IntroductionEpidemiological models benefit from incorporating detailed time-to-event data to understand how disease risk evolves. For example, decompensation risk in liver cirrhosis depends on sojourn time spent with cirrhosis. Semi-Markov and related models capture these details by modeling time-to-event distributions based on published survival data. However, implementations of semi-Markov processes rely on Monte Carlo sampling methods, which increase computational requirements and introduce stochastic variability. Explicitly calculating the evolving transition likelihood can avoid these issues and provide fast, reliable estimates.MethodsWe present the sojourn time density framework for computing semi-Markov and related models by calculating the evolving sojourn time probability density as a system of partial differential equations. The framework is parametrized by commonly used hazard and models the distribution of current disease state and sojourn time. We describe the mathematical background, a numerical method for computation, and an example model of liver disease.ResultsModels developed with the sojourn time density framework can directly incorporate time-to-event data and serial events in a deterministic system. This increases the level of potential model detail over Markov-type models, improves parameter identifiability, and reduces computational burden and stochastic uncertainty compared with Monte Carlo methods. The example model of liver disease was able to accurately reproduce targets without extensive calibration or fitting and required minimal computational burden.ConclusionsExplicitly modeling sojourn time distribution allows us to represent semi-Markov systems using detailed survival data from epidemiological studies without requiring sampling, avoiding the need for calibration, reducing computational time, and allowing for more robust probabilistic sensitivity analyses.HighlightsTime-inhomogeneous semi-Markov models and other time-to-event–based modeling approaches can capture risks that evolve over time spent with a disease.We describe an approach to computing these models that represents them as partial differential equations representing the evolution of the sojourn time probability density.This sojourn time density framework incorporates complex data sources on competing risks and serial events while minimizing computational complexity.

BackgroundWhile structured expert elicitation (SEE) is gaining traction in health technology assessment in situations in which data are scarce, its application in practice remains limited. Co-designing a practical and fit-for-purpose SEE with experts could enhance its acceptability and feasibility in clinical research.ObjectivesAn SEE was co-designed with clinicians to elicit expert opinions on 3 uncertain quantities of interest (QoIs) for a decision-analytic model in exercise oncology.MethodsA series of co-design meetings was convened to design 6 elicitation stages. Individual elicitation was conducted using the variable interval method (VIM), via videoconferencing. Linear pooling was adopted to generate group estimates. Semi-structured interviews were conducted after the elicitation exercise to gather the experts’ first-hand experience of the elicitation process and to identify areas for improvement. Qualitative data were transcribed and content analyzed.ResultsTwelve experts participated in the co-designed SEE. Three beta distributions were derived and estimated from the experts’ responses: the relative risk reduction of cardiovascular events of exercise for women who survived early-stage endometrial cancer (Mean: 0.362, SD: 0.15), the probability that a clinician would refer a patient to the exercise program (Mean: 0.457, SD: 0.218), and the probability that a cancer patient would use such a health service upon referral (Mean: 0.446, SD: 0.203). Most of the experts’ first-hand experience of the co-designed SEE was positive. The qualitative feedback highlighted critical aspects of the elicitation process that should be designed and executed with caution when targeting clinicians with no prior experience of SEE.ConclusionsThis is the first expert elicitation conducted in exercise oncology. Engaging diverse stakeholders through co-design meetings and incorporating qualitative feedback proved effective and practical in introducing expert elicitation into clinical research.HighlightsRecent SEE guidelines aim to facilitate the conduct of expert elicitation in model-based economic evaluation, but its application in practice remains limited.Engaging experts in the design of SEE could enhance its acceptability and feasibility in clinical research.This is the first co-designed expert elicitation involving clinicians in the field of exercise oncology.This practical approach to conducting SEE could promote a wider adoption to inform health care policy decisions when the evidence is lacking or uncertain.

BackgroundEvidence is lacking on the most effective values clarification methods (VCMs) in patient decision aids (PtDAs). We tested the effects of an adaptive conjoint analysis (ACA)–based VCM compared with a ranking-based VCM and no VCM on several decision-related outcomes, with the decisional conflict and its subscale “perceived values clarity” as primary outcomes.DesignOnline experimental study with 3 conditions: no VCM versus ranking-based VCM versus ACA-based VCM (N = 282; Mage = 63.11 y, s = 12.12), with the latter 2 conditions including attributes important for a lung cancer treatment decision. We assessed 1) decisional conflict, 2) perceived values clarity (decisional conflict subscale), 3) perceived cognitive load, 4) anticipated regret, 5) ambivalence, 6) preparedness for decision making, 7) hypothetical treatment preference, and 8) values congruence (proxy). We performed analysis of covariance and linear regression. Age and level of deliberation were included as potential moderators, and we controlled for subjective numeracy (covariate). We exploratively tested the moderating effects of subjective numeracy and health literacy (without covariates).ResultsWe found no significant effect of type of VCM on overall decisional conflict or perceived values clarity. Age had a moderating effect: in younger participants, no VCM (v. ranking-based VCM) led to more values clarity, while in older participants, a ranking-based VCM (v. no VCM) led to more values clarity. Completing the ACA-based VCM, compared with no VCM, resulted in more values congruence.LimitationsThe hypothetical choice situation might have induced lower levels of cognitive/affective involvement in the decision.ConclusionsThis study found mixed effects of an ACA-based VCM. It did not decrease decisional conflict or increase perceived values clarity, yet it did improve values congruence.ImplicationsCompletion of an ACA-based VCM in a PtDA may increase values congruence.HighlightsAn adaptive conjoint analysis or a ranking-based values clarification method did not decrease analog patients’ decisional conflict nor did it increase their perceived values clarity.In younger participants, no VCM (v. ranking-based VCM) led to more values clarity, while in older participants, a ranking-based VCM (v. no VCM) led to more values clarity.An adaptive conjoint analysis task for values clarification resulted in more values congruence.

This chapter shows how to combine exponential survival models and actuarial survival models to establish a credible representation of long life will last for a patient. Returning to what will be covered in the chapter, the discussion starts with the common properties of all survival models. The uncertainty about the length of the patient's life with an outcome is expressed by the lifetime probabilities. One group consists of women dying from breast cancer. The other group consists of women dying from other causes. Like all parametric models, the validity of the exponential survival model rests on an assumption. The chapter describes an actuarial survival model that can be combined with the exponential survival model to more closely matched survival observations. Determining an outcome utility requires knowing the probabilities for the possible lengths of life.

Restricted accessAbstractFirst published online January 11, 202418th Biennial European Conference of the Society for Medical Decision Making Berlin, Germany, May 21–23, 2023OnlineFirsthttps://doi.org/10.1177/0272989X231225267

Health utilities from value sets for the EQ-5D-5L are commonly used in economic evaluations. We examined whether modeling spatial correlation among health states could improve the precision of the value sets. Using data from 7 EQ-5D-5L valuation studies, we compared the predictive precision of the published linear model, a recently proposed cross-attribute level effects (CALE) model, and 2 Bayesian models with spatial correlation. Predictive precision was quantified through the root mean squared error (RMSE) for out-of-sample predictions of state-level mean utilities on omitting individual states, as well as omitting blocks of states. In all 7 countries, on omitting single health states, Bayesian models with spatial correlation improved upon the published linear model: the RMSEs for the originally published models, 0.050, 0.051, 0.060, 0.061, 0.039, 0.050, and 0.087 for Canada, China, Germany, Indonesia, Japan, Korea, and the Netherlands, respectively, could be reduced to 0.043, 0.042, 0.051, 0.054, 0.037, 0.037, and 0.085, respectively. On omitting blocks of health states, Bayesian models with spatial correlation led to smaller RMSEs in 3 countries, while the CALE model led to smaller RMSEs in the remaining 4 countries. Bayesian models incorporating spatial correlation and CALE models are promising for improving the precision of value sets for the EQ-5D-5L. The differential performance of the Bayesian models on omitting single states versus blocks of states suggests that designing valuation studies to capture more health states may further improve precision. We suggest that Bayesian and CALE models be considered as candidates when creating value sets and that alternative designs be explored; this is vital as the prediction errors in value sets need to be smaller than the minimal important difference of the instrument. The accuracy of value sets of multi-attribute utility instruments is typically of the same order of magnitude as the instrument's minimal important difference and would benefit from improvement.Bayesian models with spatial correlation have been shown to improve value set accuracy in isolated cases.We showed that Bayesian approaches with spatial correlation improved predictive precision in 7 EQ-5D-5L valuation studies.We recommend that Bayesian models incorporating spatial correlation be considered when creating value sets and have provided code for fitting them.

Significant health disparities exist for trans and gender-nonconforming patients, in part caused by a reduced likelihood of seeking health care due to experienced discrimination in health care settings. Increasing inclusivity in patient-provider interactions can decrease barriers to care seeking for these patients, but such advice to providers must be simple to implement and account for potential backlash among certain patient subgroups. In 3 studies, we use online samples to experimentally test patient reactions to 2 methods of provider inclusivity signaling: verbal sharing of the provider's pronouns and request for the patient's pronouns (a method frequently suggested in the literature) and the provider wearing a pin indicating their pronouns without verbal sharing or requesting that the patient share their pronouns. Study 1 finds that political orientation significantly moderated the effect of pronoun request on patient satisfaction with the visit, P = .003; satisfaction with the doctor, P = .003; and willingness to return to the provider, P = .007, with politically liberal participants being more satisfied and more likely to return when pronouns were requested and politically conservative participants less satisfied and less likely to return. Study 2 replicated these findings and demonstrated that such backlash among conservative participants did not occur when inclusivity was indicated via the provider wearing a pin indicating their pronouns. Study 3 showed that while pronoun pins may be subtle enough to not cause backlash, they are noticed by the target community and increase satisfaction and willingness to return among participants who identify as trans, nonbinary, and genderqueer. These studies were hypothetical scenario studies run online rather than field studies. These results suggest that recommendations for increasing inclusivity should account for potential negative impacts on the practice from more politically conservative patients and that more subtle methods of indicating an inclusive health care environment are likely to be more implementable for practitioners in the field while still positively affecting the target population. Increasing inclusivity in health care settings by verbally sharing and requesting pronouns can cause backlash and aggression among politically conservative patients.More subtle methods of inclusivity signaling such as providers wearing pins sharing their own pronouns do not have this type of backlash effect and increase satisfaction and willingness to return among non-cisgender patients.Recommendations for increasing inclusivity in health care settings should consider negative reactions from other patient subgroups to ensure recommendations are those providers are likely to be willing and able to implement long term.

BackgroundIn decision modeling with time-to-event data, there are a variety ofparametric models that can be used to extrapolate the survival function.Each model implies a different hazard function, and in situations in whichthere is moderate censoring, this can result in quite different survivalprojections. External information such as expert opinion on long-termsurvival can more accurately characterize the uncertainty in theseextrapolations.ObjectiveWe present a general and easily implementable approach to incorporate varioustypes of expert opinions into parametric survival models, focusing onopinions about survival at various landmark time points.MethodsExpert opinion is incorporated into parametric survival models using Bayesianand frequentist approaches. In the Bayesian method, expert opinion isincluded through a loss function and in the frequentist approach bypenalizing the likelihood function, although in both cases the core approachis the same. The issue of aggregating multiple expert opinions is alsoconsidered.ResultsWe apply this method to data from a leukemia trial and use previouslyelicited expert opinion on survival probabilities for that particular trialpopulation at years 4 and 5 to inform our analysis. We take a robustapproach to modeling expert opinion by using pooled distributions and fit abroad class of parametric models to the data. We also assess statisticalgoodness of fit of the models to both the observed data and expertopinion.ConclusionsExpert opinions can be implemented in a straightforward manner using thisnovel approach; however, more work is required on the correct elicitation ofthese quantities.HighlightsPresentation of a novel and open-source method to incorporate expertopinion into decision modeling.Extends upon earlier work in that expert opinion can be incorporatedinto a wide range of parametric models.Provides methodological guidance for directly including expertopinion in decision modeling, which is a research focus area in NICETSD 21.1

PurposeThe goal of this study is to develop an approach for estimating nationallyrepresentative quality-adjusted life-year (QALY) loss from injury andpoisoning conditions using data collected in the Medical Expenditure PanelSurvey (MEPS) and the National Health Interview Survey (NHIS).MethodsThis study uses data from the 2002–2015 NHIS and MEPS surveys. Injuries wereidentified in the MEPS medical events file and through self-reporting ofmedical conditions. We restricted our model to 163,731 adults, for which wepredict a total of 294,977 EQ-5D scores using responses to theself-administered questionnaire. EQ-5D scores were modeled using age, sex,comorbidities, and binary indicators of the presence and duration of injuryat the time of the health status questionnaire. These models considernonlinearity over time during the first 3 y following the injury event.ResultsInjuries are identified in MEPS using medical events that provide areasonable proxy for the date of injury occurrence. Health-related qualityof life (HRQL) decrements can be estimated using binary indicators of injuryduring different time periods. When grouped into 29 injury categories, mostcategories were statistically significant predictors of HRQL scores in thefirst year after injury. For these groups of injuries, mean first-year QALYloss estimates range from 0.005 (sprains and strains of joints and adjacentmuscles, n = 7067) to 0.109 (injury to nerves and spinalcord, n = 71). Fewer estimates are significant in thesecond and third years after injury, which may reflect a return to baselineHRQL.ConclusionThis research presents both a framework for estimating QALY loss forshort-lived medical conditions and nationally representative,community-based HRQL scores associated with a wide variety of injury andpoisoning conditions.HighlightsThis research provides a catalog of nationally representative,preference-based EQ-5D score decrements associated with surviving alarge set of injuries, based on patient-reported health status.Mean first-year QALY loss estimates range from 0.005 (sprains andstrains of joints and adjacent muscles, n = 7067)to 0.109 (injury to nerves and spinal cord, n =71).This article presents a novel methodology for assessingquality-of-life impacts for acute conditions by calculating the timeelapsed between injury and health status elicitation. Researchersmay explore adapting these methods to study other short-livedconditions and health states, such as COVID-19 or chemotherapy.

BackgroundDigital health interventions (DHIs) can improve the provision of health care services. To fully account for their effects in economic evaluations, traditional methods based on measuring health-related quality of life may not be appropriate, as nonhealth and process outcomes are likely to be relevant too.PurposeThis systematic review identifies, assesses, and synthesizes the arguments on the analytical frameworks and outcome measures used in the economic evaluations of DHIs. The results informed recommendations for future economic evaluations.Data SourcesWe ran searches on multiple databases, complemented by gray literature and backward and forward citation searches.Study SelectionWe included records containing theoretical and empirical arguments associated with the use of analytical frameworks and outcome measures for economic evaluations of DHIs. Following title/abstract and full-text screening, our final analysis included 15 studies.Data ExtractionThe arguments we extracted related to analytical frameworks (14 studies), generic outcome measures (5 studies), techniques used to elicit utility values (3 studies), and disease-specific outcome measures and instruments to collect health states data (both from 2 studies).Data SynthesisRather than assessing the quality of the studies, we critically assessed and synthesized the extracted arguments. Building on this synthesis, we developed a 3-stage set of recommendations in which we encourage the use of impact matrices and analyses of equity impacts to integrate traditional economic evaluation methods.LimitationsOur review and recommendations explored but not fully covered other potentially important aspects of economic evaluations that were outside our scope.ConclusionsThis is the first systematic review that summarizes the arguments on how the effects of DHIs could be measured in economic evaluations. Our recommendations will help design future economic evaluations.HighlightsUsing traditional outcome measures based on health-related quality of life (such as the quality-adjusted life-year) may not be appropriate in economic evaluations of digital health interventions, which are likely to trigger nonhealth and process outcomes.This is the first systematic review to investigate how the effects of digital health interventions could be measured in economic evaluations.We extracted and synthesized different arguments from the literature, outlining advantages and disadvantages associated with different methods used to measure the effects of digital health interventions.We propose a methodological set of recommendations in which 1) we suggest that researchers consider the use of impact matrices and cost-consequence analysis, 2) we discuss the suitability of analytical frameworks and outcome measures available in economic evaluations, and 3) we highlight the need for analyses of equity impacts.