the growing volume of cases. 15 Compounded by the observation that a significant proportion of patients have multiple admissions and investigations prior to their diagnosis, it is paramount if emerging therapies are on the horizon, that we ensure timely, equitable access for as many patients as possible.
Reconfiguring the care of these patients and collaboration through a network of hub and spoke centers to manage the variable stages of disease progression, may offer a prudent avenue for the NAC to manage these patients. However, gaps still remain discerning the clinical journey towards diagnosis within centers that manage these patients on a day-today basis, and therefore the associated healthcare utilization and economic burden. A closer review of these aspects, particularly in centers able to deliver long-term or complex specialized care in addition to urgent and ambulatory services will further help our understanding and assist in restructuring clinically relevant pathways. This could ultimately help minimize delays to diagnosis, enable more streamlined input via the NAC and importantly, improve patient-centered outcomes for this life-threatening disease.

| OBJECTIVES
The objectives of this study were 2-fold. First, we aimed to understand the healthcare resource use and the attributed costs associated with ATTR cardiomyopathy patients at our center, by mapping and describing their clinical journey pre-and post-diagnosis. Second, we sought to estimate the mean pre-diagnosis cost per patient and explore the hypothesis that a theoretical optimal pathway would have a direct impact on delayed diagnosis and cost accrued pre-diagnosis.
To test this hypothesis, we conducted quantitative analysis of the diagnostic pathway using real-world data.

| Design and context
This is a retrospective, cross-sectional, descriptive study within the King's Health Partners (KHP), an academic health science center based in London. This was predominantly focused on patients assessed at Guy's and St Thomas' hospitals (GSTT), and any subsequent healthcare resource use at King's College Hospital (KCH). We first registered a retrospective audit on both sites that was approved and conducted to evaluate patients seen with a diagnosis of amyloidosis, alongside their investigation and referral process to the NAC.
Clinical information of patients with ATTRwt or ATTRv with cardiac involvement, diagnosed by GSTT (all of whom were referred to the NAC where the diagnosis was confirmed) between 2010 and 2018 were retrospectively evaluated. Diagnosis was based on the 2019 expert consensus recommendations for the suspicion and diagnosis of ATTR cardiac amyloidosis and included Electrocardiogram (ECG), plasma biomarkers, echocardiography, cardiac Magnetic Resonance Imaging (cMRI) and screening for monoclonal proteins. 17 Use of radiotracers was also included, and this was done exclusively at the NAC. Invasive methods included endomyocardial, fat, bladder, rectal, bone marrow, nerve, or salivary gland biopsy with histological testing to confirm the preliminary diagnosis. This was followed by characterization of the amyloid type which always occurred at the NAC.

| Patient and public involvement
No patient/public involvement.

| Inclusion criteria
Patients had to meet all the following criteria to be eligible for inclusion in this study: 1. Have a definitive diagnosis of ATTR cardiomyopathy 2. Adult patients (aged 18 years or older)

Diagnosed between January 2010 and December 2018
Data sources examined at GSTT are summarized in Figure 1.
ATTR cardiomyopathy status was determined by examination of clinical notes, referral letters to NAC and biopsies by one clinician, and a second clinician for data source verification.
We extracted data on all Accident and Emergency (A&E), inpatient, outpatient, and community heart failure contacts through the relevant information systems at GSTT. Some of the cohort (n = 18) also utilized health care resources at KCH. This was manually reviewed via the electronic health record system to identify their activity.

| Exclusion criteria
1. Patients on inotersen, patisiran or diflunisal at point of referral to GSTT.

Amyloid patients without confirmed ATTR cardiomyopathy.
This resulted in the exclusion of two patients, due to the absence of a confirmed ATTR cardiomyopathy diagnosis. No patients were on inotersen, patisiran or diflunisal at point of referral to GSTT.

| Catchment area
Patients within the GSTT catchment area will have received the majority (if not all) of their secondary care at GSTT, and a few also utilized KCH. Initial review of the data suggested that approximately half of the GSTT patients were referred from out of area.

| Referrals to the NAC
All confirmed ATTR cardiomyopathy patients were referred to the NAC. When a presumed diagnosis was made from imaging or biopsies, a referral was made to the NAC. In more recent years, patients have also been directly referred to the NAC without cardiac biopsy if there is strong evidence for cardiac amyloid presence on clinical information and imaging alone. Most referrals were made through the cardiology service; however, some patients may have been seen in the hematology service, and referred directly.

| Variables
The clinical dataset included demographics, clinical contact points at GSTT during the study period and diagnostic investigations utilized toward diagnosis. The New York Heart Association (NYHA) classification of heart failure was used to describe each patient's symptom severity throughout the pathway. This is widely used in clinical practice, uniform, and an accepted classification of heart failure symptoms.
F I G U R E 1 Flow chart summarizing data sources examined and extracted to determine number of transthyretin amyloidosis (ATTR) cardiomyopathy patients. All hospital attendances at Guy's and St Thomas' hospitals (GSTT) are fully coded by the clinical coding team, and entered a data warehouse. This was accessed by our data analyst and any patient coded between 2010 and 2018 with an amyloidosis code were identified. ATTR cardiomyopathy status was determined by examination of clinical notes, referral letters to the National Amyloid Centre (NAC) and biopsies at GSTT. Myocardial biopsy list was provided by the Cardiovascular Quality Improvement and Patient Safety Manager who maintained a database of all cardiac biopsies between 2010 and 2018. Every patient was reviewed using their unique identifier within the electronic health record system to identify the final diagnosis. NAC provided their available referral list from GSTT since 2010. Those with ATTR cardiomyopathy were extracted and crosschecked in the data warehouse search and biopsy lists to confirm that all with ATTR cardiomyopathy were referred to the NAC 3.8 | Mortality data GSTT sent the list of confirmed ATTR cardiomyopathy patients to the Office for National Statistics (ONS), who provided latest mortality status with date of death (where relevant) for each patient.

| Costs
We used NHS reference costs to estimate the aggregated total cost per patient prior and post diagnosis confirmation. This was inclusive of healthcare resource use of outpatient cardiology visits, inpatient admissions, A&E visits, paid-for community care, with echocardiogram and cMRI as diagnostic tests. We did not consider the cost of biopsy and scintigraphy as we assumed all patients were tested equally at the NAC for diagnosis confirmation.

| Economic impact of optimal diagnostic pathway
We hypothesized implementing an optimal pathway may have a positive effect on earlier diagnosis and cost reduction. This analysis consisted of three stages. Firstly, we determined a quantitative definition of an optimal diagnostic pathway using the retrospective data collected. Clinical expertise suggested the minimum number of actions required to determine a diagnosis was an exploratory consultation and investigation upon symptoms onset followed by a confirmatory investigation and consultation before referral to the NAC. We defined a diagnostic cycle as two consultation visits and two tests (or any combination thereof) (see Appendix S1). Consultation options were considered as either outpatient, A&E visits, or inpatient admissions.
Echocardiogram and cMRI were considered as diagnostic tests.
We then estimated the effect of the number of cycles on time and cost prior to diagnosis after adjusting for a number of covariates using generalized linear modeling (GLM) (see Appendix S1 for details). Finally, we employed the recycled predictions, 18 to test the hypothesis of optimal management pre-diagnosis, assuming optimal management is equivalent to having only one diagnostic cycle. We present the potential average savings of time to diagnosis and total pre-diagnosis costs. We performed sensitivity analysis to test violation of parametrical assumptions and to quantify uncertainty around point estimates (see Appendix S1 for details).
Microsoft Excel, R studio and Stata 15 were used for the analyses.

| Study population
We identified 38 patients with a diagnosis of ATTR cardiomyopathy (n = 20; 53% for ATTRwt, and n = 18; 47% for ATTRv). The mean age at diagnosis was 77 ± 8 years and most patients were male (n = 28, 74%), and of African-Caribbean descent (n = 23, 64%). The pathogenic variant, V122I also known as p.V142l, 19 was found in 15 (94%) patients with ATTRv, all of whom were of African-Caribbean descent. One individual with ATTRv who identified as White-British was affected by the V30M mutation, also referred to as p.V50M. 20 The majority of patients with ATTR cardiomyopathy assessed at GSTT were via referral (n = 28, 76%) from primary or secondary care rather than from acute admissions to the trust. The average time between first presentation and final diagnosis was 2.74 years.

| Outcome
At time of symptom onset, most had mild heart failure symptoms (57%). Those who were asymptomatic or had mild heart failure symptoms, had on average, a longer period of time under the care of GSTT cardiovascular service prior to a diagnosis. Sixty percentage of patients died at the end of the study period (n = 22), and the mortality rate was similar between the groups (61% for ATTRv; 60% for ATTRwt). Baseline characteristics and outcome are summarized in

| Costs
Four patients were missing symptom onset date which were dealt with in the costing analysis by using the first known cardiac-related activity for each of those patients. When we calculated the total costs pre-diagnosis, we excluded the four patients without symptom onset date as this was required for each section of this calculation. Only one patient was excluded from the calculation of total costs post diagnosis, as there was insufficient information on their healthcare resource use during the study period. Tables 2 and 3 summarize the costs from symptom onset to diagnosis and from diagnosis to study exit date. The average cost per patient per month was higher with progressive heart failure symptoms, and driven mostly by the inpatient admission costs. Heart failure community nurses' costs were notably higher from diagnosis to study exit period, presumably relating to more heart failure symptoms as the disease progresses.  Table 4 shows the results of the optimal pathway analysis for time to diagnosis and total costs.

| Time to diagnosis
The GLM in the analysis of time to diagnosis provided the same aver-

| Total cost prior to diagnosis
The GLM estimate for total costs was £ 8927 (SE = 1020.31).
Implementing an optimal pathway generates mean total costs of

| DISCUSSION
This study is the first UK based study to examine the relationship between healthcare utilization throughout the patient journey with ATTR cardiomyopathy and the associated economic burden occurring within a local institution outside the NAC.
The predominant route of contact with patients seen at our trust was via referral; a significant contribution from primary care for further specialist review of presumed cardiac symptoms. There was also a contribution from other "spoke" secondary care hospitals, due to Abbreviations: A&E, Accident and Emergency; NYHA, New York Heart Association classification of heart failure.
From diagnosis, there was additional usage of the community heart failure services, and this was partly the source for the higher overall costs following a diagnosis of ATTR cardiomyopathy. This is consistent with the disease profile over time, that this condition is progressive and debilitating, necessitating increased supportive care and careful titration of heart failure medical therapy. 5,7,21 Given that traditional medical therapy in the form of beta blockers and aceinhibitors are sometimes poorly tolerated in patients with ATTR cardiomyopathy, community heart failure nurses are well placed to monitor patients in between local services and the NAC, and to potentially prevent treatment-related setbacks. Nationwide, it is recognized that community heart failure nurses contribute to a substantial reduction in hospital admissions related to heart failure, and the associated overall costs per patient. 24,25 This may suggest, that in areas where this service has already been decommissioned, one might anticipate a considerable upsurge in hospital admissions and its total costs associated with ATTR cardiomyopathy.
Results from a UK-based study by Lane et al, 15  We acknowledge our definition of optimality may be an oversimplification of a rather complex diagnostic journey. Our analysis is not intended to provide clinical recommendations nor to inform on the value for money of specific diagnostic algorithms. Instead, our study explores the potential benefits and optimistic impact of a streamlined pathway that minimizes healthcare resource utilization. Therefore, the results stemming from our findings are only illustrative. Research on the implications of achieving an earlier diagnosis, including costs and cost-effectiveness of novel diagnostic algorithms and therapeutics, is more likely to determine the overall effect within the entire clinical pathway.

| CONCLUSIONS
ATTR cardiomyopathy is associated with increased healthcare resource utilization in conjunction with worsening heart failure symptoms. Inpatient admissions and excess bed days account for a significant proportion of this use and overall costs, with a notable contribution from engagement with community heart failure teams.
By review of the current patient clinical journey, we estimated that a hypothetical optimal pathway could result in considerable time savings to diagnosis of 1.65 to 1.74 years, and total costs pre-diagnosis between £3000 and £4800 per patient. In order to appreciate this, it requires a high index of suspicion in "at-risk" individuals who present T A B L E 4 Optimal pathway analysis for time to diagnosis (years) and mean total costs pre-diagnosis (£)  mostly within primary or secondary care, with selective diagnostic tools available at highly specialized hubs such as our center, to elucidate the etiology early, manage progressive disease and enable timely, efficient links with the NAC. There is room for further research incorporating primary and social care resource use, but also within data from a national registry to confirm our hypothesis.