Abstract
Background
In 2024, the UK National Institute of Health and Care Excellence (NICE) recommended TriageHF alerts as an option for remote monitoring of patients with heart failure (HF) and a compatible cardiac implantable electronic device (CIED). Data on the cost-effectiveness of this approach has not been published. This research evaluates the cost-effectiveness of TriageHF Plus in public hospital settings, using data from the TriageHF Plus multicentre study (758 participants, NCT04177199).
Methods
An economic model was developed to capture the lifetime cost and benefits of TriageHF Plus versus usual care, based on a site of 300 eligible patients. Analysis on individual patient-level data informed model efficacy and resource use parameters. EuroQol five-dimensional questionnaire data and unit costs were obtained from published, peer-reviewed literature and national databases respectively. Costs and benefits were discounted at 3.5% per annum to adjust future costs and benefits to present value.
Results
In a site size of 300, TriageHF Plus was predicted to prevent 384 (363–405) hospitalisations over 5 years. In total, TriageHF Plus saved £3989 (£1812–£5563) per person-lifetime versus usual care and was more effective and cost-saving in 99.4% of simulations. Results were robust to changes in key input parameters. Modelling showed that to avoid one hospitalisation, 1.7 people would need lifetime access to TriageHF Plus.
Conclusion
The TriageHF Plus pathway is cost-effective for the remote monitoring of HF in patients with CIEDs.
Trial registration number
Keywords: Health Care Economics and Organizations; Heart Failure; Telemedicine; Defibrillators, Implantable
WHAT IS ALREADY KNOWN ON THIS TOPIC
The TriageHF Plus remote monitoring clinical pathway is effective at reducing unplanned hospitalisations in patients with heart failure (HF) and a cardiac implantable electronic device. The cost-effectiveness for use in the UK National Health Service has not yet been published.
WHAT THIS STUDY ADDS
TriageHF Plus is not only cost-effective (as defined by the National Institute of Health and Care Excellence), but cost-saving by approximately £4000 per person-lifetime compared with care without device-based remote monitoring.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This analysis demonstrates how adoption of TriageHF Plus can save the NHS money while maintaining safe, effective patient care.
Introduction
Heart failure (HF) is the leading cause of unplanned hospitalisations in the UK, costing approximately £400 million per year.1 Many HF hospitalisations are potentially avoidable if deterioration is identified early.2–5
Patients with HF are often eligible for ‘complex’ cardiac implantable electronic devices (CIED), namely, cardiac resynchronisation therapy with pacemaker (CRT-P) CRT combined with a defibrillator (CRT-D) and implantable cardioverter-defibrillators (ICD).6 There are a variety of multi-parametric algorithm-based remote monitoring systems available, notably in the UK; TriageHF (Medtronic), HeartLogic (Boston Scientific) and HeartInsight (Biotronik). TriageHF is a HF algorithm which leverages routinely monitored health data from cardiac devices to calculate a patient’s risk of HF events in the next 30-days (low, medium, high; this has been described previously).7–9 The more abnormal the physiological parameters, the higher the risk status. ‘High’ risk status triggers a HF alert to be notified to the clinical team via Carelink. TriageHF Plus is a remote HF monitoring pathway for individuals with compatible CIEDs. As visually presented in figure 1, the pathway is triggered by a high-risk transmission (HF alert), signifying elevated risk of HF hospitalisation in the next 30 days.9 10 The remote monitoring platform is monitored by the HF team (usually an HF nurse) on a twice-weekly basis. After reviewing all HF alerts, patients are assessed using a structured phone-call based assessment. Clinical interventions thereafter are at the discretion of the assessor in line with guideline-recommended medical therapy. Prompts are provided to encourage proactive practice, commonly temporary up-titration of diuretics for congestive symptoms, optimisation of guideline-recommended medical therapies or arrangements for further investigations such as blood tests or echocardiogram or face-to-face clinical review (see figure 2). All HF alerts are followed up in 30days with both a repeat transmission and telephone assessment. A repeat 21-day transmission is scheduled for all medium-risk cases.11
Figure 1. Schematic of the TriageHF Plus pathway.
Figure 2. TriageHF Plus clinical pathway - principles of care poster. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor/neprilysin inhibitor; BHF, British Heart Foundation; BP, blood pressure; HF, heart failure; HFrEF, HF with reduced ejection fraction; HFRS, heart failure risk score; MRA, mineralocorticoid receptor antagonists; NHS, National Health Service; NICE, National Institute of Health and Care Excellence; NYHA, New York Heart Association, SR, sinus rhythm.
In 2021, European Society of Cardiology HF guidelines did not recommend routine remote monitoring of HF diagnostics from cardiac devices due to lack of evidence of impact on clinical outcomes, such as hospital admissions and mortality.12 Since then, emerging evidence has changed the landscape of HF remote monitoring, and the subsequent 2023 HRS/EHRA/APHRS/LAHRS Expert Consensus Statement gave a 2 a recommendation to remotely monitor HF diagnostics.13 A recent pooled analysis of multi-sensor CIED-based remote monitoring reported reduced HF hospitalisations;14 however, only TriageHF and HeartLogic have been tested as part of a structured clinical care pathway for worsening HF.15 The TriageHF Plus Study (ClinicalTrials.gov Identifier NCT04177199), published in early 2024, demonstrated that using HF alerts as part of a remote patient management pathway (TriageHF Plus) reduced all-cause hospitalisations between 31–58% compared with usual care without adverse treatment-related effects.16 In response to new evidence, the United Kingdom (UK) National Institute of Health and Care Excellence (NICE) recommended HF alerts (specifically TriageHF Plus and HeartLogic) to be used as an option for people with HF and a compatible CIED in the UK in October 2024.17–19
Although the general practice of remote monitoring of CIEDs has been shown to be cost-effective, the economic value of a HF clinical care pathway driven by multiparametric HF alerts has not been evaluated.20–22
In this research, a health economic model was developed to conduct a cost-effectiveness analysis of the TriageHF Plus pathway using real-world data from the TriageHF Plus study.16
Methods
Study design, setting and participants
TriageHF Plus (NCT04177199) was a real-world, multicentre study of 758 participants across three National Health Service (NHS) sites in Greater Manchester, UK, including two cohorts with patients aged 18+ years, with an existing TriageHF compatible CIED capable of automated transmissions (Medtronic, Minneapolis, Minnesota, USA).16 Patients were recruited regardless of HF status; however, given the nature of the CIED population, the vast majority had a diagnosis of HF. Inclusion criteria were designed to represent a typical cohort if a hospital activated all devices to facilitate remote monitoring. The TriageHF Plus cohort (n=443) was actively recruited where, in addition to usual care, participants underwent remote monitoring following the TriageHF Plus clinical pathway after activation of device alerting capabilities. The time-matched ‘usual care’ cohort (n=315) was identified from hospital records as patients meeting study criteria but who did not undergo device alert activation. To be included in the analysis cohort, a minimum of 90 days of transmission data were required during the time horizon. Follow-up ran from September 2019 until February 2021, mean 433.6 days. Hereafter, this full-time horizon is synonymous with ‘total follow-up’. The primary outcome was the difference in unplanned, all-cause hospital admissions (hospitalisations). Prespecified minimum sample size was 450 patients followed up over at least 12 months; however, recruitment was expanded due to reduced hospitalisation rates during COVID-19.
Usual care was defined as compliance with national HF guidelines and usual CIED follow-up, but without action of HF alerts.6 23 Current NICE guidance recommends remote monitoring for people with HF; therefore, a scenario analysis was conducted for the 91.6% of the study population with HF. As the study spanned the COVID-19 pandemic, to examine the impact of the pandemic on results, a second scenario analysis was conducted using data from the pre-pandemic follow-up period, mean 173.0 days. Hereafter, this is called the ‘pre-COVID-19 follow-up’.16 24
Characteristics for study participants across the total follow-up and the two scenario analyses are presented in table 1.
Table 1. Patient characteristics of the TriageHF Plus study population.
| Total follow-up (n=758) | Pre-COVID-19 follow-up (n=583) | HF population (n=694) | |
|---|---|---|---|
| Age (mean (SD)) | 67.8 (12.7) | 67.7 (13.1) | 68.7 (11.7) |
| Sex (female) (N(%)) | 210 (27.7%) | 171 (29.3%) | 185 (26.7%) |
| Prior hospitalisation* (N(%)) | 144 (19.0%) | 108 (18.5%) | 132 (19.0%) |
| CIED type: CRT-D (N(%)) | 395 (52.1%) | 314 (53.9%) | 386 (55.6%) |
| CIED type: CRT-P (N(%)) | 222 (29.3%) | 182 (31.2%) | 193 (27.8%) |
| CIED type: ICD (N(%)) | 141 (18.6%) | 87 (14.9%) | 115 (16.6%) |
| NYHA status: no HF (N(%)) | 44 (5.8%) | 32 (5.5%) | N/A |
| NYHA status: type I (N(%)) | 155 (20.4%) | 106 (18.2%) | 155 (22.3%) |
| NYHA status: type II (N(%)) | 328 (43.3%) | 258 (44.3%) | 328 (47.3%) |
| NYHA status: type III or IV (N(%)) | 211 (27.8%) | 170 (29.2%) | 211 (30.4%) |
| NYHA status: unknown (N(%)) | 20 (2.6%) | 17 (2.9%) | N/A |
| LVEF >50% (N(%)) | 194 (25.6%) | 159 (27.3%) | 153 (22.0.%) |
| LVEF 35–50% (N(%)) | 287 (37.9%) | 222 (38.1%) | 275 (39.6%) |
| LVEF <35% (N(%)) | 268 (35.4%) | 195 (33.4%) | 260 (37.5%) |
| AF/atrial flutter (N(%)) | 369 (48.7%) | 286 (49.1%) | 340 (49.0%) |
| IHD (N(%)) | 436 (57.5%) | 349 (59.9%) | 420 (60.5%) |
| ACHD (N(%)) | 37 (4.9%) | 30 (5.2%) | 27 (3.9%) |
| Prior AV node ablation (N(%)) | 101 (13.3%) | 81 (13.9%) | 83 (12.0%) |
| Diabetes (N(%)) | 193 (25.5%) | 149 (25.6%) | 184 (26.5%) |
| CKD stage ≥3 (N(%)) | 283 (37.3%) | 220 (37.7%) | 281 (40.5%) |
At least one non-elective hospitalisation in the 6 months prior to enrolment.
ACHD, adult congenital heart disease; AF, atrial fibrillation; ARNI, angiotensin receptor neprolysin inhibitor; AV, atrioventricular; CIED, Cardiac implantable electronic device; CKD, chronic kidney disease; CRT-D, cardiac resynchronisation therapy with defibrillator; CRT-P, cardiac resynchronisation therapy with pacemaker; HF, heart failure; HFrEF, Heart failure with reduced ejection fraction; HFRS, Heart Failure Risk Score; ICD, implanted cardiac defibrillator; IHD, ischaemic heart disease; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association functional classification; SR, sinus rhythm.
Economic model structure
An economic model was conceptualised in accordance with the decision problem (table 2) and developed in adherence with the NICE reference case25 in Microsoft Excel (Microsoft Corporation, Washington, USA). The model had a two-state, regression-based structure which estimated the cost-effectiveness of TriageHF Plus compared with usual care. The cohort entered the model in the ‘alive’ state. A 1-month cycle length was required to capture the 30-day hospitalisation risk predicted by the TriageHF algorithm. In each cycle, a proportion of the cohort remained in the ‘alive’ state, and a proportion moved to the ‘dead’ health state. The proportion of the cohort in each health state at any point in time in the model was informed by survival analysis. The cohort accrued costs and benefits while in the ‘alive’ state over their lifetime. The ‘base case’/primary model was based on a typical hospital site—serving a population of 300 eligible patients for CIED remote monitoring. The probabilistic base case and key scenario analyses used published data from the TriageHF Plus study to inform the population characteristics, the per-month probability of a HF alert and the per-month hospitalisation event probability. Additional scenarios investigated the cost-effectiveness by CIED subgroup (CRT-P; CRT-D; ICD). These were informed by an R-based (http://www.r-project.org) post-hoc analysis of the individual patient-level data (IPD). These analyses used two covariate adjusted regression models to estimate the impact of TriageHF Plus on per-month probabilities of hospitalisations and HF alerts (covariables presented in table 1). A description of the IPD analysis methods and results is presented in the online supplemental file 1 (Section 1).
Table 2. Decision problem and pharmacoeconomic guidelines.
| Element | Decision |
|---|---|
| Model structure | Two-state (alive/dead) regression-based model |
| Populations | A cohort of people with TriageHF-compatible CIED |
| Scenarios |
|
| Intervention | TriageHF Plus |
| Comparator | Usual care |
| Outcomes |
|
| Perspective | England and Wales NHS and PSS |
| Cycle length | One month |
| Time horizon | Lifetime |
| Discount rates | 3.5% for both costs and benefits |
| Cost-effectiveness threshold | £20 000 per QALY gained |
CIED, cardiac implanted electronic device; HF, heart failure; IPD, individual patient data; NHB, net health benefit; NHS, National Health Service; PSS, personal social services.
Considered costs were NHS and personal social services. An annual discount rate of 3.5% was applied to both costs and benefits. Benefits were expressed as quality-adjusted life years (QALYs).
Model parameters
Derivation of inputs for the economic model is described below in the methods section.
Hospitalisations
Regression models, using inverse probability of treatment weighting, were built to compare hospitalisation rates between cohorts. An annual hospitalisation rate of 0.58 in the usual care cohort was observed, equating to a per-month hospitalisation probability of 4.69%, and this was applied to the model. The incident rate ratios are provided in Section 1 of online supplemental file 1. The TriageHF Plus cohort hospitalisation rate was derived by applying the incidence rate ratios (IRRs) to the usual care cohort hospitalisation rate.
The estimated IRR in the CIED analysis across the total follow-up was 0.57 (95% CI 0.40 to 0.81). An exploratory, deterministic, breakeven analysis was conducted to determine the hospitalisation rate IRR at which TriageHF Plus was cost-neutral and cost-effective compared with usual care.
HF alerts
Per-month probability of a HF alert from the TriageHF Plus group was 2.91%. This probability was assumed to be equivalent for the pre-COVID-19 follow-up and the HF population scenario analysis and constant over the cohort’s lifetime. In the CIED scenario analyses, a logistic regression model estimated the probability of a HF alert. Regression equations and per-month probabilities for all scenario analyses are provided in online supplemental file 1 (Section 1).
Mortality
Six conventional parametric regression models were fit to IPD-based time-to-all-cause mortality data. The long-term predictive plausibility of each regression model, presented in the online supplemental file 1 (Section 2), was externally validated using published, peer-reviewed literature and clinical expert opinion. The Weibull model was used in the base case analyses. Mortality rates for the general population were sourced from life tables, compared with the estimates from the Weibull model and applied where appropriate. Due to a paucity of data, equivalent mortality estimates across both cohorts were assumed. The model employed a lifetime horizon (table 2). It was assumed life expectancy could not exceed 100 years. The total time horizon was 32.2 years (386 cycles).
Health-related quality of life
QALY gains were calculated from time spent by the cohort at a certain health-related quality of life (HRQoL). HRQoL is measured in utilities, a value of one represents perfect health and zero is equivalent to death. The model employed a multiplicative approach to calculate QALYs as per NICE guidance.26
Annual EuroQol five-dimensional questionnaire (EQ-5D-3L) utility estimates for each NYHA class were sourced from published literature.27 These were compared with UK-specific age-adjusted and sex-adjusted general population utility norms.28 The baseline population distribution across NYHA classes (table 1) was used to derive a weighted average annual utility decrement for the HF population (−0.109) and the weighted average per-event hospitalisation decrement (−0.070). This hospitalisation decrement was applied in the cycle the hospitalisation occurred (decrements provided in online supplemental file 1, section 3).
Costs and resource use
NHS-specific publicly available sources informed the costs in the model where possible (cost year 2022/2023). A summary of costs is presented in table 3; additional detailed breakdowns are provided in online supplemental file 1 (Section 2). Annual infrastructure costs included access to remote monitoring platforms, brief induction training when required and technical support. Staff training on using the pathway was minimal, as responses to an alert aligned with guideline-recommended management and clinicians retained the discretion to provide usual clinical care in response to HF alert (see figure 2).
Table 3. Costs in the model.
| Scenarios | Source | |||
|---|---|---|---|---|
| Base case | Pre-COVID-19 follow-up | HF population | ||
| Hospitalisation costs | ||||
| All-cause hospitalisation cost (per event) | £2762 | £2762 | £2762 | NHS England data using SUSHRG costing codes. |
| Annual Triage Plus infrastructure costs | ||||
| Annual infrastructure cost of TriageHF Plus per person. | £100 | £100 | £100 | NICE DAP.17 Please note that this is equivalent regardless of site size. |
| Per-cycle monitoring costs for usual care and TriageHF Plus (without HF alerts) cohorts | ||||
| Professional cost | £96.91 | £96.91 | £96.91 | Frequency informed by published literature. Costs were from the PSSRU and NHS reference costs.14 15 |
| Investigation costs | £6.83 | £6.83 | £6.83 | Frequency informed by published literature.6 23 Unit costs from NHS reference costs.14 |
| Drug acquisition costs | £2.87 | £2.90 | £3.02 | The posology and cost of ramipril (ACE inhibitor); carvedilol (beta blocker); spironolactone (MRA) and furosemide (diuretic) were extracted from BNF20 and eMIT.20 The percentage requiring each drug was informed by the TriageHF study. |
| Per-cycle monitoring costs for TriageHF Plus cohort (with HF alert) | ||||
| Drug acquisition costs | £3.32 | £3.24 | £3.38 | The posology and cost of ramipril (ACE inhibitor); carvedilol (beta blocker); spironolactone (MRA) and furosemide (diuretic) were extracted from BNF20 and eMIT.20 The percentage requiring each drug was informed by the TriageHF study. |
| Cost of initial call (52.2%) | £81.00 | £83.88 | £81.00 | The staff time required (by nurse band) per population was informed by the TriageHF study. Cost per staff time was informed by the PSSRU.26 |
| Cost of follow-up calls (43.8%) | £69.79 | £57.87 | £71.59 | |
| Cost of further investigations (7.1%) | £73.46 | £73.46 | £73.46 | Frequency informed by the TRIAGE-HF study and assumptions. Unit costs from NHS reference costs.14 |
| Cost of an outpatient cardiologist visit (10%) | £186.38 | £186.38 | £186.38 | NHS reference costs.14 |
ACE, angiotensin-converting enzyme; BNF, British National Formulary; DAP, Diagnostics Assessment Programme; eMIT, UK Government Drugs and pharmaceutical electronic market information tool; HF, heart failure; MRA, mineralocorticoid receptor antagonists; NHS, National Health Service; NICE, National Institute of Health and Care Excellence; PSSRU, Personal Social Services Research Unit; SUSHRG, Secondary Uses Services Healthcare Resource Group.
Usual care monitoring and drug acquisition costs
Per-cycle clinical investigations and HF-related visits to medical professionals were based on national HF guidelines.6 23 Data from the TriageHF Plus study were used to determine the market share across the main drug classes available at the time of the study (table 3). The per-cycle resource and drug requirements were applied to NHS costs14 15 to derive monitoring and drug acquisition costs. These per-cycle monitoring costs were also accrued by those in the TriageHF Plus cohort without a HF alert.
TriageHF Plus resource use and costs
Infrastructure/remote patient management costs for TriageHF Plus were established at an annual cost of £100 per-person per-year.17
Additional resource costs associated with a HF alert were accrued in the cycle the HF alert occurred. The mean staff time requirements associated with each consultation (initial and 30-day follow-up calls) were captured. Interim consultations only occurred if clinically indicated and were not captured in the model.
It was assumed that those with an investigation would have three blood tests (full blood count, plasma creatinine and cardiac enzyme) plus one of either chest X-ray, ECG or echocardiogram. Healthcare utilisation (HCU) touchpoint analyses and expert opinion informed the percentage of the population who experienced a cardiology outpatient (OP) appointment (10% of the population with HF alerts).
HCU touchpoints
HCU touchpoints captured were hospitalisations, emergency department (ED) attendances and OP cardiology clinics. Lack of established guidance on costings processes prohibited the inclusion of ED attendances in the economic model. Instead, ED attendance rates and costings for TriageHF Plus and usual care were compared in a complementary analysis.
Costings for hospitalisations were obtained using Secondary Uses Services Healthcare Resource Group (SUSHRG) costing codes. Minimum requirements for episode cost extraction were admission date, discharge date and valid SUSHRG code. Where applicable, short stay emergency adjustments tariffs were applied to account for differences in length of stay. Where the SUSHRG codes lacked associated cost, two clinicians (JKT and FZA) selected the most clinically suitable alternative diagnosis using a conservative approach to admission complexity and comorbidity status. The median (IQR) in-trial cost of a hospitalisation was £2366 (£785–£4786) and £3140 (£933–£5036) for the TriageHF Plus and usual care cohorts, respectively. The total median (IQR) hospitalisation cost of £2762 (£826–£4846) was conservatively used to inform both cohorts in the model.
ED attendance and OP data were similarly obtained (detailed episode selection and costings provided in Section 4 of online supplemental file 1).
Model outputs
Outcomes per person over a life-time horizon (32.2 years) are presented in the model results. A probabilistic sensitivity analysis (PSA) with 10 000 simulations was used in the base case to allow results to converge. The PSA captured and quantified uncertainties in the model parameters and assumptions. Each simulation used a different set of values for the uncertain parameters sampled from an appropriate statistical distribution, displayed in the online supplemental file 1 (Section 2).
The PSA estimated the mean (95% uncertainty intervals (UI)) lifetime costs and QALYs for TriageHF Plus and usual care. The 2.5th and 97.5% percentiles were used to estimate the uncertainty limits. Incremental cost and QALYs from each simulation are presented on the cost-effectiveness plane. These outcomes were compared against the NICE threshold of £20 000 per QALY gained to estimate the incremental cost-effectiveness ratio (ICER), net monetary benefit (NMB) and net health benefit.
Patient and public involvement
The TriageHF Plus study was reviewed by the Manchester University NHS Foundation Trust Cardiovascular Patient and Public Involvement panel prior to opening.
Results
The entire study population had a mean age of 67.8 years, 27.7% were female and a mix of CIED types (52.1% CRT-D, 29.3% CRT-P and 18.6% ICD). Per-month hospitalisation probabilities for the TriageHF Plus cohort were 2.00% for the total follow-up and 2.56% for the pre-COVID-19 follow-up.
Base case results
TriageHF Plus was found to be both dominant (increased QALYs and lower costs) and cost-effective (ICER<£20 000) in 99.4% of lifetime probabilistic simulations (presented in table 4), where a lifetime was found to be an average of 5.49 years (discounted) or 6.32 years (undiscounted) from study enrolment.
Table 4. Probabilistic base case and model results (mean, 95% UI).
| Patient-level summary | TriageHF plus | Usual care | Incremental |
|---|---|---|---|
| Total average costs | £11 563 (£9629 to £13 987) |
£15 551 (£14 560 to £16 619) |
–£3989 (–£5563 to –£1812) |
| Total average QALYs | 4.10 (3.77 to 4.32) |
4.09 (3.76 to 4.31) |
0.01 (0.00 to 0.01) |
| ICER (per QALY gained) | –£445 557 | ||
| Probability of being cost-effective (threshold of £20 000 per QALY gained) | 99.4% | ||
| Probability of being cost saving | 99.4% | ||
| Incremental NMB | £4168 (£1909 to £5800) |
||
| Incremental NHB | 0.21 (0.10 to 0.29) |
||
ICER, incremental cost-effectiveness ratio; NHB, net health benefit; NMB, net monetary benefit; QALY, quality-adjusted life years; UI, uncertainty interval.
Lifetime total costs for a person in the TriageHF Plus cohort were estimated to be £11 563 (£9629–£13 987). This was £3989 (£1812–£5563) less than usual care (£15 551 (£14 560–£16 619)). The TriageHF Plus cohort improved HRQoL, with a gain of 0.01 (0.00–0.01) QALYs per person and was dominant over usual care, indicating it is cost-saving and cost-effective. Incremental NMB was estimated at £4168 (£1909–£5800) per QALY gained per lifetime. This finding was consistent across shorter time horizons, indicated by a positive NMB after 1 year and 5 years of monitoring, at £740 (£337–£1034) and £2885 (£1433–£4049), respectively (see table 5 below).
Table 5. Additional scenario analyses: TriageHF Plus versus usual care (2000 probabilistic iterations).
| Parameter | Per-person outcomes: Mean (95% Credible Interval) | Incremental hospitalisations: Mean (95% UI) | |||||
|---|---|---|---|---|---|---|---|
| Total incremental costs | ICER | NMB | Per-person | Site size: 50 | Site size: 100 | Site size: 300 | |
| Base-case parameters* | −£3989 (−£5563 to −£1812) | −£445 557 (dominating) |
£4168 (£1909 to £5800) | −1.97 (−2.62 to −1.07) | −99 (−112 to −85) | −197 (−215 to −179) | −591 (−623 to −559) |
| 1-year time horizon | −£707 (−£991 to −£318) | −£437 962 (dominating) | £740 (£337 to £1034) | −0.31 (−0.41 to −0.17) | −16 (−18 to −13) | −31 (−34 to −28) | −93 (−98 to −88) |
| 5-year time horizon | −£2760 (−£3871 to −£1362) | −£441,500 (dominating) | £2,885 (£1433 to £4049) | −1.28 (−1.71 to −0.73) | −64 (−73 to −55) | −128 (−140 to −116) | −384 (−405 to −363) |
| Pre-COVID-19 | −£2901 (−£4971 to −£130) | −£407,417 Dominating |
£3,043 (£162 to £5199) | −1.54 (−2.39 to −0.37) | −77 (−87 to −67) | −154 (−168 to −140) | −462 (−486 to −438) |
| HF population | −£4220 (−£5858 to −£1876) | −£428 958 (dominating) | £4,417 (£1982 to £6118) | −2.11 (−2.79 to −1.11) | −106 (−119 to −92) | −211 (−231 to −192) | −633 (−667 to −599) |
| CRT-D subgroup | −£2277 (−£3946 to −£805) | −£405 501 (dominating) | £2,389 (£863 to £4119) | −1.24 (−1.93 to −0.63) | −62 (−71 to −53) | −124 (−137 to −111) | −372 (−394 to −350) |
| CRT-P subgroup | −£2410 (−£4361 to −£700) | −£404 581 (dominating) | £2,530 (£754 to £4562) | −1.31 (−2.13 to −0.60) | −66 (−75 to −56) | −131 (−145 to −117) | −393 (−417 to −369) |
| ICD subgroup | −£3441 (−£6412 to −£1066) | −£447 694 (dominating) | £3,595 (£1139 to £6685) | −1.69 (−2.92 to −0.70) | −85 (−97 to −72) | −169 (−187 to −151) | −507 (−538 to −476) |
Total follow-up, total population, lifetime horizon.
CRT-D, cardiac resynchronisation therapy with defibrillator; CRT-P, cardiac resynchronisation therapy with pacemaker; HF, heart failure; ICD, implanted cardiac defibrillator; ICER, incremental cost-effectiveness ratio; NMB, net monetary benefit; UI, uncertainty interval.
The cost-effectiveness plane (figure 3) showed TriageHF Plus saved costs and accrued QALYs in 99.4% of the 10 000 probabilistic simulations, suggesting that TriageHF Plus is dominant compared with usual care. TriageHF Plus was also found to be cost-effective in 99.4% of simulations at a threshold of £20 000 per QALY gained. This finding was consistent across all thresholds of cost-effectiveness, indicated by the cost-effectiveness acceptability curve (online supplemental file 1, Section 5).
Figure 3. Cost-effectiveness plane. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year.
Within the TriageHF cohort, 6.4% of the lifetime healthcare costs were attributed to the pathway infrastructure and resource use costs. This increase in infrastructure cost was offset by the reduction in hospitalisations (1.97 per person). Lifetime hospitalisation costs are displayed in the donut chart (figure 4).
Figure 4. Breakdown of probabilistic costs for TriageHF Plus compared with usual care.
TriageHF Plus infrastructure cost was fixed at £100 per person. Benefits and cost savings estimated increases in proportion with site size. In the base case analysis, of a 300-patient population, TriageHF Plus was estimated to reduce the number of hospitalisations by 93 (88–98) in the first year with an NMB of £740 (£337–£1034) per person. This amounted to 591 (559–623) lifetime hospitalisations avoided for a 300-patient population. To avoid one hospitalisation, costing £2762 (£862–£4856) per event, 1.7 people would need lifetime access to TriageHF Plus.
Scenario analysis results
Key results for the scenario analyses are presented in figure 5 and table 5. NMB remained positive in favour of TriageHF Plus across all scenarios. The greatest per-person lifetime cost-saving and NMB was observed in the HF population, £4417 (£1982–£6118); however, the effect of TriageHF Plus on NMB appeared to be consistent and cost effective regardless of baseline scenario.
Figure 5. Net monetary benefit per person with TriageHF compared with usual care. CRT-D, cardiac resynchronisation therapy with a defibrillator; CRT-P, cardiac resynchronisation therapy device with a pacemaker; HF, heart failure; ICD, implantable cardioverter defibrillator.
Scenario analyses by CIED type used the regression equations derived from the IPD analysis to inform hospitalisation rate and probability of HF alerts. NMB remained positive across all scenarios, and incremental costs remained negative, consistent with the primary results (see figure 5 and table 5). The greatest cost savings and NMB were incurred in the ICD scenario; however, wide UIs suggested high variability.
Exploratory breakeven analysis
In the base case, TriageHF Plus needed to reduce hospitalisation rates by 8.9% (IRR: 0.911) and 8.6% (IRR 0.914) to be cost-saving and cost-effective at a threshold of £20 000 per QALY, respectively. Using base-case parameters, NMB for an average 50-patient site was £208 400 (£194 643–£222 157) and £1 250 400 (£1 216 703–£1 284 097) for a 300-patient site. A full breakdown is provided in the online supplemental file 1 (Section 5).
Complementary HCU analysis
A significant reduction in ED rates in the TriageHF Plus cohort compared with usual care was observed over the total follow-up, IRR 0.80 (0.64–1.00) p=0.046, and pre-COVID-19 follow-up, IRR 0.49 (0.29–0.82), p=0.007. Each ED visit was estimated to cost £165. See online supplemental file 1 (Section 6) for full breakdown.
A total of 4864 cardiology OP clinics occurred during total follow-up. After propensity score adjustment, there was a higher rate of clinics in the TriageHF Plus cohort across total follow-up; IRR 1.47 (1.39–1.56) p<0.001; but a lower rate in the pre-COVID-19 follow-up; IRR 0.82 (0.74–0.91) p<0.001. The impact on the rate of OP appointments is inconclusive.
Discussion
In this prespecified analysis of the TriageHF Plus study, we report that utilisation of TriageHF Plus within the NHS had a 99.4% likelihood of being both cost-effective and cost-saving compared with current usual care. The key driver of cost savings was reduced hospitalisations. TriageHF Plus was estimated to save £3989 per person. Both small and large-sized sites were predicted to benefit from cost savings, with NMB ranging from £208 400 (£194 643–£222 157) for a 50-patient site, to £1 250 400 (£1 216 703–£1 284 097) for a 300-patient site. Adoption of TriageHF Plus could prevent 384 emergency hospitalisations over 5 years for a 300-participant site.
Breakeven analysis suggested TriageHF Plus only needed to reduce hospitalisation rates by 8.6% to be cost-effective, well short of the 43–58% reduction observed in the TriageHF Plus study. The HF subgroup, which accounted for 92% of the population studied, was associated with greatest cost savings, as expected given the intervention primarily focuses on optimising HF care and managing periods of disease instability. This said, TriageHF Plus was found to be cost effective in all scenario analyses and subpopulations, and thus a simple non-restrictive approach to patient selection would be reasonable.
This is the first paper to report on the cost-effectiveness of a device-based remote monitoring clinical pathway driven by multiparametric HF alerts. Both TriageHF and HeartLogic are now recommended by NICE as an option for algorithm-based remote monitoring in people with cardiac devices who have HF.19 Direct comparison of algorithm performance and efficacy is challenging due to differences in populations and outcomes assessed. Although there are differences in the physiological parameters assessed, there remains significant overlap.15 TriageHF Plus offers a major advantage to other systems as it is available across CRT-P, CRT-D and ICD devices and is backwards compatible with many legacy devices.
This analysis comes at a time of great need for cost-effective approaches to delivering HF care. Over the past three decades, the number of hospital beds in the UK has more than halved and continues to decline, a pattern replicated across many global healthcare systems.29 This reduction poses a challenge for HF care, a leading cause of bed occupancy in public health systems. Despite medication and treatment advancements for HF, it is projected that there will be 50% more hospitalisations due to HF by 2035 than in 2010.30 The reality of these projections and trends in disease burden highlights the need for innovative care models. In 2024, the NHS set out a vision to shift from analogue to digital, from sickness to prevention and from hospital to community.31 Although a vision for the NHS, the strategy is universally relevant. Our findings support upfront investment in resources to build capacity in remote monitoring pathways such as TriageHF Plus to benefit both patient and the NHS. This paper, along with the recently published NICE review, would also support the adoption of device-based HF remote monitoring across UK hospital sites.
For patients with CIEDs, remote monitoring technologies can promote safe and effective HF management when integrated into a structured clinical pathway. HF alerts are now recommended as an option for HF patients with a device by NICE.17 The infrastructure to facilitate nationwide adoption of the TriageHF Plus pathway is well-established, the only limitation being clinical staffing and re-organisation of workflows. The TriageHF Plus pathway was designed to be executed by HF specialist nurses; however, with the increasing use of devices to manage HF in the NHS and national shortage of specialist nurses, upskilling of arrhythmia nurses and cardiac physiologists offers a viable alternative for resourcing. New sites may choose to adapt the pathway to their local needs, considering key variables such as geography, local provision of services and staff experience.
Considering the generalisability of cost-savings observed in this analysis to other countries, extrapolation of the magnitude of effect on outcomes and cost-effectiveness should be approached with caution. While published data on HF hospitalisation risk reduction are consistent,2 14 32 factors such as differing healthcare structures, reimbursement models and population risk are likely to impact on the magnitude of savings which can be expected. Nevertheless, the potential for comparable cost savings in other healthcare systems is great, and expansion of infrastructure is relatively straightforward given this is already operational. The main barrier to implementation observed in the UK was the adaptation of working practices within the HF team to receiving HF alerts from the remote monitoring platforms. However, once embedded, most teams opted to retain the practice beyond the evaluation period.33
Limitations
The TriageHF Plus study involved three hospitals across Greater Manchester, a relatively urban geographical area of the UK; therefore, generalisability to other areas would need consideration. The study population may also not be generalisable to all patients with eligible CRT/ICD devices as there was a tendency to recruit patients with a confirmed HF diagnosis, while very frail individuals and those with language barriers may be under-represented. This was a real-world time-matched cohort design, using propensity matching for comparative analysis, which is considered less robust than a randomised control design. The study follow-up period was impacted by the COVID-19 pandemic,16 24 hence inclusion of the pre-COVID-19 scenario analysis to partially mitigate this.
Various assumptions were made to run the economic model due to lack of data availability. Equivalent mortality was applied to both cohorts, based on lack of evidence to demonstrate otherwise. It was assumed the probability of hospitalisation remained constant over a patient’s lifetime, which is unlikely given older patients are more likely to be hospitalised. The impact of HF events on an individual patient’s subsequent long-term risk was not captured and therefore assumed to be nil. This is also unlikely to be true given the detrimental impact of hospitalisations, and both assumptions could be explored in future analyses utilising a larger sample size. These conservative assumptions may have led to an underestimation of the impact observed. Mean utilities for the small non-HF device population were assumed to be equivalent to NYHA class I; therefore, the utility decrements may be marginally overestimated. Statistical assumptions related to survival modelling may have had a confounding effect on the results. Additionally, although both cohorts were similarly affected by stay-at-home guidance during the COVID-19 pandemic, current-era hospitalisation rates are expected to be higher, potentially altering the magnitude of benefit and cost savings. An analysis of contemporaneous data is currently underway to explore this further.
Attempts to provide accurate costings were made; however, where data were lacking, conservative assumptions were used. It was assumed HRQoL utility decrement applied only to the month a hospitalisation occurred in, where the real-life impact of a hospitalisation may be much longer.1 The cost of ED attendances was not included in the model due to lack of robust costing methods; this was unlikely to significantly impact results given small costs observed (approximately £18.05 per patient per year). Costings in this model were based on published NHS reference costings. Enumeration costs and the £100 per patient flat cost of TriageHF Plus may differ from country to country. Future microcosting studies could be undertaken to enhance accuracy.
Conclusion
The TriageHF Plus pathway is cost-effective for the remote monitoring of HF in patients with CIEDs. These economic benefits support the case for adding this remote-based monitoring pathway to standard care for eligible CIED recipients.
Supplementary material
Acknowledgements
We would like to acknowledge the work of all members of the TriageHF Plus team (a full list is provided in the online supplemental material).
Footnotes
Funding: The TriageHF Plus Service Evaluation study was funded by Medtronic, Inc. JKT was funded by the British Heart Foundation (Grant Ref FS/19/34163) and received support as a Peter Mount awardee at Manchester University NHS Foundation Trust.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study involves human participants and was approved by Health Research Authority permission was granted for the TriageHF Plus service evaluation study. For the usual care cohort, the Health Research Authority’s Confidentiality Advisory Group (CAG) granted a confidentiality waiver (Section 251) under the National Health Service Act to link data from electronic health records, cardiac devices and NHS Digital (as outlined above). This study complies with the Declaration of Helsinki. Participants gave informed consent to participate in the study before taking part.
Data availability free text: Data for this study obtained from NHS England are under strict information governance controls. Additional data requests will be dealt with on a case-by-case basis; however, participant level data cannot be shared.
Map disclaimer: The inclusion of any map (including the depiction of any boundaries therein) or of any geographic or locational reference does not imply the expression of any opinion whatsoever on the part of BMJ concerning the legal status of any country, territory, jurisdiction or area or of its authorities. Any such expression remains solely that of the relevant source and is not endorsed by BMJ. Maps are provided without any warranty of any kind, either express or implied.
Patient and public involvement: Patients and/or the public were involved in the design, conduct, reporting or dissemination plans of this research. Refer to the Methods section for further details.
Data availability statement
Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.
References
- 1.McMurray JJV, Trueman D, Hancock E, et al. Cost-effectiveness of sacubitril/valsartan in the treatment of heart failure with reduced ejection fraction. Heart. 2018;104:1006–13. doi: 10.1136/heartjnl-2016-310661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ezer P, Farkas N, Szokodi I, et al. Automatic daily remote monitoring in heart failure patients implanted with a cardiac resynchronisation therapy-defibrillator: a single-centre observational pilot study. Arch Med Sci. 2023;19:73–85. doi: 10.5114/aoms/131958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Health at a glance: Europe 2018: state of the health in the EU cycle. https://www.oecd.org/en/publications.html n.d. Available.
- 4.Gilotra NA, Shpigel A, Okwuosa IS, et al. Patients Commonly Believe Their Heart Failure Hospitalizations Are Preventable and Identify Worsening Heart Failure, Nonadherence, and a Knowledge Gap as Reasons for Admission. J Card Fail. 2017;23:252–6. doi: 10.1016/j.cardfail.2016.09.024. [DOI] [PubMed] [Google Scholar]
- 5.McDonagh TA, Metra M, Adamo M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: Developed by the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) With the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J. 2021;42:3599–726. doi: 10.1093/eurheartj/ehab670. [DOI] [Google Scholar]
- 6.Nicholls MG, Richards AM, Christchurch Cardioendocrine Research Group Disease monitoring of patients with chronic heart failure. Heart. 2007;93:519–23. doi: 10.1136/hrt.2005.078519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ahmed FZ, Sammut-Powell C, Callan P, et al. Abstract 13787: Use of a Novel Cardiac Implantable Electronic Device Digital Care Pathway (TriageHF-Plus) is Associated With a Reduction in All-Cause Unplanned Hospitalizations. Circulation. 2021;144:A13787. doi: 10.1161/circ.144.suppl_1.13787. [DOI] [Google Scholar]
- 8.Ahmed FZ, Sammut-Powell C, Kwok CS, et al. Remote monitoring data from cardiac implantable electronic devices predicts all-cause mortality. EP Europace. 2022;24:245–55. doi: 10.1093/europace/euab160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cowie MR, Sarkar S, Koehler J, et al. Development and validation of an integrated diagnostic algorithm derived from parameters monitored in implantable devices for identifying patients at risk for heart failure hospitalization in an ambulatory setting. Eur Heart J. 2013;34:2472–80. doi: 10.1093/eurheartj/eht083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sammut-Powell C, Taylor JK, Motwani M, et al. Remotely Monitored Cardiac Implantable Electronic Device Data Predict All-Cause and Cardiovascular Unplanned Hospitalization. J Am Heart Assoc. 2022;11:e024526. doi: 10.1161/JAHA.121.024526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ahmed FZ, Taylor JK, Green C, et al. Triage-HF Plus: a novel device-based remote monitoring pathway to identify worsening heart failure. ESC Heart Fail. 2020;7:107–16. doi: 10.1002/ehf2.12529. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jones KC, Weatherly H, Birch S, et al. Unit costs of health and social care 2022 manual. 2022
- 13.Ferrick AM, Raj SR, Deneke T, et al. 2023 HRS/EHRA/APHRS/LAHRS Expert Consensus Statement on Practical Management of the Remote Device Clinic. Europace. 2023;25 doi: 10.1093/europace/euad123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zwaenepoel BAC, Kluft A, Feijen M, et al. Impact of Multisensor CIED-based Heart Failure Monitoring on Mortality, Heart Failure Hospitalizations and Outpatient Visits: A Systematic Review. Curr Heart Fail Rep. 2025;22:21. doi: 10.1007/s11897-025-00707-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Taylor JK, Ahmed FZ. Clinical Pathways Guided by Remotely Monitoring Cardiac Device Data: The Future of Device Heart Failure Management? Arrhythm Electrophysiol Rev . 2023;12:e15. doi: 10.15420/aer.2022.13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ahmed FZ, Sammut-Powell C, Martin GP, et al. Association of a device-based remote management heart failure pathway with outcomes: TriageHF Plus real-world evaluation. ESC Heart Fail. 2024;11:2637–47. doi: 10.1002/ehf2.14821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.National Health Service (NHS) England National cost collection 2021/22. 2023. https://www.england.nhs.uk/publication/2021-22-national-cost-collection-data-publication Available.
- 18.Boehmer JP, Hariharan R, Devecchi FG, et al. A Multisensor Algorithm Predicts Heart Failure Events in Patients With Implanted Devices: Results From the MultiSENSE Study. JACC Heart Fail. 2017;5:216–25. doi: 10.1016/j.jchf.2016.12.011. [DOI] [PubMed] [Google Scholar]
- 19.Heart failure algorithms for remote monitoring in people with cardiac implantable electronic devices. National Institute for Health and Care Excellence (NICE) diagnostics guidance: DG61. 2024 https://www.nice.org.uk/guidance/dg61 Available.
- 20.Ferrick AM, Raj SR, Deneke T, et al. 2023 HRS/EHRA/APHRS/LAHRS expert consensus statement on practical management of the remote device clinic. Heart Rhythm. 2023;20:e92–144. doi: 10.1016/j.hrthm.2023.03.1525. [DOI] [PubMed] [Google Scholar]
- 21.de Graaf G, Timmermans I, Meine M, et al. Economic evaluation of remote monitoring of patients with an implantable cardiac defibrillator (REMOTE-CIED study) J Telemed Telecare. 2024;30:1173–85. doi: 10.1177/1357633X221129176. [DOI] [PubMed] [Google Scholar]
- 22.Burri H, Sticherling C, Wright D, et al. Cost-consequence analysis of daily continuous remote monitoring of implantable cardiac defibrillator and resynchronization devices in the UK. Europace . 2013;15:1601–8. doi: 10.1093/europace/eut070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Abraham WT, Compton S, Haas G, et al. Intrathoracic impedance vs daily weight monitoring for predicting worsening heart failure events: results of the Fluid Accumulation Status Trial (FAST) Congest Heart Fail. 2011;17:51–5. doi: 10.1111/j.1751-7133.2011.00220.x. [DOI] [PubMed] [Google Scholar]
- 24.Taylor JK, Ndiaye H, Daniels M, et al. Lockdown, slow down: impact of the COVID-19 pandemic on physical activity-an observational study. Open Heart. 2021;8:e001600. doi: 10.1136/openhrt-2021-001600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.National Institute of Health and Care Excellence. NICE health technology evaluations: the manual. NICE process and methods. PMG36. 2022. https://www.nice.org.uk/process/pmg36/ Available.
- 26.Department of Health and Social Care eMIT national database. 2011. https://www.gov.uk/government/publications/drugs-and-pharmaceutical-electronic-market-information-emit Available.
- 27.Griffiths A, Paracha N, Davies A, et al. The cost effectiveness of ivabradine in the treatment of chronic heart failure from the U.K. National Health Service perspective. Heart. 2014;100:1031–6. doi: 10.1136/heartjnl-2013-304598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Alava MH, Pudney S, Wailoo A. NICE DSU report; 2022. Estimating EQ-5D by age and sex for the UK (PDF, 428kb)https://www.sheffield.ac.uk/nice-dsu/methods-development/estimating-eq-5d Available. [Google Scholar]
- 29.The Kings fund website. [03-Feb-2025]. https://www.kingsfund.org.uk/insight-and-analysis/data-and-charts/number-hospital-beds Available. Accessed.
- 30.Heidenreich PA, Albert NM, Allen LA, et al. Forecasting the Impact of Heart Failure in the United States. Circ: Heart Failure. 2013;6:606–19. doi: 10.1161/HHF.0b013e318291329a. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Policy Paper from Department of Health and Social Care; 2025. Road to recovery: the government’s 2025 mandate to NHS England.https://www.gov.uk/government/publications/road-to-recovery-the-governments-2025-mandate-to-nhs-england/road-to-recovery-the-governments-2025-mandate-to-nhs-england Available. [Google Scholar]
- 32.Treskes RW, Beles M, Caputo M-L, et al. Clinical and economic impact of HeartLogic™ compared with standard care in heart failure patients. ESC Heart Fail. 2021;8:1541–51. doi: 10.1002/ehf2.13252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Case studies published alongside National Institute for Health and Care Excellence (NICE) diagnostics guidance: DG61: heart failure algorithms for remote monitoring in people with cardiac implantable electronic devices. [05-Sep-2025]. https://www.nice.org.uk/guidance/dg61/resources Available. Accessed.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.