ObjectiveTo examine trends in the inclusion of societal costs in published cost-effectiveness analyses (CEAs), factors associated with their inclusion, and the impact of societal costs on incremental costs and incremental cost-effectiveness ratios (ICERs).MethodsWe analyzed 7,800 CEAs from 2013 to 2023 using the Tufts Medical Center CEA registry. The inclusion of societal costs in CEAs was evaluated across study characteristics. Associations between study characteristics and the inclusion of societal costs were analyzed using multivariate logistic regression. For studies reporting health care and societal perspectives, we assessed the impact of including societal costs on incremental costs and ICERs.ResultsFrom 2013 to 2023, CEAs including societal costs increased from 19% to 28%. Productivity was the most frequently reported component (12%), followed by transportation (8%), caregiver time (6%), patient time (5%), and consumption costs (1%). Compared with US-based analyses, studies from Scandinavian countries (adjusted odds ratio [OR]: 3.6) and the Netherlands (5.6) had higher odds of including societal costs, whereas studies from Canada (0.7), Australia (0.6), and the United Kingdom (0.4) had lower odds. Studies on mental health disorders (6.2) and immunization (4.1) had the highest odds of including societal costs. Compared with CEAs focused on adults, CEAs targeting pediatric populations had higher odds (OR: 1.6), while those targeting the elderly had lower odds (OR: 0.7). Upon inclusion of societal costs, incremental costs decreased in 72% and increased in 28% of studies; the ICER decreased in 74% and increased in 26% of studies.ConclusionDespite the increase in recent years, societal costs are infrequently included in CEAs, with substantial variation by country, disease, and population. Including societal costs can meaningfully improve value assessments and should be guided by relevance, evidence, and decision context.HighlightsBuilding on prior work by Kim et al. (2020), which analyzed approximately 6,900 cost-effectiveness analyses (CEAs), this study examined a larger and more recent sample of 7,800 CEAs from 2013 to 2023. In addition to updating the evidence base, we conducted new analyses to assess trends, associated factors, and the effect of including societal costs on incremental cost-effectiveness ratios (ICERs), thus providing insights that were not explored in prior work and addressing a key evidence gap in health economics.The inclusion of societal costs in CEAs rose modestly from 19% to 28% from 2013 to 2023, with substantial variation across countries, diseases, and intervention types. In some cases, the inclusion of societal costs affected incremental costs and ICERs enough to cross commonly used cost-effectiveness thresholds.The inclusion of societal costs can help improve value assessments in health care interventions, but it should be guided by relevance, available evidence, and the potential to influence decision making. Identifying when and where societal costs meaningfully affect outcomes can support more consistent and appropriate use.

BackgroundMedication adherence is a critical factor in hypertension management, which remains a challenge for public health systems.MethodsGraded-pair questions were used to quantify the perception of how much nonadherence to antihypertensives increases the risk of serious cardiovascular events. A discrete-choice experiment was used to quantify the relative importance of medication outcomes (e.g., reduction in cardiovascular event risk and medication side effects). Rating questions were used to assess perspectives of the effect of treatment nonadherence on treatment side effects. Results were combined to assess how preferences and outcome expectations influence adherence.ResultsPatients perceived treatment adherence as the most significant contributor to cardiovascular event risk. A reduction in cardiovascular risk was the most significant consideration when choosing medication. Missing consecutive (v. alternate) doses was associated with greater perceived cardiovascular risk and fewer side effects. The differences between complete adherence and any level of nonadherence were significantly larger for side effects than for changes in the risk of cardiovascular events, suggesting that side effects are perceived to be more sensitive to nonadherence than treatment efficacy.LimitationsOur study relied on hypothetical scenarios, which may not fully capture real-world decision making. While our findings shed light on the relationship between adherence patterns and treatment perceptions, it is essential to recognize the complexity of adherence behavior.ConclusionsPatients believe that they can manage medication side effects by skipping doses without compromising the efficacy to the same degree and that they can offset compromises in efficacy by avoiding missing consecutive doses for prolonged periods.ImplicationsHealth care providers should understand the importance of patient education and counseling to address misconceptions and promote realistic expectations regarding treatment efficacy and the consequences of nonadherence.HighlightsThe average patient believes that they can manage medication side effects by skipping doses without compromising the efficacy to the same degree.There is a belief that patients can offset some of the impact of nonadherence on their cardiovascular event risk, particularly if they avoid missing consecutive doses for prolonged periods of time.This highlights the importance of patient education and counseling to address misconceptions and promote realistic expectations regarding treatment efficacy and the consequences of nonadherence.

Network meta-analysis (NMA) synthesizes data from randomized controlled trials to estimate the relative treatment effects among multiple interventions. When treatments can be grouped into classes, class effect NMA models can be used to inform recommendations at the class level and can also address challenges with sparse data and disconnected networks. Despite the potential of NMA class effects models and numerous applications in various disease areas, the literature lacks a comprehensive guide outlining the range of class effect models, their assumptions, practical considerations for estimation, model selection, checking assumptions, and presentation of results. In addition, there is no implementation available in standard software for NMA. This article aims to provide a modeling framework for class effect NMA models, propose a systematic approach to model selection, and provide practical guidance on implementing class effect NMA models using the multinma R package. We describe hierarchical NMA models that include random and fixed treatment-level effects and exchangeable and common class-level effects. We detail methods for testing assumptions of heterogeneity, consistency, and class effects, alongside assessing model fit to identify the most suitable models. A model selection strategy is proposed to guide users through these processes and assess the assumptions made by the different models. We illustrate the framework and structured approach for model selection using an NMA of 41 interventions from 17 classes for social anxiety.HighlightsProvides a practical guide and modelling framework for network meta-analysis (NMA) with class effects.Proposes a model selection strategy to guide researchers in choosing appropriate class effect models.Illustrates the strategy using a large case study of 41 interventions for social anxiety.

ObjectivesBayesian multiparameter evidence synthesis (B-MPES) can improve the reliability of long-term survival extrapolations by leveraging registry data. We extended the B-MPES framework to also incorporate historical trial data and examined the impact of alternative external information sources on predictions from early data cuts for a trial in metastatic non-small-cell lung cancer (mNSCLC).MethodsB-MPES models were fitted to survival data from the phase III CheckMate 9LA study of nivolumab plus ipilimumab plus 2 cycles of chemotherapy (NIVO+IPI+CHEMO, v. 4 cycles of CHEMO) in first-line mNSCLC, with 1 y of minimum follow-up. Trial observations were supplemented by registry data from the Surveillance, Epidemiology, and End Results program, general population data, and, optionally, historical trial data with extended follow-up for first-line NIVO+IPI (v. CHEMO) and/or second-line NIVO monotherapy in advanced NSCLC, via estimated 1-y conditional survival. Predictions from the 3 alternative B-MPES models were compared with those from standard parametric models (SPMs).ResultsB-MPES models better anticipated the emergent survival plateau with NIVO+IPI+CHEMO that was apparent in the 4-y data cut compared with SPMs, for which short-term extrapolations in both treatment arms were overly conservative. However, the B-MPES model incorporating NIVO+IPI data slightly overestimated 4-y NIVO+IPI+CHEMO survival owing to a confounding effect on estimated hazards that could not be accounted for a priori until later data cuts of CheckMate 9LA. Extrapolations were relatively robust to the choice of external data sources provided that the prior data had been adjusted to attenuate confounding.ConclusionsIncorporating historical trial data into survival models can improve the plausibility and interpretability of lifetime extrapolations for studies of novel therapies in metastatic cancers when data are immature, and B-MPES provides an appealing method for this purpose.HighlightsLeveraging historical trial data with extended follow-up to extrapolate survival from early study data cuts in a Bayesian evidence synthesis framework can realize anticipated longer-term effects that are characteristic of a novel therapy or class thereof.Using moderately confounded external data sources can improve the reliability of survival extrapolations from B-MPES models provided that the prior information is adjusted and rescaled appropriately, but it is essential to rationalize the implicit assumptions surrounding longer-term treatment effects in the current study.B-MPES models are an attractive option to conduct informed lifetime survival extrapolations based on transparent clinical assumptions via leveraging multiple external data sources, but model flexibility and a priori confidence in external data must be specified carefully to avoid overfitting.

IntroductionSubgroup analyses are vital components of health technology assessments, but randomized controlled trials (RCTs) do not commonly report survival distributions for subgroups. This study developed an analytical framework to elicit unreported subgroup-specific survival curves from aggregate RCT data.MethodsAssuming exponentially distributed subgroup survival durations, we developed an optimization model that approximates the restricted mean survival time (RMST) for the overall population via the weighted average of the RMSTs of 2 subgroups in each arm. Reported hazard ratios from the forest plots between the arms were used to enforce the relationship among subgroups' hazard rates in the model. The performance of the model was tested in a real-life test set of 8 RCTs in advanced-stage gastrointestinal tumors, which also reported KM curves for overall survival (OS) for 40 subgroups as well as in 42 synthetic test cases with 168 subgroups as a benchmark. For each subgroup, predicted median survival, OS rates, and the RMSTs were compared against their actual counterparts as well as their 95% confidence intervals (CIs).ResultsPredicted median survivals and RMSTs were within the 95% CIs of the reported values in 32 (80%) and 34 (85%) of 40 subgroups in real-life test cases and in 163 (97%) and 146 (87%) of 168 subgroups in synthetic test cases, respectively. Across all cases, on average, the predicted survival curves laid within the 95% CIs of reported KM curves 71% and 97% of the time in real-life and synthetic test cases, respectively.DiscussionOur study offers a useful and scalable method for extracting subgroup-specific survival from aggregate RCT data to enable subgroup-specific indirect comparisons, and cost-utility and meta-analyses.HiglightsMost randomized controlled trials report survival curves for the overall patient population but do not provide subgroup-specific survival curves, which are crucial for cost-effectiveness analyses and meta-analyses focusing on these subgroups.This study developed an optimization modeling approach to elicit unreported subgroup-specific survival curves from aggregate trial data.The proposed modeling approach accurately predicted the reported subgroup-specific survival curves in 42 simulated test cases with 168 subgroups overall, in which each subgroup-specific survival curve was assumed to followed an exponential distribution.The performance of the proposed modeling approach was sensitive to the assumptions when it was tested using a real-life test set of 8 oncology trials, which also reported survival curves for a total of 40 subgroups.

PurposeWe conducted a distributional cost-effectiveness analysis (DCEA) using routinely collected data to estimate the population health and health inequality impacts of the National Abdominal Aortic Aneurysm Screening Programme (NAAASP) in England.MethodsAn existing discrete event simulation model of AAA screening was adapted to examine differences between socioeconomic groups defined by Index of Multiple Deprivation, obtained from an analysis of secondary data sources. We examined the distributional cost-effectiveness of being invited versus not invited at age 65 y to screen using a National Health Service perspective. Changes in inequality were valued using a measure of equally distributed equivalent health.ResultsThe net health benefits of population screening (317 quality-adjusted life-years [QALYs] gained) were disproportionately accounted for by the effects on those living in more advantaged areas. The NAAASP improved health on average compared with no screening, but the health opportunity cost of the programme exceeded the QALY gains for people living in the most deprived areas, resulting in a negative net health impact for this group (106 QALYs lost) that was driven by differences in the need for screening. Consequently, the NAAASP increased health inequality at the population level. Given current estimates for inequality aversion in England, screening for AAA remains the optimal strategy.ConclusionExamination of the distributional cost-effectiveness of the NAAASP in England using routinely collected data revealed a tradeoff between total population health and health inequality. Study findings suggest that the NAAASP provides value for money despite health impacts being disseminated to those who are more advantaged.HighlightsThis study examines the population health and health inequality effects of the National Abdominal Aortic Aneurysm Screening Programme (NAAASP) between socioeconomic groups defined by Index of Multiple Deprivation.Findings suggest a tradeoff between total population health and health inequality.Given current estimates for inequality aversion in England, screening remains the optimal strategy relative to not screening.Opportunities remain to reduce inequality effects for those most vulnerable through targeted approaches.

BackgroundCervical cancer, driven predominantly by persistent high-risk human papillomavirus (HPV) infection, ranks as the fourth most common malignancy in women worldwide. China faces barriers to achieving the World Health Organization (WHO) 2030 elimination targets due to low vaccination rates and complex demographics. Strategic intervention optimization is critical for accelerating elimination.MethodsWe developed an age-stratified deterministic compartmental model integrating demographic data and HPV transmission dynamics, capturing heterogeneity in age, sex, sexual activity, and intervention efficacy. The model simulated cervical cancer natural history, including HPV infection, progression to precancerous lesions, and invasive cancer and was calibrated using epidemiological data from the Global Burden of Disease. We evaluated multiple vaccination scenarios (varying coverage rates, age groups, and durations) to project incidence trajectories, estimate elimination timelines, and calculate the reproduction number. Sensitivity analyses were conducted to assess parameter effects.ResultsWithout vaccination, HPV infection becomes endemic (R = 1.38), causing 2.92 million cervical cancer cases in China during 2021 to 2070. Maintaining the 2020 vaccination rate would prevent 1.01 million cases in this period. While prioritizing females aged 15 to 26 y maximizes the per-dose impact, expanding vaccination to all females aged ≥15 y is essential for achieving elimination before 2040. Even single-year vaccination would confer >50-y protection. A higher vaccination rate accelerates elimination: annual rates of 0.09, 0.15, and 0.21 among females aged ≥15 y achieve elimination by 2037, 2035, and 2034, respectively, accelerating timelines by 15 to 20 y compared with strategies targeting only 15- to 26-y-olds.ConclusionsHPV vaccination is pivotal for reducing cervical cancer burden in China, with prioritizing women aged 15 to 26 y as the optimal strategy. Expanding vaccination to all women aged ≥15 y can accelerate the achievement of WHO elimination targets.HighlightsAn age-stratified model simulates HPV transmission patterns and assesses cervical cancer interventions.Without intervention, HPV remains endemic (R = 1.38), causing 2.92 million cervical cancer cases in China (2021-2070).Prioritizing 15- to 26-y-olds maximizes the per-dose impact, but expanding to 15+ y cohorts is essential for elimination.Even a single year of vaccination offers >50 y of protection.Females ≥15 y vaccinated annually at rates of 0.09, 0.15, and 0.21 achieve elimination by 2037, 2035, and 2034, respectively.

BackgroundThe Australian National Bowel Cancer Screening Program (NBCSP), which provides 2-yearly screening to people aged 50 to 74 y, had a phased rollout from 2006 and was fully implemented in 2020. To measure the effectiveness of the NBCSP accounting for age-specific trends, we aimed to develop a novel integrative method to project colorectal cancer (CRC) incidence rates from 2006 to 2045 in the absence of the NBCSP (referred to as "no-NBCSP projections") while addressing the challenge of complex age-specific trends in CRC incidence.MethodsWe constructed a new dataset by replacing the observed data for NBCSP-eligible individuals aged 50 to 74 y with intermediate projections based on pre-NBCSP data from 1982 to 2005. We compared the no-NBCSP CRC incidence projected using a standard age-period-cohort (APC) model, age-stratified APC models, and the integrative modeling approach.ResultsThe integrative modeling approach captured complex age-specific trends better than the standard and age-stratified APC models did. Without the NBCSP, the overall CRC incidence rates would be expected to decline from 2005 to 2025, followed by increases from 2026 to 2045. The incidence rates for those aged <50 y would be projected to continue increasing to 2045, and an increase in incidence rates for older age groups would be projected to occur from 2020 for ages 50 to 54 y, from 2030 for ages 65 to 74 y, and from 2035 for ages 75 y and older.ConclusionsThese no-NBCSP projections provide a counterfactual benchmark against which to measure the impact of the NBCSP on CRC incidence in Australia, and they have been used as new calibration targets for a simulation model of CRC and screening in Australia. The methods developed here could be used to generate comparators to assess the impact of other public health interventions.HighlightsWe constructed counterfactual projections of colorectal cancer (CRC) incidence rates in the absence of the National Bowel Cancer Screening Program (no-NBCSP projections).To do this, we developed a new integrative modeling approach to capture complex age-specific colorectal cancer incidence trends.These no-NBCSP projections provide a counterfactual benchmark against which to measure the impact of the NBCSP on CRC incidence in Australia.These projections stress the need for ongoing assessment of the starting age for the NBCSP, to tackle the increasing incidence for people younger than 50 y.

BackgroundAn artificial intelligence (AI)-enabled rule-out device may autonomously remove patient images unlikely to have cancer from radiologist review. Many published studies evaluate this type of device by retrospectively applying the AI to large datasets and use sensitivity and specificity as the performance metrics. However, these metrics have fundamental shortcomings because sensitivity will always be negatively affected in retrospective studies of rule-out applications of AI.MethodWe reviewed 2 performance metrics to compare the screening performance between the radiologist-with-rule-out-device and radiologist-without-device workflows: positive/negative predictive values (PPV/NPV) and expected utility (EU). We applied both methods to a recent study that reported improved performance in the radiologist-with-device workflow using a retrospective US dataset. We then applied the EU method to a European study based on the reported recall and cancer detection rates at different AI thresholds to compare the potential utility among different thresholds.ResultsFor the US study, neither PPV/NPV nor EU can demonstrate significant improvement for any of the algorithm thresholds reported. For the study using European data, we found that EU is lower as AI rules out more patients including false-negative cases and reduces the overall screening performance.ConclusionsDue to the nature of the retrospective simulated study design, sensitivity and specificity can be ambiguous in evaluating a rule-out device. We showed that using PPV/NPV or EU can resolve the ambiguity. The EU method can be applied with only recall rates and cancer detection rates, which is convenient as ground truth is often unavailable for nonrecalled patients in screening mammography.HighlightsSensitivity and specificity can be ambiguous metrics for evaluating a rule-out device in a retrospective setting. PPV and NPV can resolve the ambiguity but require the ground truth for all patients. Based on utility theory, expected utility (EU) is a potential metric that helps demonstrate improvement in screening performance due to a rule-out device using large retrospective datasets.We applied EU to a recent study that used a large retrospective mammography screening dataset from the United States. That study reported an improvement in specificity and decrease in sensitivity when using their AI as a rule-out device retrospectively. In terms of EU, we cannot conclude a significant improvement when the AI is used as a rule-out device.We applied the method to a European study that reported only recall rates and cancer detection rates. Since there is no established EU baseline value in European mammography screening workflow, we estimated the EU baseline using data from previous literature. We cannot conclude a significant improvement when the AI is used as a rule-out device for the European study.In this work, we investigated the use of EU to evaluate rule-out devices using large retrospective datasets. This metric, used with retrospective clinical data, could be used as supporting evidence for rule-out devices.

BackgroundUpdated estimates of the productivity losses per HIV infection due to premature HIV mortality are needed to help quantify the economic burden of HIV and inform cost-effectiveness analyses.MethodsWe used the human capital approach to estimate the productivity loss due to HIV mortality per HIV infection in the United States, discounted to the time of HIV infection. We incorporated published data on age-specific annual productivity, life expectancy at HIV diagnosis, life-years lost from premature death among persons with HIV (PWH), the number of years from HIV infection to diagnosis, and the percentage of deaths in PWH attributable to HIV. For the base case, we used 2018 life expectancy data for all PWH in the United States. We also examined scenarios using life expectancy in 2010 and life expectancy for cohorts on antiretroviral therapy (ART). We conducted sensitivity analyses to understand the impact of key input parameters.ResultsWe estimated the base-case overall average productivity loss due to HIV mortality per HIV infection at $65,300 in 2022 US dollars. The base-case results showed a 45% decrease in the estimated productivity loss compared with the results when applying life expectancy data from 2010. Productivity loss was 83% lower for cohorts of PWH on ART compared with the base-case scenario. Results were sensitive to assumptions about percentage of deaths attributable to HIV and heterogeneity in age at death.ConclusionThis study provides valuable insights into the economic impact of HIV mortality, illustrating reductions in productivity losses over time due to advancements in treatments.HighlightsUpdated estimates of productivity losses per HIV infection due to premature HIV mortality can help assess the total economic burden of HIV in the United States.This study estimates productivity losses per HIV infection for overall, by sex, and by varying ages of HIV infection.Advancement in treatment has contributed to a significant reduction in productivity losses due to premature HIV mortality in the United States over the past decade.

This article critically examines the application of minimal important differences (MIDs) to health state values or utilities. The concept of MIDs aims to guide clinical and research decisions by identifying important changes in health-related quality-of-life (HRQoL) indicators. However, this cannot be used without additional information not contained within the indicator itself, so that the MID cannot be regarded as a property of the indicator. First, MIDs defined at the individual patient level cannot be meaningfully aggregated for groups without additional context. Second, any improvement in HRQoL is important for patients themselves, so decision making using an MID also requires context, such as resource costs for effecting change. Third, health state values incorporate a measure of importance according to patient preferences, so the only change that is unimportant is zero. Calculating and reporting MIDs for health state values is not only unhelpful but also misleading.HighlightsThe minimal important difference (MID) for health-related quality of life and patient-reported outcome measures is widely used but arguably is not only of limited use but also usually misleading because it lacks context-specific meaning.MIDs for individuals cannot be aggregated without judgments about the distribution of outcomes over patient groups, and quality-of-life indicators need context; thus, the MID cannot be regarded as a property of an indicator.Quality-of-life indicators that generate health state values or utilities incorporate importance based on patient preferences, so the only unimportant change is zero.Published research into MIDs for health state values is unhelpful and even misleading.

I30, J17.

BackgroundPrevious research suggests that physicians' inclination to refer patients for suspected cancer is a relatively stable characteristic of their decision making. We aimed to identify its psychological determinants in the presence of a risk-prediction algorithm.MethodsWe presented 200 UK general practitioners with online vignettes describing patients with possible colorectal cancer. Per the vignette, GPs indicated the likelihood of referral (from highly unlikely to highly likely) and level of cancer risk (negligible/low/medium/high), received an algorithmic risk estimate, and could then revise their responses. After completing the vignettes, GPs responded to questions about their values with regard to harms and benefits of cancer referral for different stakeholders, perceived severity of errors, acceptance of false alarms, and attitudes to uncertainty. We tested whether these values and attitudes predicted their earlier referral decisions.ResultsThe algorithm significantly reduced both referral likelihood ( = -0.06 [-0.10, -0.007], = 0.025) and risk level ( = -0.14 [-0.17, -0.11], < 0.001). The strongest predictor of referral was the value GPs attached to patient benefits ( = 0.30 [0.23, 0.36], < 0.001), followed by benefits ( = 0.18 [0.11, 0.24], < 0.001) and harms ( = -0.14 [-0.21, -0.08], < 0.001) to the health system/society. The perceived severity of missing a cancer vis-à-vis overreferring also predicted referral ( = 0.004 [0.001, 0.007], = 0.009). The algorithm did not significantly reduce the impact of these variables on referral decisions.ConclusionsThe decision to refer patients who might have cancer can be influenced by how physicians perceive and value the potential benefits and harms of referral primarily for patients and the moral seriousness of missing a cancer vis-à-vis over-referring. These values contribute to an internal threshold for action and are important even when an algorithm informs risk judgments.HighlightsPhysicians' inclination to refer patients for suspected cancer is determined by their assessment of cancer risk but also their core values; specifically, their values in relation to the perceived benefits and harms of referrals and the seriousness of missing a cancer compared with overreferring.We observed a moral prioritization of referral decision making, in which considerations about benefits to the patient were foremost, considerations about benefits but also harms to the health system or the society were second, while considerations about oneself carried little or no weight.Having an algorithm informing assessments of risk influences referral decisions but does not remove or significantly reduce the influence of physicians' core values.

BackgroundEconomic evaluations in health typically assume a nonwelfarist framework, arguably better served by preferences elicited from a social perspective than a personal one. However, most health state valuation studies elicit personal preferences, leading to a methodological inconsistency. No studies have directly compared social and personal preferences for outcomes using otherwise identical scenarios, leaving their empirical relationship unclear.AimThis unique study examines whether the choice of eliciting preferences from a social or personal perspective influences valuations of health and well-being outcomes.MethodsUsing discrete choice experiments, social and personal preferences for health and well-being attributes were elicited from the UK general public recruited from an internet panel ( = 1,020 personal, = 3,009 social surveys). Mixed logit models were estimated, and willingness-to-pay (WTP) values for each attribute were calculated to compare differences between the 2 perspectives.ResultsWhile no significant differences were observed in the effects of physical and mental health, loneliness, and neighborhood safety across the 2 perspectives, significant differences emerged in WTP values for employment and housing quality. For instance, other things being the same, personal preferences rate being retired as more preferable than being an informal caregiver, but the social preferences rate them in the reverse order.ConclusionOur findings demonstrate that the perspective matters, particularly for valuing outcomes such as employment and housing. These findings indicate that the exclusive use of personal preferences to value states such as employment and housing quality may potentially lead to suboptimal resource allocation, given that such valuations reflect individual rather than societal benefit. This highlights the importance of considering perspective especially in the resource allocation of public health interventions.HighlightsPersonal preferences were not aligned with social preferences for employment and housing quality outcomes.Respondents valued health outcomes the same in both social and personal perspectives.Using personal preferences in public health resource allocation decisions may not reflect societal priorities.

BackgroundAffective forecasting errors (i.e., errors in people's predictions about future emotions) are common in health decision making and can negatively affect health outcomes. Although narrative interventions have been used to mitigate these errors, many studies did not clearly identify the specific errors targeted or examine the impact of different narrative types on affective forecasting. We applied the narrative immersion model (NIM) to capture the nuances of narratives on mitigating specific affective forecasting errors in health decision making.MethodsUsing a narrative review of existing narrative affective forecasting interventions, we investigated the potential of experience, process, and outcome narratives to reduce specific affective forecasting errors (e.g., focalism, immune neglect).ResultsDifferent narrative types-experience, process, and outcome-may play distinct roles in mitigating affective forecasting errors. Experience narratives may reduce affective forecasting errors by describing what people most likely (targeted) or might (representative) experience, process narratives by modeling optimal decision making, and outcome narratives by broadening people's understanding of possible emotional outcomes. We further discuss how narrative characteristics related to content and structure (e.g., perspective taking, transportation, etc.) may advance narrative effects on affective forecasting.ConclusionsOur findings have implications for intervention design as they facilitate the selection of narrative types tailored to specific affective forecasting errors (e.g., framing, misconstruals, or impact bias).HighlightsSpecific affective forecasting errors may be reduced through different types of narratives, but greater understanding is needed regarding the exact mechanisms.The narrative immersion model is a useful framework to investigate the potential of experience, process, and outcome narratives to reduce specific types of affective forecasting errors.We describe the pathways through which narrative types most likely influence affective forecasting and facilitate the choice of narrative message type for a specific affective forecasting error.Narratives designed for affective forecasting interventions should include detailed and realistic descriptions of people's emotional health care experiences.Other narrative characteristics (e.g., realism, perspective taking, transportation) might affect a person's ability to imagine future emotional health states, and future research should consider their effects on affective forecasting.

IntroductionBreast cancer survivors have a higher risk of interval cancers relative to the screening population. Patient characteristics including features of the primary cancer and its treatment can help predict interval second breast cancer risk, but patient and physician perspectives on how risk prediction tools might enhance surveillance decision making are not well characterized.DesignWe conducted a qualitative study of women with breast cancer who had completed primary treatment and multispecialty physicians recruited through Breast Cancer Surveillance Consortium registries. We conducted semi-structured focus groups with 5 to 7 breast cancer survivors and individual physician interviews. All participants were presented with information about an interval cancer risk prediction tool. We elicited participant perspectives on aspects of the tool's design, relevance, and use for surveillance decision making. Data coding, thematic analysis, and interpretation were guided by the principles of theoretical thematic analysis.ResultsForty physician interviews and 4 focus groups involving 23 breast cancer survivors were analyzed. Two prominent areas of focus emerged: 1) perspectives on how a risk prediction tool would enhance and add value to patient-centered care and 2) risk prediction tools can be a means to improve communication about risk of in-breast recurrence or new breast cancer.ConclusionsThis study provides data on breast cancer survivor and physician perceptions of a new risk prediction tool to support surveillance imaging decisions among breast cancer survivors.ImplicationsAn interval second breast cancer risk prediction tool may promote evidence-based care across an array of physicians and different clinical settings. Future research should identify care delivery settings and features that promote adoption and support use in ways that improve shared decision making and patient outcomes.HighlightsThis qualitative study of breast cancer survivors and physicians found that risk prediction tools to support surveillance decisions were perceived positively when positioned as a supplement to the patient-physician relationship.Both patients and physicians said that a tool supported by strong evidence and accessible outputs would be valuable for shared decision making.

BackgroundCorrectional facilities can act as amplifiers of infectious disease outbreaks. Small community outbreaks can cause larger prison outbreaks, which can in turn exacerbate the community outbreaks. However, strategies for epidemic control in communities and correctional facilities are generally not closely coordinated. We sought to evaluate different strategies for coordinated control.MethodsWe developed a stochastic simulation model of an epidemic spreading across a network of communities and correctional facilities. We parameterized it for the initial phases of the COVID-19 epidemic for 1) California communities and prisons based on community data from covidestim, prison data from the California Department of Corrections and Rehabilitation, and mobility data from SafeGraph, and 2) a small, illustrative network of communities and prisons. For each community or prison, control measures were defined by the intensity of 2 activities: 1) screening to detect and isolate cases and 2) nonpharmaceutical interventions (e.g., masking and social distancing) to reduce transmission. We compared the performance of different control strategies including heuristic and reinforcement learning (RL) strategies using a reward function, which accounted for both the benefit of averted infections and nonlinear cost of the control measures. Finally, we performed analyses to interpret the optimal strategy and examine its robustness.ResultsThe RL control strategy robustly outperformed other strategies including heuristic approaches such as those that were largely used during the COVID-19 epidemic. The RL strategy prioritized different characteristics of communities versus prisons when allocating control resources and exhibited geo-temporal patterns consistent with mitigating prison amplification dynamics.ConclusionRL is a promising method to find efficient policies for controlling epidemic spread on networks of communities and correctional facilities, providing insights that can help guide policy.HighlightsFor modelers, we developed a stochastic simulation model of an epidemic spreading across a network of communities and correctional facilities, and we parameterized it for the initial phases of the COVID-19 epidemic for California communities and prisons in addition to an illustrative network.We compared different control strategies using a reward function that accounted for both the benefit of averted infections and cost of the control measures; we found that reinforcement learning robustly outperformed the other strategies including heuristic approaches such as those that were largely used during the COVID-19 epidemic.For policy makers, our work suggests that they should consider investing in the further development of such methods and using them for the control of future epidemics.We offer qualitative insights into different factors that might inform resource allocation to communities versus prisons during future epidemics.

Individual-level data are routinely used in trial-based economic evaluations to assess the effectiveness and costs of a given intervention. While effectiveness measures are often expressed via utility scores derived from health-related quality-of-life instruments (e.g., EQ-5D questionnaires), information on different types of health care resource use (HRU) measures (e.g., number and types of services) are collected to compute the costs. Partially complete HRU data, particularly for self-reported questionnaires, are handled via ad hoc methods that rely on some assumptions (fill in a zero) that are typically hard to justify. Although methods have been proposed to account for the uncertainty surrounding missing data, particularly in the form of multiple imputation or Bayesian methods, these have mostly been implemented at the level of costs at different times or over the entire study period, while little attention has been given to how missing values at the level of HRUs should be addressed and their implications on the final analysis. We present a general Bayesian framework for the analysis of partially observed HRUs in trial-based economic evaluations, which can accommodate the typical complexities of the data (e.g., excess zeros, skewness, missingness) and quantify the impact of missingness uncertainty on the results. We show the benefits of our approach with a motivating example and compare the results to those from more standard analyses fitted at the level of cost variables after adopting some ad hoc imputation. This article highlights the importance of adopting a comprehensive modeling approach to handle partially observed HRU data in economic evaluations and the strategic advantages of building these models within a Bayesian framework.HighlightsMissing health care service data in trial-based economic evaluations are often removed or imputed using quite restrictive assumptions (e.g., no use of service).We propose a flexible Bayesian approach to account for missing health care service uncertainty and compare the results with models fitted at more aggregated levels (e.g., total costs) using a real case study.Our results show that, depending on the (assumed) missingness assumptions and the level of data aggregation at which analyses are performed, results may be considerably changed.When feasible, analyses should be conducted at the most disaggregated level to ensure that all available information collected in the trial is used in the analysis without relying on (often) restrictive ad hoc imputation approaches.

BackgroundReal-world data can inform health care decisions by allowing the evaluation of nuanced treatment strategies. Longitudinal observational data enable the assessment of dynamic treatment regimes (DTRs), strategies that adapt treatment over time based on patient history, but require causal inference methods to address time-varying confounding. Longitudinal targeted minimum loss-based estimation (LTMLE) is a machine learning-based double-robust approach for improved causal effect estimation.MethodsWe applied LTMLE to longitudinal registry data to evaluate the impact of erythropoiesis-stimulating agents (ESAs) in the clinical management of low to intermediate-1 risk myelodysplastic syndrome (MDS). We defined DTRs based on clinically relevant decision rules (e.g., commencing treatment when the hemoglobin level falls below a threshold) and compared them to static treatment regimes (always or never giving ESAs). Outcomes include mortality and health-related quality of life measured by EQ-5D scores.ResultsThe static regime of never administering ESAs resulted in declining counterfactual EQ-5D scores and increasing mortality risk over time. In contrast, both the static regime of continuous administration of ESAs and the use of dynamic regimes improved the EQ-5D scores and tended to reduce mortality, although the mortality differences were not statistically significant.ConclusionsThe article provides a case study application of the LTMLE method to evaluate realistic treatment policies under time-varying confounding. The findings support the potential benefits of dynamic treatment strategies for the management of MDS, highlighting the importance of personalized treatment adaptation. The study contributes methodological insights into the applications of LTMLE in small-sample, long-follow-up settings relevant to health technology assessment and policy making.HighlightsThis study applies the longitudinal targeted minimum loss estimation (LTMLE) method to evaluate the causal effect of static and dynamic treatment strategies using longitudinal observational data.We demonstrate the use of the LTMLE method to assess the impact of erythropoiesis stimulating agents (ESAs) on quality of life and mortality in patients with low to intermediate-1 risk myelodysplastic syndromes.The findings suggest that patients treated under dynamic ESA treatment regimes show an improved quality of life measured by EQ-5D scores and survival compared with those treated under the static treatment regime of never administering ESAs.This study contributes to the methodological literature by showcasing the application of the LTMLE method in a small-sample, long-follow-up setting with time-varying confounding, informing health technology assessment and policy decisions.