Developing Markov Models From Real-World Data: A Case Study of Heart Failure Modeling Using Administrative Data

Economic Burden and Healthcare Trajectories of Patients Awaiting Heart Transplantation in a French Tertiary Center

Heart transplantation (HT) is the gold standard treatment of end-stage heart failure, but organ shortage remains a challenge. This retrospective cohort study assesses the economic burden and healthcare pathways of patients awaiting HT in a French tertiary center. Direct healthcare resources were collected and valued, and a state sequence analysis was performed. Ninety-two adult patients were included, with 67 (73%) undergoing HT within a median waiting time of 2 months. The mean cost per patient was €21,324.05 with an average of 2.71 hospitalizations. Four clusters were identified. Type 1 patients (n = 43) underwent HT within 1 month, with a mean cost of €5,820.12 per patient. Only 4 (25%) Type 2 patients (n = 16) underwent HT within 30 months, as they were not prioritized for HT, with a mean cost of €22,285.32 per patient. Type 3 patients (n = 20) underwent HT within 10 months, but incurred higher costs (€27,541.11) compared to Type 2 patients over a shorter period. Despite high transplant priority, Type 4 patients (n = 13) died before HT within 3 months, with a mean cost of €61,858.45 and 3 hospitalizations. This work highlights the economic burden of organ shortage. The use of novel heart preservation devices (such as ex-vivo perfusion systems) could help to expand the donor pool and alleviate this burden, but these aspects need to be further investigated.

Health Economic Evaluations of Circulating Tumor DNA Testing for Cancer Screening: Systematic Review

BACKGROUND

Cancer detection remains a significant global healthcare challenge, and circulating tumor DNA (ctDNA) is a biomarker for noninvasive cancer screening.

OBJECTIVE

This systematic review aimed to describe health economic evaluations of ctDNA for cancer screening.

METHODS

A comprehensive literature search was performed (following PRISMA guidelines) across MEDLINE, Embase, APA PsycINFO, Cochrane Library, Web of Science, and the Center for Review and Dissemination. The review included full-scale health economic analyses such as cost-effectiveness, cost-utility, cost-benefit, and cost-consequence analyses. The quality of the included reports was assessed using CHEERS 2022 standards, and each report was categorized as excellent, very good, good, or insufficient.

RESULTS

Eighteen studies were selected, including four ctDNA tests (EBV-DNA, cf-DNA, mSEPT9, and mt-sDNA) for three types of cancer screening: nasopharyngeal carcinoma (NPC) (2; 11.11%), breast cancer (BC) (1; 5.56%), and colorectal cancer (CRC) (15; 83.33%). Five studies (27.78%) found ctDNA cost-effective for CRC screening (mt-sDNA (with higher uptake than conventional tests) versus fecal immunochemical testing (FIT) or colonoscopy (n = 4); mSEPT9 versus computed tomography colonoscopy (CTC) (n = 1)). Thirteen studies (72.22%) found ctDNA not cost-effective for NPC (EBV-DNA versus no screening (n = 2)); BC (cf-DNA versus conventional testing (n = 1)); CRC (mSEPT9 versus FIT or colonoscopy (n = 2)); mt-sDNA versus FIT or colonoscopy (n = 5); mSEPT9 or mt-sDNA versus conventional tests (n = 3)). The CHEERS assessment found all reports in the "very good" category.

CONCLUSION

All ctDNA tests were generally not cost-effective comparing to conventional screening methods, except when the mt-sDNA uptake was higher than the comparators or when mSEPT9 was compared with CTC.

TRIAL REGISTRATION

CRD42023477732.

Economic Evaluations of Digital Health Interventions for Patients With Heart Failure: Systematic Review

BACKGROUND

Digital health interventions (DHIs) have shown promising results in enhancing the management of heart failure (HF). Although health care interventions are increasingly being delivered digitally, with growing evidence on the potential cost-effectiveness of adopting them, there has been little effort to collate and synthesize the findings.

OBJECTIVE

This study's objective was to systematically review the economic evaluations that assess the adoption of DHIs in the management and treatment of HF.

METHODS

A systematic review was conducted using 3 electronic databases: PubMed, EBSCOhost, and Scopus. Articles reporting full economic evaluations of DHIs for patients with HF published up to July 2023 were eligible for inclusion. Study characteristics, design (both trial based and model based), input parameters, and main results were extracted from full-text articles. Data synthesis was conducted based on the technologies used for delivering DHIs in the management of patients with HF, and the findings were analyzed narratively. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines were followed for this systematic review. The reporting quality of the included studies was evaluated using the CHEERS (Consolidated Health Economic Evaluation Reporting Standards) guidelines.

RESULTS

Overall, 27 economic evaluations were included in the review. The economic evaluations were based on models (13/27, 48%), trials (13/27, 48%), or a combination approach (1/27, 4%). The devices evaluated included noninvasive remote monitoring devices (eg, home telemonitoring using digital tablets or specific medical devices that enable transmission of physiological data), telephone support, mobile apps and wearables, remote monitoring follow-up in patients with implantable medical devices, and videoconferencing systems. Most of the studies (24/27, 89%) used cost-utility analysis. The majority of the studies (25/27, 93%) were conducted in high-income countries, particularly European countries (16/27, 59%) such as the United Kingdom and the Netherlands. Mobile apps and wearables, remote monitoring follow-up in patients with implantable medical devices, and videoconferencing systems yielded cost-effective results or even emerged as dominant strategies. However, conflicting results were observed, particularly in noninvasive remote monitoring devices and telephone support. In 15% (4/27) of the studies, these DHIs were found to be less costly and more effective than the comparators (ie, dominant), while 33% (9/27) reported them to be more costly but more effective with incremental cost-effectiveness ratios below the respective willingness-to-pay thresholds (ie, cost-effective). Furthermore, in 11% (3/27) of the studies, noninvasive remote monitoring devices and telephone support were either above the willingness-to-pay thresholds or more costly than, yet as effective as, the comparators (ie, not cost-effective). In terms of reporting quality, the studies were classified as good (20/27, 74%), moderate (6/27, 22%), or excellent (1/27, 4%).

CONCLUSIONS

Despite the conflicting results, the main findings indicated that, overall, DHIs were more cost-effective than non-DHI alternatives.

TRIAL REGISTRATION

PROSPERO CRD42023388241; https://tinyurl.com/2p9axpmc.

Markov modeling for cost-effectiveness using federated health data network

OBJECTIVE

To introduce 2 R-packages that facilitate conducting health economics research on OMOP-based data networks, aiming to standardize and improve the reproducibility, transparency, and transferability of health economic models.

MATERIALS AND METHODS

We developed the software tools and demonstrated their utility by replicating a UK-based heart failure data analysis across 5 different international databases from Estonia, Spain, Serbia, and the United States.

RESULTS

We examined treatment trajectories of 47 163 patients. The overall incremental cost-effectiveness ratio (ICER) for telemonitoring relative to standard of care was 57 472 €/QALY. Country-specific ICERs were 60 312 €/QALY in Estonia, 58 096 €/QALY in Spain, 40 372 €/QALY in Serbia, and 90 893 €/QALY in the US, which surpassed the established willingness-to-pay thresholds.

DISCUSSION

Currently, the cost-effectiveness analysis lacks standard tools, is performed in ad-hoc manner, and relies heavily on published information that might not be specific for local circumstances. Published results often exhibit a narrow focus, central to a single site, and provide only partial decision criteria, limiting their generalizability and comprehensive utility.

CONCLUSION

We created 2 R-packages to pioneer cost-effectiveness analysis in OMOP CDM data networks. The first manages state definitions and database interaction, while the second focuses on Markov model learning and profile synthesis. We demonstrated their utility in a multisite heart failure study, comparing telemonitoring and standard care, finding telemonitoring not cost-effective.

Cost-effectiveness analysis of digital therapeutics for home-based cardiac rehabilitation for patients with chronic heart failure: model development and data analysis

BACKGROUND

In recent years, numerous guidelines and expert consensus have recommended the inclusion of digital technologies and products in cardiac rehabilitation. Digital therapeutics (DTx) is an evidence-based medicine that uses digital means for data collection and monitoring of indicators to control and optimize the treatment, management, and prevention of disease.

OBJECTIVE

This study collected and reviewed real-world data and built a model using health economics assessment methods to analyze the potential cost-effectiveness of DTx applied to home-based cardiac rehabilitation for patients with chronic heart failure. From the perspective of medical and health decision-makers, the economic value of DTx is evaluated prospectively to provide the basis and reference for the application decision and promotion of DTx.

METHODS

Markov models were constructed to simulate the outcomes of DTx for home-based cardiac rehabilitation (DT group) compared to conventional home-based cardiac rehabilitation (CH group) in patients with chronic heart failure. The model input parameters were clinical indicators and cost data. Outcome indicators were quality-adjusted life years (QALYs) and incremental cost-effectiveness ratios (ICERs). The robustness of the evaluation methods and results was tested using sensitivity analyses. Clinical indicators, cost data, and health utility values were obtained from real-world data, including clinical study data, published literature, and public website information.

RESULTS

The Markov model simulated a time span of 10 years, with a cycle set at one month, for 120 cycles. The results showed that the per capita cost of the CH group was 38,442.11 CNY/year, with a QALY of 0.7196 per person per year. The per capita cost of the DT group was 42,300.26 CNY/year, with a QALY of 0.81687 per person per year. The ICER per person was 39,663.5 CNY/QALY each year, which was below the willingness-to-pay threshold of 85,698 CNY (China's GDP per capita in 2022).

CONCLUSIONS

DTx for home-based cardiac rehabilitation is an extremely cost-effective rehabilitation option compared with conventional home-based cardiac rehabilitation. DTx for home-based cardiac rehabilitation is potentially valuable from the perspective of healthcare decision-makers.

The cost-effectiveness of pegcetacoplan in complement treatment-naïve adults with paroxysmal nocturnal hemoglobinuria in the USA

Aim: Paroxysmal nocturnal hemoglobinuria (PNH) is a rare blood disorder characterized by hemolytic anemia, bone marrow failure and thrombosis, and is associated with high healthcare burden. We evaluated the cost-effectiveness of pegcetacoplan, a proximal complement-3 inhibitor (C3i), compared with the C5i, eculizumab and ravulizumab, in complement treatment-naive adults with PNH, from the US healthcare payer perspective. Materials & methods: A de novo cost-effectiveness model based on a Markov cohort structure evaluated lifetime (55-year) PNH costs and outcomes. The 6-month cycles of the model reflected the follow-up period of PRINCE (NCT04085601), an open-label trial of pegcetacoplan compared with eculizumab in C5i-naive patients. Data from PRINCE informed the clinical, safety and health-related quality of life outcomes in the model. Results: Pegcetacoplan was associated with lifetime cost savings of USD1,176,808 and USD213,062 relative to eculizumab and ravulizumab, respectively (largely attributed to reduced drug costs and blood transfusions), and additional quality-adjusted life years (QALYs) of 0.25 and 0.24. Conclusion: In patients with PNH who are treatment-naive, the base-case cost-effectiveness analysis, scenario analysis and sensitivity analysis showed both lifetime cost savings and increased QALYs associated with pegcetacoplan compared with eculizumab or ravulizumab in the USA.

A Proposal of a Cost-Effectiveness Modeling Approach for Heart Failure Treatment Assessment: Considering the Short- and Long-Term Impact of Hospitalization on Event Rates

BACKGROUND

The rate of events such as recurrent heart failure (HF) hospitalization and death are known to dramatically increase directly after HF hospitalization. Furthermore, the number of HF hospitalizations is associated with irreversible long-term disease progression, which is in turn associated with increased event rates. However, cost-effectiveness models of HF treatments commonly fail to capture both the short- and long-term association between HF hospitalization and events.

OBJECTIVE

The aim of this study was to provide a decision-analytic model that reflects the short- and long-term association between HF hospitalization and event rates. Furthermore, we assess the impact of omitting these associations.

METHODS

We developed a life-time Markov cohort model to evaluate HF treatments, and modeled the short-term impact of HF hospitalization on event rates via a sequence of tunnel states, with transition probabilities following a parametric survival curve. The corresponding long-term impact was modeled via hazard ratios per HF hospitalization. We obtained baseline event rates and utilities from published literature. Subsequently, we assessed, for a hypothetical HF treatment, how omitting the modeled associations (through a simple two-state model) affects incremental quality-adjusted life-years (QALYs).

RESULTS

We developed a model that incorporates both short- and long-term impacts of HF hospitalizations. Based on an assumed treatment effect of a 20% risk reduction for HF hospitalization (and associated reductions in all-cause mortality of 15%), omitting the short-term, the long-term, or both associations resulted in a 5%, 1%, and 22% decrease in QALYs gained, respectively.

CONCLUSION

For both modeling components, i.e., the short- and long-term implications of HF hospitalization, the impact on incremental outcomes associated with treatment was substantial. Considering these aspects as proposed within this modeling approach better reflects the natural course of this progressive condition and will enhance the evaluation of future HF treatments.

Dynamic risk stratification using Markov chain modelling in patients with chronic heart failure

AIMS

Risk changes with the progression of disease and the impact of treatment. We developed a dynamic risk stratification Markov chain model using artificial intelligence in patients with chronic heart failure (CHF).

METHODS AND RESULTS

We described the pattern of behaviour among 7496 consecutive patients assessed for suspected HF. The following mutually exclusive health states were defined and assessed every 4 months: death, hospitalization, outpatient visit, no event, and leaving the service altogether (defined as no event at any point following assessment). The observed figures at the first transition (4 months) weres 427 (6%), 1559 (21%), 2254 (30%), 1414 (19%), and 1842 (25%), respectively. The probabilities derived from the first two transitions (i.e. from baseline to 4 months and from 4 to 8 months) were used to construct the model. An example of the model's prediction is that at cycle 4, the cumulative probability of death was 14%; leaving the system, 37%; being hospitalized between 12 and 16 months, 10%; having an outpatient visit, 8%; and having no event, 31%. The corresponding observed figures were 14%, 41%, 10%, 15%, and 21%, respectively. The model predicted that during the first 2 years, a patient had a probability of dying of 0.19, and the observed value was 0.18.

CONCLUSIONS

A model derived from the first 8 months of follow-up is strongly predictive of future events in a population of patients with chronic heart failure. The course of CHF is more linear than is commonly supposed, and thus more predictable.

Economic Evaluations of Remote Patient Monitoring for Chronic Disease: A Systematic Review

OBJECTIVES

This study aimed to systematically review and summarize economic evaluations of noninvasive remote patient monitoring (RPM) for chronic diseases compared with usual care.

METHODS

A systematic literature search identified economic evaluations of RPM for chronic diseases, compared with usual care. Searches of PubMed, Embase, CINAHL, and EconLit using keyword synonyms for RPM and economics identified articles published from up until September 2021. Title, abstract, and full-text reviews were conducted. Data extraction of study characteristics and health economic findings was performed. Article reporting quality was assessed using the Consolidated Health Economic Evaluation Reporting Standards checklist.

RESULTS

This review demonstrated that the cost-effectiveness of RPM was dependent on clinical context, capital investment, organizational processes, and willingness to pay in each specific setting. RPM was found to be highly cost-effective for hypertension and may be cost-effective for heart failure and chronic obstructive pulmonary disease. There were few studies that investigated RPM for diabetes or other chronic diseases. Studies were of high reporting quality, with an average Consolidated Health Economic Evaluation Reporting Standards score of 81%. Of the final 34 included studies, most were conducted from the healthcare system perspective. Eighteen studies used cost-utility analysis, 4 used cost-effectiveness analysis, 2 combined cost-utility analysis and a cost-effectiveness analysis, 1 used cost-consequence analysis, 1 used cost-benefit analysis, and 8 used cost-minimization analysis.

CONCLUSIONS

RPM was highly cost-effective for hypertension and may achieve greater long-term cost savings from the prevention of high-cost health events. For chronic obstructive pulmonary disease and heart failure, cost-effectiveness findings differed according to disease severity and there was limited economic evidence for diabetes interventions.

Risk of fatal and nonfatal coronary heart disease and stroke events among adult patients with hypertension: basic Markov model inputs for evaluating cost-effectiveness of hypertension treatment: systematic review of cohort studies

Objectives

Hypertension is a risk factor for a number of vascular and cardiac complications. A Markov like simulation based on cardiovascular disease (CVD) policy model is being used for evaluating cost-effectiveness of hypertension treatment. Stroke, angina, myocardial infarction (MI), cardiac arrest and all-cause mortality were only included CVD outcome variables in the model. Therefore this systematic review was conducted to evaluate completeness of CVD policy model for evaluation of cost-effectiveness across different regions.

Key findings

Fourteen cohort studies involving a total of 1 674 773 hypertensive adult population and 499 226 adults with treatment resistant hypertension were included in this systematic review. Hypertension is clearly associated with coronary heart disease (CHD) and stroke mortality, unstable angina, stable angina, MI, heart failure (HF), sudden cardiac death, transient ischemic attack, ischemic stroke, sub-arachnoid hemorrhage, intracranial hemorrhage, peripheral arterial disease (PAD), and abdominal aortic aneurism (AAA). Lifetime risk of developing HF is higher among hypertensives across all ages, with slight variation among regions. Treatment resistant hypertension is associated with higher relative risk of developing major CVD events and mortality when compared with the non-resistant hypertension.

Summary

The CVD policy model can be used in most of the regions for evaluation of cost-effectiveness of hypertension treatment. However, hypertension is highly associated with HF in Latin America, Eastern Europe, and Sub-Saharan Africa. Therefore, it is important to consider HF in CVD policy model for evaluating cost-effectiveness of hypertension treatment in these regions. We do not suggest the inclusion of PAD and AAA in CVD policy model for evaluating cost-effectiveness of hypertension treatment due to lack of sufficient evidence. Researchers should consider the effect of treatment resistant hypertension either through including in the basic model or during setting the model assumptions.