Medical Decision Making

Background. The Decisional Conflict Scale (DCS) measures 5 dimensions of decision making (feeling: uncertain, uninformed, unclear about values, unsupported; ineffective decision making). We examined the use of the DCS over its initial 20 years (1995 to 2015). Methods. We conducted a scoping review with backward citation search in Google Analytics/Web of Science/PubMed, followed by keyword searches in Cochrane Library, PubMed, Ovid MEDLINE, EMBASE, CINAHL, AMED, PsycINFO, PRO-Quest, and Web of Science. Eligible studies were published between 1995 and March 2015, used an original experimental/observational research design, concerned a health-related decision, and provided DCS data (total/subscales). Author dyads independently screened titles, abstracts, full texts, and extracted data. We performed narrative data synthesis. Results. We included 394 articles. DCS use appeared to increase over time. Three hundred nine studies (76%) used the original DCS, and 29 (7%) used subscales only. Most studies used the DCS to evaluate the impact of decision support interventions (n = 238, 59%). The DCS was translated into 13 languages. Most decisions were made by people for themselves (n = 353, 87%), about treatment (n = 225, 55%), or testing (n = 91, 23%). The most common decision contexts were oncology (n = 113, 28%) and primary care (n = 82, 20%). Conclusions. This is the first study to descriptively synthesize characteristics of DCS data. Use of the DCS as an outcome measure for health decision interventions has increased over its 20-year existence, demonstrating its relevance as a decision-making evaluation measure. Most studies failed to report when decisional conflict was measured during the decision-making process, making scores difficult to interpret. Findings from this study will be used to update the DCS user manual.

Background. We explored decisional conflict as measured with the 16-item Decisional Conflict Scale (DCS) and how it varies across clinical situations, decision types, and exposure to decision support interventions (DESIs). Methods. An exhaustive scoping review was conducted using backward citation searches and keyword searches. Eligible studies were published between 1995 and March 2015, used an original experimental/observational research design, concerned a health-related decision, and provided DCS data. Dyads independently screened titles/abstracts and full texts, and extracted data. We performed narrative syntheses and calculated average or median DCS scores. Results. We included 246 articles reporting on 253 studies. DCS scores ranged from 2.4 to 79.7 out of 100. Highest baseline DCS scores were for care planning (44.8 ± 8.9, median = 47.0) and treatment decisions (32.5 ± 12.6, median = 31.9), in contexts of primary care (40.6 ± 18.3), and geriatrics (39.8 ± 11.2). Baseline scores were high among decision makers who were ill (33.2 ± 14.1, median = 30.2) or making decisions for themselves (33.4 ± 13.8, median = 32.0). Total DCS scores <25 out of 100 were associated with implementing decisions. Without DESIs, DCS scores tended to increase shortly after decision making (>37.4). After DESI use, DCS scores decreased short-term but increased or remained the same long-term (>6 months). Conclusions. DCS scores were highest at baseline and decreased after decision making. DESIs decreased decisional conflict immediately after decision making. The largest improvements after DESIs were in decision makers who were ill or made decisions for themselves. Further meta-analyses are needed for decision type, contexts, and interventions to inform hypotheses about the expected effects of DESIs, the best timing for measurement, and interpretation of DCS scores.

Background. Musculoskeletal conditions are leading causes of disability. Management options are plentiful, but the current evidence base suggests many are ineffective or unproven. Online decision aids can help support patients make informed health care choices. However, there are little data on the quality of online decision aids for common musculoskeletal conditions such as knee or low back pain. Purpose. To identify all publicly available online decision aids for knee osteoarthritis and low back pain and evaluate them against the International Patient Decision Aids Standards Inventory (IPDASi). Data Sources. Google Australia. Study selection. Two reviewers independently screened websites for inclusion and assessed the quality of included online decision aids between April and May 2018. Included online decision aids were free, provided information about knee osteoarthritis or low back pain, and written in English. Online decision aids that required payment, targeted health professionals, addressed rheumatoid arthritis, or addressed a screening decision were excluded. Data Extraction. IPDASi Version 4. Data Synthesis. Twenty-five online decision aids were identified: 15 knee osteoarthritis and 10 low back pain. Only 3 online decision aids (12%) provided a “wait-and-see” option. Nineteen (75%) met IPDASi criteria to be considered a decision aid and 3 (12%) met IPDASi criteria to state that the online decision aid was unbiased. Limitations. Dynamic nature of Google searches may not be replicable easily. Conclusions. Few good-quality online decision aids are available for people with knee osteoarthritis or low back pain. Most online decision aids failed to explicitly provide a wait-and-see option, suggesting a bias toward intervention. These online decision aids would benefit from explicitly highlighting a wait-and-see option to support informed choice.

Kidney exchanges were developed to match kidney failure patients with willing but incompatible donors to other donor-patient pairs. Finding a match in a large candidate pool can be modeled as an integer program. However, these exchanges accumulate participants with characteristics that increase the difficulty of finding a match and, therefore, increase patients’ waiting time. Therefore, we sought to fine-tune the formulation of the integer program by more accurately assigning priorities to patients based on their difficulty of matching. We provide a detailed formulation of prioritized kidney exchange and propose a novel prioritization algorithm. Our approach takes advantage of the global knowledge of the donor-patient compatibility within a pool of pairs and calculates an iterative, paired match power (iPMP) to represent the donor-patient pairs’ abilities to match. Monte Carlo simulation shows that an algorithm using the iPMP reduces the waiting time more than using paired match power (PMP) for the difficult-to-match pairs with hazard ratios of 1.3480 and 1.1100, respectively. Thus, the iPMP may be a more accurate assessment of the difficulty of matching a pair in a pool than PMP is, and its use may improve matching algorithms being used to match donors and recipients.

Background. The expected value of sample information (EVSI) determines the economic value of any future study with a specific design aimed at reducing uncertainty about the parameters underlying a health economic model. This has potential as a tool for trial design; the cost and value of different designs could be compared to find the trial with the greatest net benefit. However, despite recent developments, EVSI analysis can be slow, especially when optimizing over a large number of different designs. Methods. This article develops a method to reduce the computation time required to calculate the EVSI across different sample sizes. Our method extends the moment-matching approach to EVSI estimation to optimize over different sample sizes for the underlying trial while retaining a similar computational cost to a single EVSI estimate. This extension calculates the posterior variance of the net monetary benefit across alternative sample sizes and then uses Bayesian nonlinear regression to estimate the EVSI across these sample sizes. Results. A health economic model developed to assess the cost-effectiveness of interventions for chronic pain demonstrates that this EVSI calculation method is fast and accurate for realistic models. This example also highlights how different trial designs can be compared using the EVSI. Conclusion. The proposed estimation method is fast and accurate when calculating the EVSI across different sample sizes. This will allow researchers to realize the potential of using the EVSI to determine an economically optimal trial design for reducing uncertainty in health economic models. Limitations. Our method involves rerunning the health economic model, which can be more computationally expensive than some recent alternatives, especially in complex models.

Research reveals a biased preference for natural v. synthetic drugs; however, this research is based on self-report and has not examined ways to reduce the bias. We examined these issues in 5 studies involving 1125 participants. In a pilot study (N = 110), participants rated the term natural to be more positive than the term synthetic, which reveals a default natural-is-better belief. In studies 1 (N = 109) and 2 (N = 100), after a supposed personality study, participants were offered a thank you “gift” of a natural or synthetic pain reliever. Approximately 86% (study 1) and 93% (study 2) of participants chose the natural v. synthetic pain reliever, which provides a behavioral choice confirmation of the natural drug bias. In studies 3 (N = 350) and 4 (N = 356), participants were randomly assigned to a control or experimental condition and were asked to consider a scenario in which they had a medical issue requiring a natural v. synthetic drug. The experimental condition included a stronger (study 3) or weaker (study 4) rational appeal about the natural drug bias and a statement suggesting that natural and synthetic drugs can be good or bad depending on the context. In both studies, the natural bias was reduced in the experimental condition, and perceived safety and effectiveness mediated this effect. Overall, these data indicate a bias for natural over synthetic drugs in preferences and behavioral choices, which might be reduced with a rational appeal.

Cost-utility analysis (CUA) is a widely recommended form of health economic evaluation worldwide. The outcome measure in CUA is quality-adjusted life-years (QALYs), which are calculated using health state utility values (HSUVs) and corresponding life-years. Therefore, HSUVs play a significant role in determining cost-effectiveness. Formal adoption and endorsement of CUAs by reimbursement authorities motivates methodological advancement in HSUV measurement and application. A large body of evidence exploring various methods in measuring HSUVs has accumulated, imposing challenges for investigators in identifying and applying HSUVs to CUAs. First, large variations in HSUVs between studies are often reported, and these may lead to different cost-effectiveness conclusions. Second, issues concerning the quality of studies that generate HSUVs are increasingly highlighted in the literature. This issue is compounded by the limited published guidance and methodological standards for assessing the quality of these studies. Third, reimbursement decision making is a context-specific process. Therefore, while an HSUV study may be of high quality, it is not necessarily appropriate for use in all reimbursement jurisdictions. To address these issues, by promoting a systematic approach to study identification, critical appraisal, and appropriate use, we are developing the Health Utility Book (HUB). The HUB consists of an HSUV registry, a quality assessment tool for health utility studies, and a checklist for interpreting their use in CUAs. We anticipate that the HUB will make a timely and important contribution to the rigorous conduct and proper use of health utility studies for reimbursement decision making. In this way, health care resource allocation informed by HSUVs may reflect the preferences of the public, improve health outcomes of patients, and maintain the efficiency of health care systems.

Background. The EQ-5D instrument has 5 dimensions. This article reports on the effects of manipulating a) the order in which the 5 dimensions are presented (appearing first v. last), b) splitting of the composite dimensions (“pain or discomfort” and “anxiety or depression”), and c) removing or “bolting off” 1 of the 5 EQ-5D dimensions at a time. The effects were examined in 2 contexts: 1) self-reporting health and 2) health state valuations. Methods. Three different types of discrete choice experiments (DCE) including a duration attribute were designed. An online survey with 12 subtypes, each with 10 DCE tasks, was designed and completed by 2494 members of the UK general public. Results. Of the 3 manipulations in the self-reporting context, only b) splitting anxiety or depression had a significant effect. In the health state valuation context, b) splitting level 5 pain or discomfort (relative to pain) and splitting level 5 anxiety or depression (relative to anxiety) had significant effects as did c) bolting off dimensions. Conclusions. We find that the values given to certain health dimensions are sensitive to the way in which it is described and the other health dimensions presented. Of particular interest is the effect of splitting composite dimensions: a given EQ-5D(-5L) profile may mean different things depending on whether the profile is used to self-report one’s health or to value hypothetical states, so that the health state values of EQ-5D(-5L) in population tariffs may not correspond to the states that patients self-report themselves in.

Objectives. To evaluate the impact of adding a respiratory dimension (a bolt-on dimension) to the EQ-5D-5L health state valuations. Methods. Based on extensive regression and principal component analyses, 2 respiratory bolt-on candidates were formulated: R1, limitations in physical activities due to shortness of breath, and R2, breathing problems. Valuation interviews for the selected bolt-ons were performed with a representative sample from the Dutch general public using the standardized interview protocol and software of the EuroQol group. Hybrid models based on the combined time-tradeoff (TTO) and discrete choice experiment (DCE) data were estimated to assess whether the 5 levels of the respiratory bolt-on led to significant changes in utility values. Results. For each bolt-on candidate, slightly more than 200 valuation interviews were conducted. Mean TTO values and DCE choice probabilities for health states with a level 4 or 5 for the respiratory dimension were significantly lower compared with the same health states in the Dutch EQ-5D-5L valuation study without the respiratory dimension. Results of hybrid models showed that for the bolt-on “limitations in physical activities,” the utility decrements were significant for level 3 (–0.055), level 4 (–0.087), and level 5 (–0.135). For “breathing problems,” the utility decrements for the same levels were greater (–0.086, –0.219, and –0.327, respectively). Conclusions. The addition of each of the 2 respiratory bolt-ons to the EQ-5D-5L had a significant effect on the valuation of health states with severe levels for the bolt-on. The bolt-on dimension “breathing problems” showed the greatest utility decrements and therefore seems the most appropriate respiratory bolt-on dimension.

Background. Microsimulation models have been extensively used in the field of cancer modeling. However, there is substantial uncertainty regarding estimates from these models, for example, overdiagnosis in prostate cancer. This is usually not thoroughly examined due to the high computational effort required. Objective. To quantify uncertainty in model outcomes due to uncertainty in model parameters, using a computationally efficient emulator (Gaussian process regression) instead of the model. Methods. We use a microsimulation model of prostate cancer (microsimulation screening analysis [MISCAN]) to simulate individual life histories. We analyze the effect of parametric uncertainty on overdiagnosis with probabilistic sensitivity analyses (ProbSAs). To minimize the number of MISCAN runs needed for ProbSAs, we emulate MISCAN, using data pairs of parameter values and outcomes to fit a Gaussian process regression model. We evaluate to what extent the emulator accurately reproduces MISCAN by computing its prediction error. Results. Using an emulator instead of MISCAN, we may reduce the computation time necessary to run a ProbSA by more than 85%. The average relative prediction error of the emulator for overdiagnosis equaled 1.7%. We predicted that 42% of screen-detected men are overdiagnosed, with an associated empirical confidence interval between 38% and 48%. Sensitivity analyses show that the accuracy of the emulator is sensitive to which model parameters are included in the training runs. Conclusions. For a computationally expensive simulation model with a large number of parameters, we show it is possible to conduct a ProbSA, within a reasonable computation time, by using a Gaussian process regression emulator instead of the original simulation model.

Background. Several methods, such as the half-cycle correction and the life-table method, were developed to attenuate the error introduced in Markov models by the discretization of time. Elbasha and Chhatwal have proposed alternative “corrections” based on numerical integration techniques. They present an example whose results suggest that the trapezoidal rule, which is equivalent to the half-cycle correction, is not as accurate as Simpson’s 1/3 and 3/8 rules. However, they did not take into consideration the impact of discontinuities. Objective. To propose a method for evaluating Markov models with discontinuities. Design. Applying the trapezoidal rule, we derive a method that consists of adjusting the model by setting the cost at each point of discontinuity to the mean of the left and right limits of the cost function. We then take from the literature a model with a cycle length of 1 year and a discontinuity on the cost function and compare our method with other “corrections” using as the gold standard an equivalent model with a cycle length of 1 day. Results. As expected, for this model, the life-table method is more accurate than assuming that transitions occur at the beginning or the end of cycles. The application of numerical integration techniques without taking into account the discontinuity causes large errors. The model with averaged cost values yields very small errors, especially for the trapezoidal and the 1/3 Simpson rules. Conclusion. In the case of discontinuities, we recommend applying the trapezoidal rule on an averaged model because this method has a mathematical justification, and in our empirical evaluation, it was more accurate than the sophisticated 3/8 Simpson rule.

Background. In health economics, there has been interest in using discrete-choice experiments (DCEs) to derive preferences for health states in lieu of previously established approaches like time tradeoff (TTO). We examined whether preferences elicited through DCEs are associated and agree with preferences elicited through TTO tasks. Methods. We used data from 1073 respondents to the Canadian EQ-5D-5L valuation study. Multivariate mixed-effects models specified a common likelihood for the TTO and discrete-choice data, with separate but correlated random effects for the TTO and DCE data, for each of the 5 EQ-5D-5L dimensions. Multivariate latent class models allowed separate but associated latent classes for the DCE and TTO data. Results. Correlation between the random effects for the 2 tasks ranged from −0.12 to 0.75, with only pain/discomfort and anxiety/depression having at least a 50% posterior probability of strong (>0.6) correlation. Latent classes for the TTO and DCE data both featured 1 latent class capturing participants attaching large disutilities to pain/discomfort, another capturing participants attaching large disutility to anxiety/depression, and the third class capturing the remainder. Agreement in class membership was poor (κ coefficient: 0.081; 95% credible interval, 0.033–0.13). Fewer respondents expressed strong disutilities for problems with anxiety/depression or pain/discomfort in the TTO than the DCE data (17% v. 55%, respectively). Conclusions. Stated preferences using TTO and DCEs show association across dimensions but poor agreement at the level of individual health states within respondents. Joint models that assume agreement between DCE and TTO have been used to develop national value sets for the EQ-5D-5L. This work indicates that when combining data from both techniques, methods requiring association but not agreement are needed.

Background. There is growing interest in multicriteria decision analysis (MCDA) for shared decision making (SDM). A distinguishing feature is that a preferred treatment should extend years of life and/or improve health-related quality of life (HRQL). Additive MCDA models are inadequate for the task. A plethora of MCDA models exist, each claiming that it can correctly solve real-world problems. However, most were developed in nonhealth fields and rely on additive models. This makes the problem of choosing an MCDA model as an aid for SDM a challenging and urgent one. Methods. A published 2017 MCDA of a hypothetical prostate cancer patient is used as a case in point of how not to do and how to do MCDA for SDM. We critically review it and analyze it using several additive linear MCDA models with years of life and HRQL as attributes and the linear quality-adjusted life-year (QALY) model. The following simple reasonableness test is presented for applicability of a method as an aid for SDM: Can a treatment that causes premature death trump a treatment that causes acceptable adverse effects? Results. Additive MCDA models and the linear QALY recommend significantly different alternatives. Additive MCDA models fail the proposed reasonableness test; the linear QALY model passes. Conclusions. MCDA possesses a strong craft element in addition to its technical aspects. MCDA practitioners and clinicians need to understand model limitations to choose models appropriate to the context. Additive MCDA models are inadequate for life-critical SDM. We advocate QALY models with additional research for increased realism as a tool for SDM.

Background In discrete-choice experiments (DCEs), choice alternatives are described by attributes. The importance of each attribute can be quantified by analyzing respondents’ choices. Estimates are valid only if alternatives are defined comprehensively, but choice tasks can become too difficult for respondents if too many attributes are included. Several solutions for this dilemma have been proposed, but these have practical or theoretical drawbacks and cannot be applied in all settings. The objective of the current article is to demonstrate an alternative solution, the fold-in, fold-out approach (FiFo). We use a motivating example, the ABC Index for burden of disease in chronic obstructive pulmonary disease (COPD). Methods Under FiFo, all attributes are part of all choice sets, but they are grouped into domains. These are either folded in (all attributes have the same level) or folded out (levels may differ). FiFo was applied to the valuation of the ABC Index, which included 15 attributes. The data were analyzed in Bayesian mixed logit regression, with additional parameters to account for increased complexity in folded-out questionnaires and potential differences in weight due to the folding status of domains. As a comparison, a model without the additional parameters was estimated. Results Folding out domains led to increased choice complexity for respondents. It also gave domains more weight than when it was folded in. The more complex regression model had a better fit to the data than the simpler model. Not accounting for choice complexity in the models resulted in a substantially different ABC Index. Conclusion Using a combination of folded-in and folded-out attributes is a feasible approach for conducting DCEs with many attributes.

Objective. This study examined whether duration of treatment effect should be considered in a benefit-risk assessment using a case study of osteoarthritis medications. Study Design and Setting. A discrete choice experiment was completed by 300 residents of the United Kingdom with hip and/or knee osteoarthritis. In 16 choice tasks, participants selected their preferred option from 2 medications. Medications were described in terms of effect on pain, stiffness, and function; duration of treatment effect; and risk of heart attack and stomach ulcer bleeding. The analysis used mixed-effects logistic regression. Results. Pain, disease severity, and duration of treatment effect had the greatest influence on medication preferences, whereas stiffness did not significantly affect medication choice. Participants were willing to accept an increase in the risk of heart attack of 2.6% (95% confidence interval: 2.0% to 3.2%) to increase the duration of treatment effect from 1 month to 12 months. Reducing pain from moderate to mild was valued the same as increasing duration of effect from 1 month to 3 months; both were seen as equivalent to an absolute reduction of 1.2% in the risk of heart attack in the next year. Subgroup analysis suggested disease severity influenced patient preferences. Conclusions. Along with treatment benefits and risks, the results suggest that duration of treatment effect is an important factor in the medication choices of people with osteoarthritis. This could have implications for the design and interpretation of clinical trials, for example, incorporating longer-term surveillance of trial participants and accounting for duration of treatment effect in risk-benefit assessments. Future research is needed to assess whether these findings are generalizable to other samples, disease areas, and levels of duration of effect.

Voluntary medical male circumcision (VMMC) is effective in reducing the risk of human immunodeficiency virus (HIV). However, countries like Tanzania have high HIV prevalence but low uptake of VMMC. We conducted a discrete-choice experiment to evaluate the preferences for VMMC service attributes in a random sample of 325 men aged 18 years or older from the general population in 2 Tanzanian districts, Njombe and Tabora. We examined the preference for financial incentives in the form of a lottery ticket or receiving a guaranteed transport voucher for attendance at a VMMC service. We created a random-parameters logit model to account for individual preference heterogeneity and a latent class analysis model for identifying groups of men with similar preferences to test the hypothesis that men who reported sexually risky behaviors (i.e., multiple partners and any condomless sex in the past 12 months) may have a preference for participation in a lottery-based incentive. Most men preferred a transport voucher (84%) over a lottery ticket. We also found that offering a lottery-based financial incentive may not differentially attract those with greater sexual risk. Our study highlights the importance of gathering local data to understand preference heterogeneity, particularly regarding assumptions around risk behaviors.