FormalPara Key Points for Decision Makers

The traditional approach to health technology assessment often overlooks the impacts of healthcare interventions on broader society beyond the healthcare sector.

The proposed conceptual framework incorporates a conventional cost-effectiveness analysis, macroeconomic approaches, and a voting scheme to capture and evaluate the broader economic and societal impacts of healthcare interventions.

The application of the framework provides a comprehensive assessment of healthcare interventions, enabling policymakers to make more informed decisions regarding reimbursement and investment in healthcare interventions.

1 Introduction

The link between economy-wide prosperity and health is increasingly recognized [1]. Compelling evidence from studies [2, 3] and the World Health Organization (WHO) report [4] highlight the reciprocal relationship between economic development and health outcomes. Health plays a crucial role in human capital, contributing to economic growth and development [5]. A healthy population is more productive, with higher labor force participation and income generation. Consequently, this leads to increased tax revenues and greater government spending on social programs [6]. Thus, healthcare interventions are of critical importance in maintaining a healthy population, promoting economic growth, and enhancing societal well-being.

Health technology assessments (HTAs) are of paramount importance in determining the value of healthcare interventions. They are used to assess the benefits and costs of new technologies, including pharmaceuticals, medical devices, and procedures, with the aim of informing decisions about their use and reimbursement [7]. Healthcare interventions may have impacts beyond direct health outcomes, such as changes in productivity and social welfare. As a result, decision makers using traditional HTA approaches may overlook significant benefits or harms associated with healthcare interventions that extend beyond the healthcare system itself [8]. For instance, a treatment that reduces disability and improves productivity among patients can generate economic gains by enhancing workforce participation and reducing long-term disability costs. Neglecting such economic impacts in HTA evaluations may undervalue the true value of interventions and hinder informed decision making [9]. Additionally, healthcare interventions have profound societal impacts beyond the individual patient. For example, interventions that improve population health can lead to reduced healthcare inequalities, increased social cohesion, and improved overall well-being [10]. These broader societal implications are often overlooked in traditional HTA frameworks, limiting the ability to assess the true societal value of healthcare interventions. Yet, current HTA frameworks fail to fully capture the impacts of healthcare interventions on broader society beyond the healthcare sector [11,12,13]. Furthermore, the limited scope of HTAs may result in an unequal distribution of benefits and harms across different population groups, leading to concerns about equity [14, 15]. This gap in the literature points to the critical need for novel frameworks that consider the broader economic and societal impacts of healthcare interventions beyond the health sector. Such frameworks would need to capture the interactions between healthcare interventions and the fabric of the economy such that economic and societal ripple effects are fully estimated. By doing so, decision makers can make more informed choices that align with societal priorities and promote equitable access to effective healthcare interventions.

In this manuscript, we document examples in which available evidence stemming from current HTA does not traditionally account for the impacts of healthcare interventions on broader society, beyond the healthcare sector. We then present a novel conceptual framework for the value assessment of healthcare interventions. The framework accounts for direct health outcomes, as well as ripple effects beyond the healthcare sector.

2 Current Health Technology Assessment (HTA) and its Limitations

HTAs play an essential role in assessing the value through quantifying their impacts on the healthcare and non-healthcare sectors of our society. However, it is crucial to recognize limitations of traditional HTA approaches in capturing the full spectrum of societal effects that extend beyond the immediate healthcare context. While current HTA approaches excel in assessing clinical effectiveness and cost effectiveness within the healthcare system, they often overlook the significant economic and productivity impacts on the broader economy. The introduction of new medical technologies can stimulate job creation, enhance productivity, and foster economic growth in related sectors, yet these broader economic impacts frequently escape the traditional HTA analysis framework [16]. This oversight underscores a critical gap in evaluating the true value of healthcare interventions.

Moreover, the traditional focus of HTAs on clinical outcomes and economic factors tends to sideline the profound social and ethical implications of health technologies. Interventions such as genetic testing and gene therapies introduce complex ethical dilemmas, including issues related to privacy, discrimination, and equity [17]. The broader social repercussions of these technologies, such as their influence on social inequalities and cultural norms, are often insufficiently considered in HTAs that prioritize clinical and economic metrics above all.

Environmental sustainability is another aspect frequently neglected in HTAs. The lifecycle of medical devices and pharmaceuticals—from production to disposal—poses significant environmental risks, including pollution and resource depletion [18]. A comprehensive evaluation of healthcare interventions demands consideration of their environmental footprint, encompassing carbon emissions, waste production, and overall ecological impact, factors traditionally excluded from HTA methodologies [19].

The reliance of HTAs on short-term, clinical trial-based evidence further limits their ability to grasp the long-term and systemic effects of healthcare interventions. The broader implications of new technologies, such as their impact on healthcare delivery, patient/provider dynamics, and the healthcare system at large, may only become apparent over time and are often overlooked in the HTA process [20]. This shortcoming highlights the need for an expanded evaluation framework that can capture a larger range of impacts.

Lastly, the reach of healthcare interventions extends well beyond the healthcare sector, touching upon areas such as education, transportation, and urban planning. Therefore, assessing the comprehensive societal impacts of healthcare innovations calls for intersectoral collaboration, a crucial element not consistently integrated into traditional HTA practices [21]. This gap underscores the necessity of adopting a more holistic approach to HTA, one that fully appreciates the multifaceted impacts of healthcare interventions.

In summary, while HTAs are invaluable for assessing healthcare interventions, their traditional methodologies fall short in several key areas. Expanding the scope of current HTAs to encompass economic, social, ethical, environmental, and systemic factors will provide a deeper understanding of healthcare innovations’ true value. For example, a more nuanced and comprehensive HTA will allow us to (1) better allocate the novel health technology to those most in need; and (2) provide incentives for continual innovation ensuring dynamic efficiency. Such an expanded approach would not only enhance the relevance of HTAs but also support more informed and equitable healthcare decision making.

3 Conceptual Framework

This section presents a multipronged approach to the assessment of technology that builds on (1) conventional cost effectiveness and inclusive of patient-centeredness and equity (microeconomic and qualitative approaches combined), followed by (2) a macroeconomic analysis of the intervention’s impacts, and (3) a voting scheme consisting of a combination of the outputs of both (1) and (2) to create decision rules for the adoption of the intervention like the four quadrants of the cost-effectiveness plane (Fig. 1).

Fig. 1
figure 1

Proposed conceptual framework

Full size image

3.1 A Conventional Cost-Effectiveness Analysis Augmented with Patient Centricity and Equity: Distributional Cost-Effectiveness Analysis (DCEA)

Distributional cost-effectiveness analysis (DCEA) enhances traditional cost-effectiveness analysis (CEA) by incorporating equity considerations, assessing how healthcare interventions impact different population subgroups and their distributional effects [22]. DCEA aims to uncover the effects of interventions on health inequalities, determining whether benefits are equitably distributed or exacerbate disparities. Implementing DCEA involves several steps, which are explained in detail elsewhere [23]. Briefly, we present steps conducive to a successful implementation.

Step 1: Defining the relevant groups of interest: This step starts with identifying relevant population groups by geography, age, or socioeconomic status [24], ensuring data availability for each. The disease’s prevalence or incidence in each group is then measured, followed by evaluating the intervention’s effectiveness in improving health outcomes, measured in quality-adjusted life-expectancy (QALE) for each subgroup.

Step 2: Establishing the baseline QALE: Establishing the baseline QALE for each socioeconomic group is the cornerstone of DCEA. This step involves collecting data on life expectancy and adjusting it for quality of life, typically derived from health surveys, life tables, and disease registries. The quality adjustment reflects the utility value of different health states, which can be gathered from the literature or through population surveys. Researchers must ensure the data are representative of the different predefined subgroups in Step 1 to accurately reflect the baseline health disparities. This step sets the stage for the analysis by providing a snapshot of the health status quo before any intervention (pre-intervention).

Step 3: Determining the average QALE: Determining the average QALE across groups is a critical step that sets the stage for identifying and quantifying health inequalities. The average QALE is calculated across all socioeconomic groups’ prespecified subgroups in Step 1 to establish the reference point against which disparities will be measured. This average serves as a benchmark to identify the extent of baseline inequalities in health outcomes.

Step 4: Applying the inequality aversion parameter (α): The inequality aversion parameter (α) is applied using an exponential function to the difference between individual QALEs and the average QALE [22]. This quantifies societal discomfort with inequality, where a higher exponential value indicates greater aversion, particularly for disadvantaged groups. This step requires an understanding of how to model preferences and may involve sensitivity analyses to explore how different values of α affect the results.

Step 5: Introducing the interventions: This step involves the application of health interventions to the model to evaluate their impact on QALE. For each socioeconomic group, we reassess QALE post-intervention. This requires detailed intervention data on effectiveness and utilization across different groups, often sourced from clinical trials, observational studies, and healthcare databases.

Step 6: Reassessing inequality with the Kolm Index: Post-intervention, inequality is reassessed using the Kolm Index, which accounts for absolute differences in health outcomes across groups. Calculating the Kolm Index post-intervention involves applying the α parameter to the differences in QALE post-intervention and the average QALE [22]. This index is particularly responsive to changes at the lower end of the health distribution, thus highlighting the intervention’s impact on the most disadvantaged groups.

Step 7: Computing the equally distributed equivalent (EDE) post-intervention: Finally, the EDE QALE post-intervention is calculated, which represents the average QALE that would result if the health gains observed were distributed equally across all socioeconomic groups. This step may utilize mathematical programming or simulation techniques to redistribute health gains and compute the EDE. It is a crucial step as it synthesizes the equity considerations into a single measure, allowing for a comparison of the equity impact of different health interventions.

Step 8: Uncertainty analysis: DCEA extends traditional CEA by incorporating differential effects of health interventions across socioeconomic strata. This methodology enables the quantification of an intervention’s influence on overall population health while simultaneously capturing its impact on health equity among distinct population groups. Central to DCEA is the assessment of uncertainty, which pertains to both the potential health benefits of an intervention and its effects on health inequality [33]. Addressing decision uncertainty is critical, as it informs the confidence in policy recommendations derived from DCEA outcomes. Two principal analytical methods are employed in this assessment: value of information (VOI) analysis and analysis of covariance (ANCOVA). VOI analysis quantifies the expected benefit of acquiring additional information, thus elucidating the value derived from mitigating uncertainty. This method aids in pinpointing where further research could yield significant impacts on policy direction, based on the existing evidence base. It provides a systematic approach to evaluating whether the potential reduction in uncertainty justifies the investment in additional data collection or research, particularly when an intervention shows promise but its distributional effects remain uncertain. ANCOVA complements VOI analysis by statistically parsing the factors contributing to outcome variability in a DCEA. Through this analysis, researchers can investigate how different covariates—such as baseline health conditions, adherence to treatment protocols, and differential effectiveness across demographic profiles—affect the cost effectiveness and equity dimensions of an intervention. This methodological approach is instrumental in identifying the sources of uncertainty and guiding subsequent research to areas that promise the greatest impact on policy decisions. The incorporation of uncertainty analysis into DCEA enhances the ability to inform HTA decisions that are both data-driven and ethically informed, ensuring that efforts to improve health outcomes are aligned with the goal of reducing disparities across socioeconomic groups.

Each parameter and step in this DCEA approach is critical to providing insight into both the effectiveness and equity of health interventions. The QALE offers a measure of the health benefits of interventions, while α and the Kolm Index bring the equity dimension into focus. Together, they facilitate a comprehensive assessment that informs which interventions can deliver the greatest health benefits across all societal groups equitably. As such, DCEA extends the focus of HTA to include patient-centered outcomes and equity considerations, addressing potential disparities in healthcare access and outcomes. However, it stops short of capturing broader societal impacts of health technologies. Thus, we recommend augmenting DCEA with macroeconomic analysis to explore the wider economic effects of healthcare interventions, providing a more complete assessment of their societal value.

3.2 Extended HTA Analysis with Input-Output Modeling

3.2.1 Overview of the Input-Output Model

The input-output model can be effectively utilized to assess the wide-ranging economic effects of healthcare interventions. The model, conceived by economist Wassily Leontief [25], is a practical tool for mapping out how different sectors of the economy interact with one another, showing us how the activities of the healthcare sector can have ripple effects throughout the economy.

Extended HTA analysis, augmented with input-output modeling, broadens the scope of traditional HTA by integrating techniques that quantify the wider economic impacts of healthcare interventions. This enhanced approach provides a thorough evaluation of secondary effects, including shifts in productivity, innovation, and social welfare. Originating from the work of economist Wassily Leontief [25], the input-output model serves as a valuable instrument for charting the intersectoral dynamics of the economy, highlighting the far-reaching impact that healthcare sector activities have beyond their direct environment.

At its core, the input-output model employs a matrix (input-output table) to manifest the transactions between sectors within the economy, painting a detailed picture of how expenditures and consumptions traverse through various industries. The input-output table is instrumental in visualizing the intersectoral flow of economic value, quantifying how healthcare investments can propagate through and impact the economy at large.

By embracing the linear relationships intrinsic to this model, we can project how changes within the healthcare sector influence the production functions of interconnected sectors. Although the model traditionally operates on the premise of fixed coefficients and an inexhaustible supply capability, assumptions that offer a degree of simplification, the analysis acknowledges and adjusts for these parameters to ensure nuanced and realistic estimations.

Furthermore, while the input-output model often considers a closed economy, the approach can expand this horizon to encapsulate the dynamism of global trade, thereby refining the understanding of healthcare interventions’ international repercussions, as needed.

This approach not only underscores the direct impact of healthcare interventions but also illuminates their extended influence on sectors seemingly remote from healthcare. It equips us with a quantitative foundation to argue for a more comprehensive appreciation of the healthcare sector’s contributions, advocating for policies that recognize and reward the full spectrum of impacts that healthcare interventions yield across the entire economy. The following subsections outline the steps necessary for the successful implementation of this approach.

Step 1: Choice of aggregation level: In an input-output model, aggregation analysis refers to the grouping of economic sectors and activities for analysis. This allows for a more manageable analysis, as many sectors can be consolidated into broader categories, rather than analyzing each sector individually. The aggregation level of the analysis depends on the research question being addressed, the availability of input-output tables and the data on the final demand, which is the spending component of the gross domestic product (GDP). Typically, input-output tables are created with data related to specific economic areas, which may be the national economy and also an economic region. One can construct a symmetric industry-by-industry input-output table (SIOT) from the ‘Make and Use’ tables in the ‘After Redefinition Tables’ of the Bureau of Economic Analysis (BEA) [26]. The latter reveals how products flow in the economy, from being sold (rows) to being purchased (columns). Each industry is identified using the North American Industry Classification System (NAICS) [27]. Such industries, which are directly relevant to the analysis, are identified based on their contribution to the economic activity under consideration. This may include industries involved in producing goods or services that impact the healthcare sector and disease state for which the intervention is devised.

Step 2: Analyzing the characteristics of a SIOT: The SIOT is divided into four sections, each serving a unique purpose (Online Resource 1 eTable 1). The top left section shows the exchange of materials and services between industries, known as intermediate inputs. It helps us understand how industries rely on one another. The bottom left section focuses on the value-added components within each industry. It reveals the portion of the final price generated within each industry, such as wages, profits, and taxes. The top right section represents the final demand. It captures the total amount of products and services consumed by households, businesses, and the government. The lower right section of the table is left empty as it is not used in this analysis for relevance purposes. In fact, the model is simplified to focus on the key sectors and flows of interest, leaving out parts of the matrix that do not contribute to the analysis at hand (Online Resurce 1 eTable 1).

Step 3: Calculating the Leontief inverse: Upon establishing the SIOT, we can calculate technical coefficients. These coefficients reflect the proportions of inputs utilized by various industries to generate their outputs, serving as a fundamental aspect of input-output analysis [28]. Unlike variable coefficients in other models, these coefficients are assumed to be constant and are organized into what is known as the direct coefficient matrix. This matrix forms the basis for analyzing the input-output relationships within the economy, following a specific production function termed the Leontief production function [29].

The Leontief production function is distinctive because it assumes fixed input ratios, simplifying the analysis of economic interdependencies. This function is operationalized by subtracting the technical coefficient matrix from an identity matrix, often referred to as the unity matrix. The identity matrix is characterized by ones on its main diagonal and zeros elsewhere, establishing a standard reference for matrix operations.

To assess the impact of a change or shock in a specific industry, economists employ multipliers. These multipliers are derived from the inverse of the matrix formed by the Leontief production function, known as the Leontief inverse. They are crucial in measuring the ripple effects of a change in one industry on the broader economy. By using these multipliers in specific equations, we can calculate the overall economic impact of various shocks or changes.

For instance, consider a scenario where there is an increase in demand in the automotive industry. The multipliers can help quantify how this demand increase affects not only the automotive supply chain but also other unrelated sectors and the economy at large. This ability to model economic interconnections is particularly valuable in scenarios and policy simulations [30].

Step 4: Applying the input-output model using simulation scenarios: Through input-output analysis, policymakers and analysts can simulate the effects of various interventions—such as drugs, adjustments in government policies, investment levels, or shifts in demand/supply patterns—to understand the net broader impacts on the economy. As a first step, the model simulates the economy’s baseline estimate using multiplier coefficients that connect economic variables. This baseline scenario represents the current economic situation without considering the intervention’s economic consequences. In the second step, the model creates an alternative scenario by introducing exogenous changes in the economic variables for the US economy over a specific period to explicitly reflect the impact of the intervention. By comparing the outputs of the control scenario with those of the simulation accounting for the exogenous changes, we can quantify the net economic impact. There could be more than one alternative scenario that reflects several interventions or policies that are envisioned to be assessed or analyzed.

Step 5: Uncertainty analysis: Given the assumptions and potential for incomplete data, uncertainty analysis is crucial for interpreting the input-output model results. It begins with identifying sources of uncertainty, such as the estimation of technical coefficients. Sensitivity analysis follows, pinpointing impactful variables. This consists of quantifying the identified sources of uncertainty using appropriate statistical methods. Moreover, VOI analysis can be utilized to gauge the potential advantages linked with diminishing parametric uncertainty [31].

The literature suggests several approaches to quantifying uncertainty in input-output modeling [32]. First, the deterministic error analysis sets boundaries for variables, exemplified by establishing limits on the costs of raw materials and the output of pharmaceutical products. This is complicated by the Leontief inversion, a non-linear mathematical process translating inputs into outputs. The econometric estimation of input-output coefficients then uses statistical methods to understand relationships within the economy. For instance, it might analyze how demand in the chemical industry affects pharmaceutical production, based on data from different business sectors. As for the error transmission method, it focuses on neutralizing random errors to improve result accuracy. This is illustrated by a pharmaceutical company predicting a 10% reduction in hospital admissions with a new drug. If the actual reduction is lower, the discrepancy, potentially caused by random factors like patient adherence, is analyzed to refine the drug’s economic impact assessment. Finally, the full probability density function approach aggregates data from various sources to create a more accurate overall picture, while the extended Monte Carlo simulations and Bayesian/Entropy methods are used for advanced equilibrium analysis and data treatment, enhancing model reliability. Additionally, scenario analysis can be conducted to explore different outcomes under varying conditions. The input-output model is iteratively refined with new data, with methods that are transparently documented, ensuring reproducibility.

3.3 Key Assumptions to Consider When Using DCEA and Input-Output

When employing DCEA, it is essential to underpin the analysis with several critical assumptions. These assumptions frame the interpretability and generalizability of the results. It is assumed that the QALYs gained from an intervention are comparable across different socioeconomic groups, which may not always reflect the complexity of real-world effectiveness [34]. This assumption of uniform QALYs valuation requires careful consideration, as the perception of health gains could be influenced by factors unique to each group [22]. Another key assumption is that the intervention’s cost effectiveness derived from trial or modeled data can be extrapolated to different population subgroups. This extrapolation assumes that the intervention’s relative effectiveness and resource utilization are constant, which may not hold true across diverse socioeconomic settings. Additionally, DCEA presupposes that the societal perspective for the value of reducing inequality, as expressed by the inequality aversion parameter (α), is accurately captured and constant. This parameter’s selection is subject to ethical considerations and may significantly influence the outcome of the analysis [35]. It is also crucial to acknowledge that the methods used to quantify health inequalities, such as the Kolm Index or the Concentration Index, are based on assumptions about the social welfare function and the distribution of health states within a population [36]. Lastly, when interpreting the results of a DCEA, it is typically assumed that policymakers will act rationally upon the evidence presented, incorporating both cost effectiveness and distributional impacts into their decision-making processes [37].

In employing an IO model for economic analysis, we must carefully consider the underlying assumptions that govern its framework. The model is predicated on the notion of linear relationships between inputs and outputs across different sectors. It assumes a proportional scale of production, where inputs and outputs increase or decrease in unison. While this linear perspective facilitates a straightforward computational approach, it may not encapsulate the complexities of real-world production where returns to scale can vary dramatically.

Furthermore, the input-output model is built on the assumption of fixed coefficients. These coefficients, which dictate the input needs per unit of output, are considered static. However, this assumption does not hold up well under the dynamic conditions of a real economy where technological advancements and shifts in relative prices can lead to substitutions among inputs, thus altering these coefficients over time.

Another critical assumption is the absence of supply constraints within the model. It posits that sectors have the unfettered capability to meet any level of demand. This ignores the finite nature of resources and the production ceilings imposed by factors such as limited natural resources or production capacity, which are very real challenges that industries face.

Lastly, the model, in its more rudimentary form, operates on the assumption of a closed economy. This means it does not account for the intricacies of international trade, which are integral to most modern economies. Although more sophisticated iterations of the model do consider trade, this simplification can lead to significant inaccuracies in economic analysis when using the basic version of the model.

These foundational assumptions are essential to the input-output model’s operation, yet they also delineate its limitations. Analysts must recognize and adjust for these limitations when applying the model to real-world economic situations, to ensure that the insights derived are both relevant and robust.

3.4 Decision Rules Through Voting Scheme

A voting scheme is established to guide the adoption of interventions. It delineates four distinct quadrants, each representing different combinations of net health benefits and net broader impacts (Fig. 2). Quadrant I embodies the epitome of an ideal choice, where interventions not only yield positive augmented net health benefits but also generate extensive net broader impacts. These interventions, characterized by their ability to provide the highest value for money, are actively sought after by organizations and individuals striving to maximize the impact of their allocated resources. Transitioning to quadrant II, we encounter interventions that manifest positive augmented net health benefits but are accompanied by a negative net broader impact. The pursuit of options in this quadrant is warranted when the augmented net health benefits outweigh the negative broader impacts, thereby justifying the investment in the intervention. While acknowledging the existence of some unfavorable consequences, the overall positive health outcomes make this quadrant a viable and prudent choice. Conversely, quadrant III encompasses interventions that exhibit both a negative augmented net health benefit and a negative net broader impact. As such, these interventions are deemed inadequate choices, as they engender detrimental outcomes. To ensure optimal decision making, options within this quadrant should be unequivocally rejected due to their deleterious effects. Lastly, quadrant IV comprises interventions with a negative augmented net health benefit but a positive net broader impact. In certain circumstances, organizations or individuals may opt to pursue options within this quadrant when the positive societal impacts outweigh the negative augmented net health benefits. Consequently, the investment in such interventions can be deemed justified, despite their inability to directly enhance health outcomes.

Fig. 2
figure 2

Health technology assessment decision quadrants

Full size image

The voting scheme is designed to integrate multiple stakeholders’ perspectives, including patients, healthcare providers, policymakers, researchers, advocacy groups, payors, and community representatives, ensuring a balanced and inclusive approach to decision making. This integration occurs through a structured voting process where each stakeholder group is given a platform to express their priorities and concerns regarding the interventions under consideration.

Patients provide insights based on their personal experiences and preferences, highlighting the direct impact of interventions on their health and well-being. Healthcare providers, such as doctors and nurses, offer a clinical perspective, evaluating the interventions’ efficacy, safety, and feasibility within the healthcare system. Policymakers consider the broader societal impacts, including economic, ethical, and social implications, ensuring that interventions align with public health goals and resource allocation principles. Stakeholders cast their votes for or against the adoption of interventions, with the results aggregated to identify the options with the broadest impact.

To reconcile differences and reach a consensus, the scheme may employ weighting systems, negotiation rounds, or expert panels to further deliberate on contentious issues. This multifaceted approach facilitates a democratic and transparent decision making process, where the diverse values and priorities of the community are reflected in the final choices [38].

A detailed step-by-step explanation of the proposed voting scheme, integrating elements of decision conferencing [38] for a compelling and inclusive approach to decision making, is as follows:

Step 1: Stakeholder identification and engagement: The first step involves identifying key stakeholders, including patients, healthcare providers, policymakers, researchers, industry representatives, advocacy groups, payors, and community leaders. Each stakeholder group brings a unique perspective and expertise to the decision-making process.

Step 2: Preparation and information sharing: Prior to voting, stakeholders are provided with comprehensive information about the healthcare interventions under consideration. This includes evidence-based data on efficacy, safety, cost effectiveness, and potential societal impacts. Decision conferencing techniques are employed to facilitate constructive dialogue and knowledge exchange among stakeholders.

Step 3: Voting process: Stakeholders are presented with a set of options for each healthcare intervention, ranging from adoption to rejection. They cast their votes based on their assessment of the options, taking into account their respective interests and expertise. Votes are submitted anonymously to encourage candid and unbiased decision making.

Step 4: Aggregation and analysis of results: Once all votes are collected, they are aggregated and analyzed to determine the level of support for each option. This involves tallying the votes and identifying patterns or trends across stakeholder groups. Decision conferencing tools such as visualizations and simulations may be used to facilitate data interpretation.

Step 5: Reconciliation and consensus building: In cases where differences in opinion arise, the voting scheme employs reconciliation mechanisms to facilitate consensus building. This may include weighting systems to account for the relative importance of stakeholder perspectives, negotiation rounds to address conflicting viewpoints, or expert panels to provide additional insights and recommendations.

Step 6: Final decision and implementation: After thorough deliberation and consensus building, a final decision is reached regarding the adoption of healthcare interventions. The decision reflects the collective wisdom and input of all stakeholders involved. Implementation plans are developed to ensure effective execution of the chosen interventions, with ongoing monitoring and evaluation to assess their impact over time

4 Case Study: Assessing the Value of Reset-O for Opioid Use Disorder Treatment

4.1 Case Presentation

The widespread occurrence of opioid use disorder (OUD) has placed significant burdens on both healthcare systems and society at large. This has triggered the need to develop alternative treatment approaches for patients suffering from OUD. reSET-O, a US FDA-approved prescription digital therapeutic, is now being considered for formulary placement by an employer-sponsored health plan in the United States (US). The employer-sponsored health plan is particularly interested in a comprehensive assessment of the broader impact and benefits associated with the utilization of reSET-O in combination with treatment as usual (TAU) when compared with TAU alone for its qualified members.

4.2 Methods

This case study employs a multipronged framework that encompasses DCEA implications and macroeconomic impacts to optimize resource allocation for OUD treatment, as previously described.

4.2.1 DCEA

Step 1: Defining the population of interest: For this case study, the population comprised patients diagnosed with OUD who were eligible for treatment with the digital therapeutic application reSET-O. To capture the nuanced effects of socioeconomic status on the outcomes of the intervention, we stratified these patients into five quintiles based on their socioeconomic status. This is defined based on household income obtained from the US census data [39]. These quintiles ranged from the lowest to the highest, including second, middle, and fourth, thereby encompassing the full spectrum of socioeconomic status. This stratification allowed us to evaluate the distributional impact of the reSET-O treatment across different socioeconomic layers and to understand how the intervention could potentially ameliorate or exacerbate existing health inequalities among patients with OUD.

Step 2: Determining baseline and average QALE by socioeconomic status: The baseline data in Online Resource 2 eTable 1, illustrates the disparities in QALE across five socioeconomic groups before any intervention. The average QALE is set at 60 years, with the lowest quintile (Group 1) having a QALE of 40 years, indicating a significant disadvantage of 20 years compared with the average. Conversely, the top quintile (Group 5) enjoys a QALE 20 years above the average.

Step 3: Applying the inequality aversion parameter (α): In our analysis, we utilized estimates of 0.25 for the α parameter, which would typically be obtained through a survey of the general public in the US. Social welfare was derived by applying the Kolm index to the mean level of health in the distribution, resulting in the determination of the EDE level of health.

Step 4: Introducing the interventions and recalculating QALEs: With the introduction of health interventions, such as reSET-O + TAU, QALEs are recalculated for each socioeconomic group. These new QALE figures reflect the health benefits or losses each intervention provides (Online Resource 2 eTables 3 and 4).

Step 5: Reassessing inequality using the Kolm Index post-intervention: Post-intervention, the Kolm Index is employed to reassess inequality levels. This index is sensitive to changes in the distribution of health gains and losses. A decrease in the Kolm Index, as observed following the reSET-O + TAU intervention, signifies a reduction in health inequality (Online Resource 2 eTables 3 and 4).

Step 6: Computing the EDE post-intervention: Finally, the EDE health measure is computed to represent what the average QALE would be if health gains post-intervention were distributed equally among all groups. An increased EDE post-intervention value indicates a movement toward a more equitable health outcome distribution. An Excel (Office 365/Microsoft 365; Microsoft Corporation, Redmond, WA, USA) model that details the calculation steps described above is made available to readers as a companion file (Online Resource 2).

4.2.2 Extended HTA of reSET-O

We used an input-output model to understand how the reSET-O treatment for OUD affects the wider economy. This approach helps us see the full range of economic changes that healthcare treatments can bring about. In this analysis, we focus on value added, income and employment impacts. We break down this analysis into several clear steps:

Step 1: Choice of aggregation level: Our analysis is built on a 2021 SIOT, which dissects the economy into distinct sectors. This delineation includes critical sectors such as OUD subsectors, amended health sectors, and non-health-related sectors. The specificity of this categorization is achieved by using NAICS codes, enabling precise tracking of economic activities.

Step 2: Analyzing the characteristics of the SIOT: The SIOT, represented in Online Resource 3 eTable 1, is a matrix showcasing the flow of economic transactions between sectors. It captures both the intermediate demand (the goods and services exchanged for production) and the final demand (the consumption by end users such as households and government). This table culminates in a total output for each sector, summing up the intermediate and final demands.

Step 3: Calculating the Leontief inverse: At the core of our analysis is the Leontief Inverse, computed by first determining the technical coefficients in Online Resource 3 eTable 2. These coefficients, reflecting the intersector production dependencies, are obtained by dividing the intermediate demand by each sector’s total output. We then constructed an identity matrix, as seen in Online Resource 3 eTable 3, which, when combined with the technical coefficients, results in the Leontief Matrix (Online Resource 3 eTable 4). Subtracting the technical coefficients from the identity matrix provides us with the Leontief Inverse, a pivotal element that illustrates how sectoral changes propagate through the economy. Online Resource 3 eTable 5 showcases the direct, indirect, and total multiplier effects of economic activities. It illustrates how spending in one sector can amplify throughout the economy, significantly in non-health-related sectors with a total multiplier effect of 2.9304. Online Resource 3 eTable 6 offers employment figures, which, when juxtaposed with changes in final demand, estimate the potential employment impacts across sectors.

Step 4: Applying the input-output model using simulation scenarios: Employing the input-output model, we examine two scenarios. Initially, we establish a baseline scenario (TAU alone), calculated by multiplying the final demand by the output ratios from Online Resource 3 eTable 7. We then simulate an alternative scenario (reSET-O + TAU) in which we hypothesize a 15% reduction in final demand within the OUD sector, diverting these resources to non-health sectors. The implications of such a shift are quantified by examining the changes in value added, income, and employment, as captured in Table 3. This comparative analysis allows us to unveil the net broader impacts of the reSET-O intervention, demonstrating its potential for enhancing economic value, augmenting income, and creating employment opportunities compared with the status quo. An Excel (Office 365/Microsoft 365) model that presents the calculation steps described above is made available to readers as a companion file (Online Resource 3).

4.2.3 Decision Rules through the Voting Scheme

The voting scheme consisted of the integration of the results of DCEA and macroeconomic impacts to provide decision makers with a comprehensive framework for informed decision making, as described in Sect. 3.4. Decision makers evaluating the body of evidence generated through this analysis would include patients, clinicians, researchers, and policymakers.

4.3 Results

4.3.1 DCEA

Upon applying the reSET-O combined with TAU, we observe an improvement in the QALE of disadvantaged groups (Groups 1 and 2) [Online Resource 1 eTable 3], where the presence of disease is indicated. The QALE increases to an average of 66 years, and the exponential function reflects decreased inequality, as shown by the Kolm Index reduction to 9.9, from 13.9. The EDE post-intervention QALE also increases, demonstrating that this intervention improves both health outcomes and equity, with a significant impact of DCEA (19.9) over the unadjusted health value (30) [Online Resource 1 eTable 3].

Conversely, when TAU is applied without the reSET-O intervention, we see an increase in the QALE for the advantaged groups (Groups 4 and 5) who already had a higher baseline QALE (Online Resource 1 eTable 4). The intervention exacerbates health inequalities, as evidenced by the Kolm Index increase to 19.9, indicating that society’s aversion to inequality is not addressed. The EDE post-intervention QALE decreases for these groups, suggesting a regressive effect, where the wealthier benefit more, thus widening the health disparity gap. This is further supported by the negative impact of DCEA (−29.9), showing a loss in health value when distributional effects are considered (Online Resource 2 eTable 4). Furthermore, the analysis demonstrated that reSET-O + TAU had a more pronounced impact on QALE gains (10 EDE QALYs gained) compared with TAU alone (0.02 EDE QALYs gained) (Table 1). By quantifying the distribution of health gains and losses, the analysis highlights the potential equity-enhancing aspects of reSET-O + TAU in OUD treatment. These findings underscore the importance of considering equity considerations when evaluating the value of healthcare interventions.

Table 1 Impact of DCEA on QALE
Full size table

The DCEA health value metric provides a nuanced view of the interventions’ effects, revealing that reSET-O + TAU is a more equitable approach that benefits the socioeconomically disadvantaged groups, whereas TAU alone favors the already advantaged, worsening health inequality. These findings underscore the importance of considering health equity in HTA decisions, as interventions can have divergent effects on different segments of the population.

4.3.2 Extended HTA of reSET-O

Table 2 represents the inverse of Leontief matrix used in the estimation of the broader impact of the reSET-O + TAU intervention. The results in Table 3 suggest an increase in value added to the economy that is substantial, amounting to $6.7 million due to the reSET-O + TAU intervention. This figure reflects the total economic value generated by the intervention, distributed over several sectors. Notably, one sector alone accounts for a significant portion of this increase, contributing $3.19 million. This uptick in value added is a clear indication of the intervention’s capacity to enhance the overall productivity and worth of the economy. As for income, the intervention has collectively raised earnings by $2.3 million across all sectors. This boost is highlighted by the largest single increase within a sector, which stands at $1.37 million. Such an increase in income underscores the potential of the reSET-O + TAU intervention to raise the financial well-being of individuals within the economy.

Table 2 Inverse of Leontief Matrix
Full size table
Table 3 Calculation of broader net impacts for reSET-O + TAU
Full size table

Employment also sees a beneficial impact from the intervention, with a creation of 11.7 jobs spread across the sectors. The most pronounced effect within a single sector has been the addition of 5.28 jobs. This growth in employment is indicative of the intervention’s role in job creation and its positive influence on the labor market. Through this nuanced and detailed approach, our input-output model transcends traditional economic assessments, enabling us to highlight the broader economic ramifications of reSET-O, thus providing valuable insights into its potential benefits for the broader economy.

4.3.3 Decision Rule Through Voting Scheme

The extended HTA analysis demonstrates positive macroeconomic impacts, including a positive net value added, an increase in income, and a labor force gain of 12 individuals (Table 3). These broader impacts contribute to the overall value of reSET-O. This placed reSET-O in the first quadrant of the voting scheme proposed. As a result of the decision conferencing, it was recommended that the employer-sponsored health plan should consider the implementation of reSET-O as part of a comprehensive treatment strategy for patients with OUD in their network. The intervention demonstrates cost effectiveness, enhances patient-centeredness, addresses equity concerns, and has positive macroeconomic impacts. These findings support the value and potential benefits of reSET-O in improving patient outcomes and optimizing resource allocation for OUD treatment.

5 Discussion

This manuscript presented a novel conceptual framework for the value assessment of healthcare interventions, aiming to address the limitations of current HTAs in capturing the broader economic and societal impacts beyond the healthcare sector. The framework combines DCEA and extended HTA analysis with input-output modeling, with a voting scheme to guide decision making.

The integration of DCEA into the framework allows for a more comprehensive evaluation of healthcare interventions by incorporating equity considerations and patient-centered outcomes. By assessing the impact of interventions on different population subgroups and evaluating disparities, DCEA enhances the understanding of the distributional impacts of interventions. This information is crucial for decision makers to ensure that the adoption of interventions does not exacerbate existing healthcare disparities, and is aligned with the goal of equitable access to effective healthcare. However, implementing DCEA requires access to reliable data on health outcomes, costs, and sociodemographic and equity variables, which may pose challenges in certain settings.

Furthermore, the inclusion of extended HTA analysis with input-output modeling is a significant advancement in capturing the broader economic impacts of healthcare interventions. By simulating the ripple effects on various economic sectors, such as changes in productivity, job creation, and innovation, this analysis provides a more holistic assessment of the societal consequences. This approach acknowledges that healthcare interventions have implications beyond the healthcare sector, impacting the overall economy and society. The input-output modeling technique offers valuable insights into the interconnectedness of industries and the potential economic impacts associated with healthcare interventions. However, it is important to note that implementing this analysis requires access to comprehensive and up-to-date economic data, which may pose challenges, particularly in resource-constrained settings.

To guide decision making, the proposed voting scheme combines the outputs of DCEA and extended HTA analysis, providing a practical approach that considers multiple dimensions and stakeholder perspectives. This inclusive approach helps balance the assessment of interventions by considering not only their health outcomes and economic impacts but also their broader societal implications. The voting scheme provides decision makers with a framework to guide the adoption of interventions based on a more comprehensive evaluation of their value. However, it is essential to ensure the representation of diverse stakeholders and their meaningful involvement in the decision-making process to avoid potential biases.

While the proposed framework shows promise, there are several challenges and considerations that need to be addressed. The availability and quality of data play a crucial role in the implementation of the framework. Reliable data on health outcomes, costs, equity-relevant sociodemographic variables, and intersectoral linkages are necessary to conduct robust assessments. Efforts should be made to improve data collection and sharing mechanisms to support the implementation of the framework in different settings. It is reassuring to see that the WHO has recently been leading the way by facilitating data collection through the released Health Inequality Data Repository (HIDR), the largest global database of publicly available disaggregated data on health inequality [40].

The integration of broader impacts beyond the health sector, stemming from the use of health technologies, necessitates careful consideration of ethical implications. Recognizing the fundamental role of ethics in healthcare, our approach adheres to key bioethical principles: justice, ensuring fair access and distribution of healthcare resources; autonomy, respecting patient choices and rights; beneficence, maximizing benefits while minimizing harm; and non-maleficence, avoiding harm to patients. The ethical challenges in our framework primarily revolve around potential conflicts between health and non-health sector impacts. For instance, while an intervention may demonstrate substantial non-health sector advantages, it must not compromise the quality of patient care or overlook the individual’s healthcare needs. This balancing act requires a nuanced understanding of both impacts and ethical priorities. To navigate these complexities, we propose developing clear guidelines that assist decision makers in evaluating healthcare interventions when broader impacts are considered. These guidelines should prioritize patient well-being, ensuring that economic considerations do not overshadow the fundamental goal of healthcare: to provide beneficial and harm-free treatment to patients. Moreover, the inclusion of stakeholder perspectives, including patients, healthcare providers, and policymakers, is crucial in ensuring that the framework remains aligned with ethical values and societal expectations. Moreover, our framework aims to address healthcare disparities without favoring one group over another, fostering equity. Equity considerations are essential in decision making but are not the sole factor. Thresholds for prioritizing equity are context-dependent and ethically guided. The complexity of implementing a voting scheme to reflect societal values is acknowledged, but its significance in democratizing decision making is emphasized. This approach begins a conversation for a more inclusive healthcare evaluation, highlighting areas for further research and refinement in our framework. In summary, our framework seeks not only to advance healthcare evaluation by incorporating economic aspects but also to uphold the highest ethical standards, ensuring that economic benefits enhance rather than detract from the primary objective of healthcare—improving patient outcomes.

Our framework’s success hinges on interdisciplinary collaboration, uniting health economics, social sciences, and macroeconomics. This collaboration, involving diverse professionals, is key to effectively implementing and interpreting the framework’s results. Establishing collaborative platforms, such as interdisciplinary committees or working groups, allows professionals to regularly meet, discuss, and align on methodologies. These platforms would serve as hubs for sharing knowledge, debating ideas, and formulating collective strategies. Additionally, we recommend the development of structured interdisciplinary training programs and workshops that focus on sharing knowledge and techniques across different fields. These training programs would aim to build a common language and understanding among various professionals involved in HTA, facilitating better communication and collaboration. Finally, our approach includes regular review and updates of collaborative strategies based on feedback and outcomes, ensuring continuous improvement and alignment with the latest research and practices in HTA. This dynamic approach ensures that our framework remains relevant and effective in the ever-evolving field of healthcare. It brings together concepts and methods from health economics, social sciences, and macroeconomics. Collaboration between researchers, policymakers, economists, and healthcare professionals is essential to ensure the successful implementation and interpretation of the framework’s results. Building capacity and fostering partnerships across disciplines will be vital in advancing the field of comprehensive value assessment. To ensure its full adoption, the framework needs to be further validated and tested in various healthcare contexts and settings. Case studies, pilot projects, and sensitivity analyses can help assess the feasibility, reliability, and generalizability of the framework. By applying the framework to different healthcare interventions and comparing the results with traditional HTAs, its added value and potential for informing decision making can be better understood.

In Sect. 2 of our manuscript, we explore the limitations of contemporary HTA and acknowledge that certain impacts, such as environmental pollution and resource depletion, are not directly addressed in our proposed conceptual framework. We want to emphasize however that the importance of these issues is not diminished in our analysis. Our framework is designed to be both innovative and encompassing. Incorporating environmental impacts into broader evaluations could significantly enhance the assessment landscape. By employing the input-output model as a key methodological tool, our framework could integrate these impacts more fully. This would involve a detailed calibration of technical coefficients within our framework, specifically adjusted to account for the environmental effects on the economic sector’s interconnections. Such an approach not only highlights the critical role of environmental considerations but also enhances the precision and relevance of our evaluations. It ensures a thorough comprehension of the complex interplay between ecological factors and economic dynamics, affirming our commitment to a holistic and forward-thinking assessment methodology.

6 Conclusion

This paper’s proposed conceptual framework presents a significant advancement in improving HTAs to account for the broader economic and societal impacts of healthcare interventions. By integrating DCEA, extended HTA analysis with input-output modeling, and a voting scheme, decision makers can make more informed choices that align with societal priorities and promote equitable access to effective healthcare interventions. However, further research, collaboration, and validation are needed to fully realize the potential of this framework and ensure its practical applicability in diverse healthcare contexts.