A Blueprint for Multi-use Disease Modeling in Health Economics: Results from Two Expert-Panel Consultations

BACKGROUND

The current use of health economic decision models in HTA is mostly confined to single use cases, which may be inefficient and result in little consistency over different treatment comparisons, and consequently inconsistent health policy decisions, for the same disorder. Multi-use disease models (MUDMs) (other terms: generic models, whole disease models, disease models) may offer a solution. However, much is uncertain about their definition and application. The current research aimed to develop a blueprint for the application of MUDMs.

METHODS

We elicited expert opinion using a two-round modified Delphi process. The panel consisted of experts and stakeholders in health economic modelling from various professional backgrounds. The first questionnaire concerned definition, terminology, potential applications, issues and recommendations for MUDMs and was based on an exploratory scoping review. In the second round, the panel members were asked to reconsider their input, based on feedback regarding first-round results, and to score issues and recommendations for priority. Finally, adding input from external advisors and policy makers in a structured way, an overview of issues and challenges was developed during two team consensus meetings.

RESULTS

In total, 54 respondents contributed to the panel results. The term 'multi-use disease models' was proposed and agreed upon, and a definition was provided. The panel prioritized 10 potential applications (with comparing alternative policies and supporting resource allocation decisions as the top 2), while 20 issues (with model transparency and stakeholders' roles as the top 2) were identified as challenges. Opinions on potential features concerning operationalization of multi-use models were given, with 11 of these subsequently receiving high priority scores (regular updates and revalidation after updates were the top 2).

CONCLUSIONS

MUDMs would improve on current decision support regarding cost-effectiveness information. Given feasibility challenges, this would be most relevant for diseases with multiple treatments, large burden of disease and requiring more complex models. The current overview offers policy makers a starting point to organize the development, use, and maintenance of MUDMs and to support choices concerning which diseases and policy decisions they will be helpful for.

Cost-Effectiveness of SGLT2 Inhibitors in a Real-World Population: A MICADO Model-Based Analysis Using Routine Data from a GP Registry

OBJECTIVE

Sodium-glucose cotransporter 2 inhibitors (SGLT2i) have been shown to reduce the risk of cardiovascular complications, which largely drive diabetes' health and economic burdens. Trial results indicated that SGLT2i are cost effective. However, these findings may not be generalizable to the real-world target population. This study aims to evaluate the cost effectiveness of SGLT2i in a routine care type 2 diabetes population that meets Dutch reimbursement criteria using the MICADO model.

METHODS

Individuals from the Hoorn Diabetes Care System cohort (N = 15,392) were filtered to satisfy trial inclusion criteria (including EMPA-REG, CANVAS, and DECLARE-TIMI58) or satisfy the current Dutch reimbursement criteria for SGLT2i. We validated a health economic model (MICADO) by comparing simulated and observed outcomes regarding the relative risks of events in the intervention and comparator arm from three trials, and used the validated model to evaluate the long-term health outcomes using the filtered cohorts' baseline characteristics and treatment effects from trials and a review of observational studies. The incremental cost-effectiveness ratio (ICER) of SGLT2i, compared with care-as-usual, was assessed from a third-party payer perspective, measured in euros (2021 price level), using a discount rate of 4% for costs and 1.5% for effects.

RESULTS

From Dutch individuals with diabetes in routine care, 15.8% qualify for the current Dutch reimbursement criteria for SGLT2i. Their characteristics were significantly different (lower HbA1c, higher age, and generally more preexisting complications) than trial populations. After validating the MICADO model, we found that lifetime ICERs of SGLT2i, when compared with usual care, were favorable (< €20,000/QALY) for all filtered cohorts, resulting in an ICER of €5440/QALY using trial-based treatment effect estimates in reimbursed population. Several pragmatic scenarios were tested, the ICERs remained favorable.

CONCLUSIONS

Although the Dutch reimbursement indications led to a target group that deviates from trial populations, SGLT2i are likely to be cost effective when compared with usual care.

A congener-specific modelling approach for the transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans and dioxin-like polychlorinated biphenyls from feed to eggs of laying hens

Calibration of a kinetic model for the transfer of PCDD/Fs and dl-PCBs from feed to the hen's body and eggs was thus far restricted to the total TEQ concentration, i.e. the summed concentrations of PCDD/Fs and dl-PCBs expressed in terms of equivalents of 2,3,7,8-TCDD. However, this approach may lead to over- or underestimation of the transfer if the mixture contains congeners with kinetic characteristics which differ considerably from those used in such a model. This paper extends a previous transfer model of PCDD/Fs and dl-PCBs from feed to egg yolk fat and abdominal fat of high production laying hens, based on the total TEQ approach, to the level of individual congeners. Both modelling approaches are compared and the new approach is presented as a webtool application. This congener-specific approach enabled the calibration of 25 of the 29 relevant PCDD/F and dl-PCB congeners with respect to their individual transfer characteristics to body fat and egg yolk fat and their clearance from the body. Limitations of the available experimental data prevented the calibration of 1,2,3,4,6,7,8-HpCDD, OCDD, OCDF and PCB 123. The fraction transferred to egg yolk fat after long-term daily intake of contaminated feed was found to be at least 0.78 for 2,3,7,8-TCDD, 0.75 for PeCDD, 0.42-0.61 for HxCDDs, 0.70 for 2,3,7,8-TCDF, 0.71 for PeCDF, 0.54-0.60 for HxCDFs, 0.18-0.24 for HpCDFs and 0.89-1.00 for dl-PCBs. Various experimental and feed incident mixtures were used to compare the total TEQ- model with the congener-specific approach. An overestimation of the transfer by the total TEQ method was shown in particular for mixtures with a substantial contribution of hexa-, hepta- and octa-PCDD/Fs to the total TEQ level.

Integrating in vitro chemical transplacental passage into a generic PBK model: A QIVIVE approach

With the increasing application of cell culture models as primary tools for predicting chemical safety, the quantitative extrapolation of the effective dose from in vitro to in vivo (QIVIVE) is of increasing importance. For developmental toxicity this requires scaling the in vitro observed dose-response characteristics to in vivo fetal exposure, while integrating maternal in vivo kinetics during pregnancy, in particular transplacental transfer. Here the transfer of substances across the placental barrier, has been studied using the in vitro BeWo cell assay and six embryotoxic compounds of different kinetic complexity. The BeWo assay results were incorporated in an existing generic Physiologically Based Kinetic (PBK) model which for this purpose was extended with rat pregnancy. Finally, as a "proof of principle", the BeWo PBK model was used to perform a QIVIVE based on developmental toxicity as observed in various different in vitro toxicity assays. The BeWo results illustrated different transport profiles of the chemicals across the BeWo monolayer, allocating the substances into two distinct groups: the 'quickly-transported' and the 'slowly-transported'. BeWo PBK exposure simulations during gestation were compared to experimentally measured maternal blood and fetal concentrations and a reverse dosimetry approach was applied to translate in vitro observed embryotoxicity into equivalent in vivo dose-response curves. This approach allowed for a direct comparison of the in vitro dose-response characteristics as observed in the Whole Embryo Culture (WEC), and the Embryonic Stem Cell test (cardiac:ESTc and neural:ESTn) with in vivo rat developmental toxicity data. Overall, the in vitro to in vivo comparisons suggest a promising future for the application of such QIVIVE methodologies for screening and prioritization purposes of developmental toxicants. Nevertheless, the clear need for further improvements is acknowledged for a wider application of the approach in chemical safety assessment.

Do background levels of the pesticide pirimiphosmethyl in plant-based aquafeeds affect food safety of farmed Atlantic salmon?

The substitution of fish oil and fishmeal with plant-based ingredients in commercial aquafeeds for Atlantic salmon, may introduce novel contaminants that have not previously been associated with farmed fish. The organophosphate pesticide pirimiphos-methyl (PM) is one of the novel contaminants that is most prevalent in commercial salmon feed. In this study, the feed-to-fillet transfer of dietary PM and its main metabolites was investigated in Atlantic salmon fillet. Based on the experimental determined PM and metabolite uptake, metabolisation, and elimination kinetics, a physiologically based toxicokinetic (PBTK) compartmental model was developed. Fish fed PM had a relatively low (~4%) PM retention and two main metabolites (2-DAMP and Desethyl-PM) were identified in liver, muscle, kidney and bile. The absence of more metabolised forms of 2-DAMP and Desethyl-PM in Atlantic salmon indicates different metabolism in cold-water fish compared to previous studies on ruminants. The model was used to simulate the long term (>1.5 years) feed-to-fillet transfer of PM + metabolite in Atlantic salmon under realistic farming conditions including seasonal fluctuations in feed intake, growth, and fat deposition in muscle tissue. The model predictions show that with the constant presence of the highest observed PM concentration in commercial salmon feed, fillet PM+ metabolite levels were approximately 5 nmol kg^-1, with highest levels for the metabolite 2-DAMP. No EU maximum residue levels (MRL) for PM and its main metabolites exist in seafood to date, but the predicted levels were lower than the MRL for PM in swine of 32.7 nmol kg^-1.

Biomonitoring of Deoxynivalenol and Deoxynivalenol-3-glucoside in Human Volunteers: Renal Excretion Profiles

Biomarkers for the determination of the dietary exposure to deoxynivalenol (DON) have been proposed in the past but so far no quantification of their use in humans has been carried out. Following a human intervention study with two mycotoxins, namely DON and deoxynivalenol-3-glucoside (DON3G), the renal excretion of these compounds, including their phase II metabolites, was analysed. The purpose was to develop biokinetic models that can be used to determine: (1) the preferred (set of) urinary biomarker(s), (2) the preferred urinary collection period, and (3) a method to estimate the dietary exposure to these mycotoxins. Twenty adult volunteers were restricted in consuming cereals and cereal-based foods for 4 days. At day 3, a single dose of 1 µg/kg body weight of DON or DON3G was orally administered to 16 volunteers; 4 volunteers served as control. All individual urine discharges were collected during 24 h after administration. The metabolism and renal excretion could be described by a biokinetic model using three physiological compartments (gastrointestinal tract, liver, and kidneys). Kinetic analysis revealed a complete recovery of the renal excretion of total DON (mainly DON and its glucuronides) within 24 h after administration of DON or DON3G. The so-called 'reverse dosimetry' factor was used to determine the preferred (set of) biomarker(s) and to estimate the dietary intake of the parent compounds in the future. The fact that DON3G was absorbed and mainly excreted as DON and its glucuronides confirms that DON3G (as well as other modified forms) should be taken into account in the exposure and risk assessment of this group of mycotoxins.

Modelling of the feed-to-fillet transfer of ethoxyquin and one of its main metabolites, ethoxyquin dimer, to the fillet of farmed Atlantic salmon (Salmon salar L.)

Ethoxyquin (EQ) is an antioxidant supplemented to feed ingredients, mainly fish meal, which is currently under re-evaluation for use in the food production chain. EQ is partly metabolized into several metabolites of which the ethoxyquin dimer (EQDM) accumulates most in the farmed fish fillet. In this study, the feed-to-fillet transfer of dietary EQ and EQDM in Atlantic salmon fillet was investigated, and a physiologically based pharmacokinetic (PBPK-) two-compartmental model was developed, based on experimental determined EQ and EQDM uptake, metabolism, and elimination kinetics. The model was verified with an external data-set and used to simulate the long term (>1.5 years) EQ and EQDM feed-to fillet transfer in Atlantic salmon under realistic farming conditions such as the seasonal fluctuations in feed intake, growth, and fillet fat deposition. The model predictions showed that initial EQDM levels in juvenile fish are the driving factor in final levels found in food-producing animals, while for EQ the levels in feed, and seasonal variations were the driving factor for food EQ levels.

Cell selection as driving force in lung and colon carcinogenesis

Carcinogenesis is the result of mutations and subsequent clonal expansions of mutated, selectively advantageous cells. To investigate the relative contributions of mutation versus cell selection in tumorigenesis, we compared two mathematical models of carcinogenesis in two different cancer types: lung and colon. One approach is based on a population genetics model, the Wright-Fisher process, whereas the other approach is the two-stage clonal expansion model. We compared the dynamics of tumorigenesis predicted by the two models in terms of the time period until the first malignant cell appears, which will subsequently form a tumor. The mean waiting time to cancer has been calculated approximately for the evolutionary colon cancer model. Here, we derive new analytic approximations to the median waiting time for the two-stage lung cancer model and for a multistage approximation to the Wright-Fisher process. Both equations show that the waiting time to cancer is dominated by the selective advantage per mutation and the net clonal expansion rate, respectively, whereas the mutation rate has less effect. Our comparisons support the idea that the main driving force in lung and colon carcinogenesis is Darwinian cell selection.

The costs, effects and cost-effectiveness of counteracting overweight on a population level. A scientific base for policy targets for the Dutch national plan for action

OBJECTIVES

To gain insight in realistic policy targets for overweight at a population level and the accompanying costs. Therefore, the effect on overweight prevalence was estimated of large scale implementation of a community intervention (applied to 90% of general population) and an intensive lifestyle program (applied to 10% of overweight adults), and costs and cost-effectiveness were assessed.

METHODS

Costs and effects were based on two Dutch projects and verified by similar international projects. A markov-type simulation model estimated long-term health benefits, health care costs and cost-effectiveness.

RESULTS

Combined implementation of the interventions--at the above mentioned scale--reduces prevalence rates of overweight by approximately 3 percentage points and of physical inactivity by 2 percentage points after 5 years, at a cost of 7 euros per adult capita per year. The cost-effectiveness ratio of combined implementation amounts to euro 6000 per life-year gained and euro 5700 per QALY gained (including costs of unrelated diseases in life years gained). Sensitivity analyses showed that these ratios are quite robust.

CONCLUSIONS

A realistic policy target is a decrease in overweight prevalence of three percentage points, compared to a situation with no interventions. In reality, large scale implementation of the interventions may not counteract the expected upward trends in The Netherlands completely. Nonetheless, implementation of the interventions is cost-effective.

Modeling predicted that tobacco control policies targeted at lower educated will reduce the differences in life expectancy

BACKGROUND AND OBJECTIVE

To estimate the effects of reducing the prevalence of smoking in lower educated groups on educational differences in life expectancy.

METHODS

A dynamic Markov-type multistate transition model estimated the effects on life expectancy of two scenarios. A "maximum scenario" where educational differences in prevalence of smoking disappear immediately, and a "policy target-scenario" where difference in prevalence of smoking is halved over a 20-year period. The two scenarios were compared to a reference scenario, where smoking prevalences do not change. Five Dutch cohort studies, involving over 67,000 participants aged 20 to 90 years, provided relative mortality risks by educational level, and smoking habits were assessed using national data of more than 120,000 persons.

RESULTS

In the reference scenario, the difference in life expectancy at age 40 between highest and lowest educated groups was 5.1 years for men and 2.7 years for women. In the "maximum scenario" these differences were reduced to 3.6 years for men and 1.7 years for women (reduction approximately 30%), and in the "policy target-scenario" differences were 4.7 years for men and 2.4 years for women (reduction approximately 10%).

CONCLUSION

Theoretically, educational differences in life expectancy would be reduced by 30% at maximum, if variations in smoking prevalence were eliminated completely. In practice, tobacco control policies that are targeted at the lower educated may reduce the differences in life expectancy by approximately 10%.

Disease prevalence estimations based on contact registrations in general practice

This paper describes how to estimate the prevalence of chronic diseases in a population using data from contact registrations in general practice with a limited time length. Instead of using only total numbers of observed patients adjusted for the length of the observation period, we propose the use of (i) the time of the first contact of patients, (ii) the joint total numbers of patients and contacts, and (iii) the sets of patients in distinct time intervals, to generate prevalence rate estimates. The three new prevalence rate estimators have been developed assuming either a homogeneous or a parameterized heterogeneous patient population. Systematic and stochastic components of the estimators have been analysed by cross-validation for five chronic diseases using data from the Dutch 'Study on Chronic Conditions'. The results show that the first two estimators work well for diseases with a relatively structured visiting behaviour, such as hypertension and diabetes mellitus, assuming a time-constant contact rate and homogeneous patient population. For diseases such as ischaemic heart disease, chronic non-specific respiratory diseases and osteoarthritis, that do not satisfy these assumptions, the methods generally result in underestimations.

Limits of applicability for the deterministic approximation of the two-step clonal expansion model

Motivated by a hypothesis published recently, the limits of applicability of the deterministic approximation of the two-step clonal expansion model are investigated. The approximate hazard increases unlimited, while the exact hazard is approaching a constant value. The approximate solution becomes inapplicable for hazards in the order of that constant rate. When the initiation rate is much larger than the cell division rate, the survival rate is small when the approximation becomes inapplicable. The simplicity of the exact solutions suggests using them in all situations.

Community-based study of the incidence of gastrointestinal diseases in The Netherlands

The incidence of gastrointestinal diseases was studied in a community-based study in four regions of The Netherlands. Two grades of severity were distinguished--1: diarrhoea or vomiting and at least 2 additional symptoms within the period of 1 week, and 2: diarrhoea or vomiting and at least 2 additional symptoms occurring on the same day lasting at least 2 days within the period of 1 week. The incidence of gastrointestinal episodes was calculated to be 630 for grade 1 and 180 for grade 2 disease per 1000 person-years, after correction for age and sex. The incidence was higher for women than for men (relative risk 1.25) and lower for those in the 19-64-year-old age group when compared to those younger or older (relative risk 0.75 and 0.40, respectively). Independent of the degree of severity of the symptoms, about 20% of the patients had consulted a general practitioner, about half in person and half by telephone. It is concluded that community studies are essential to assess the real incidence of gastrointestinal diseases in the population.