Mark Strong

Estimating Multiparameter Partial Expected Value of Perfect Information from a Probabilistic Sensitivity Analysis Sample A Nonparametric Regression Approach

The partial expected value of perfect information (EVPI) quantifies the expected benefit of learning the values of uncertain parameters in a decision model. Partial EVPI is commonly estimated via a 2-level Monte Carlo procedure in which parameters of interest are sampled in an outer loop, and then conditional on these, the remaining parameters are sampled in an inner loop. This is computationally demanding and may be difficult if correlation between input parameters results in conditional distributions that are hard to sample from. We describe a novel nonparametric regression-based method for estimating partial EVPI that requires only the probabilistic sensitivity analysis sample (i.e., the set of samples drawn from the joint distribution of the parameters and the corresponding net benefits). The method is applicable in a model of any complexity and with any specification of input parameter distribution. We describe the implementation of the method via 2 nonparametric regression modeling approaches, the Generalized Additive Model and the Gaussian process. We demonstrate in 2 case studies the superior efficiency of the regression method over the 2-level Monte Carlo method. R code is made available to implement the method.

The Prevalence of Juvenile Huntington's Disease: A Review of the Literature and Meta-Analysis.

Juvenile Huntington’s disease (JHD) is usually defined as Huntingtons disease with an onset ≤ 20 years. The proportion of JHD cases reported in studies of Huntington’s disease (HD) varies. A review of the literature found 62 studies that reported the proportion of JHD cases amongst all HD cases. The proportion of JHD cases in these studies ranged from 1% to 15%, and in a meta-analysis the pooled proportion of JHD cases was 4.92% (95% confidence interval of 4.07% to 5.84%). Limiting the analysis to the 25 studies which used multiple methods of ascertainment resulted in a similar pooled proportion of 5.32%, (95% confidence interval 4.18% to 6.60%). A small difference was observed when the meta-analysis was restricted to studies from countries defined by the World Bank as high income, that used multiple methods of ascertainment, and that were conducted since 1980 (4.81%, 95% confidence interval 3.31% to 6.58%, n=11). This contrasts with the pooled result from three post 1980 studies using multiple methods of ascertainment from South Africa and Venezuela, defined by the World Bank as upper middle income, where the estimated mean proportion was 9.95%, (95% confidence interval 6.37% to 14.22%). These results, which are expected to be more robust than those from a single study alone, may be helpful in estimating the proportion of JHD cases in a given population. Key Words: Juvenile Huntington’s disease, prevalence, epidemiology

A comparison of methods for calculating general practice level socioeconomic deprivation

BackgroundA measure of the socioeconomic deprivation experienced by the registered patient population of a general practice is of interest because it can be used to explore the association between deprivation and a wide range of other variables measured at practice level. If patient level geographical data are available a population weighted mean area-based deprivation score can be calculated for each practice. In the absence of these data, an area-based deprivation score linked to the practice postcode can be used as an estimate of the socioeconomic deprivation of the practice population. This study explores the correlation between Index of Multiple Deprivation 2004 (IMD) scores linked to general practice postcodes (main surgery address alone and main surgery plus any branch surgeries), practice population weighted mean IMD scores, and practice level mortality (aged 1 to 75 years, all causes) for 38 practices in Rotherham UK.ResultsPopulation weighted deprivation scores correlated with practice postcode based scores (main surgery only, Pearson r = 0.74, 95% CI 0.54 to 0.85; main plus branch surgeries, r = 0.79, 95% CI 0.63 to 0.89). All cause mortality aged 1 to 75 correlated with deprivation (main surgery postcode based measure, r = 0.50, 95% CI 0.22 to 0.71; main plus branch surgery based score, r = 0.55, 95% CI 0.28 to 0.74); population weighted measure, r = 0.66, 95% CI 0.43 to 0.81).ConclusionPractice postcode linked IMD scores provide a valid proxy for a population weighted measure in the absence of patient level data. However, by using them, the strength of association between mortality and deprivation may be underestimated.

An Efficient Method for Computing Single- Parameter Partial Expected Value of Perfect Information

The value of learning an uncertain input in a decision model can be quantified by its partial expected value of perfect information (EVPI). This is commonly estimated via a 2-level nested Monte Carlo procedure in which the parameter of interest is sampled in an outer loop, and then conditional on this sampled value, the remaining parameters are sampled in an inner loop. This 2-level method can be difficult to implement if the joint distribution of the inner-loop parameters conditional on the parameter of interest is not easy to sample from. We present a simple alternative 1-level method for calculating partial EVPI for a single parameter that avoids the need to sample directly from the potentially problematic conditional distributions. We derive the sampling distribution of our estimator and show in a case study that it is both statistically and computationally more efficient than the 2-level method.

Estimating the Expected Value of Sample Information Using the Probabilistic Sensitivity Analysis Sample A Fast, Nonparametric Regression-Based Method

Health economic decision-analytic models are used to estimate the expected net benefits of competing decision options. The true values of the input parameters of such models are rarely known with certainty, and it is often useful to quantify the value to the decision maker of reducing uncertainty through collecting new data. In the context of a particular decision problem, the value of a proposed research design can be quantified by its expected value of sample information (EVSI). EVSI is commonly estimated via a 2-level Monte Carlo procedure in which plausible data sets are generated in an outer loop, and then, conditional on these, the parameters of the decision model are updated via Bayes rule and sampled in an inner loop. At each iteration of the inner loop, the decision model is evaluated. This is computationally demanding and may be difficult if the posterior distribution of the model parameters conditional on sampled data is hard to sample from. We describe a fast nonparametric regression-based method for estimating per-patient EVSI that requires only the probabilistic sensitivity analysis sample (i.e., the set of samples drawn from the joint distribution of the parameters and the corresponding net benefits). The method avoids the need to sample from the posterior distributions of the parameters and avoids the need to rerun the model. The only requirement is that sample data sets can be generated. The method is applicable with a model of any complexity and with any specification of model parameter distribution. We demonstrate in a case study the superior efficiency of the regression method over the 2-level Monte Carlo method.

Cohort Profile: The Yorkshire Health Study

The Yorkshire Health Study is a longitudinal observational regional health study collecting health information on the residents from the Yorkshire and Humberside region in England. The second wave of data collection is currently under way. The study aims to inform National Health Service (NHS) and local authority health-related decision making in Yorkshire, with wider implications from findings as well. The first wave contains records for 27 806 individuals (2010-12), aged between 16 and 85, from one part of Yorkshire (South Yorkshire), with the second wave expanding data collection to the whole of the Yorkshire and Humberside region. Data were collected on current and long-standing health, health care usage and health-related behaviours, with a particular focus on weight and weight management. The majority of individuals have also given consent for record linkage with routine clinical data, allowing the linking to disease diagnosis, medication use and health care usage. The study encourages researchers to utilize the sample through the embedding of randomized controlled trials, other controlled trials and qualitative studies. To access the anonymized data or use the sample to recruit participants to studies, researchers should contact Clare Relton (c.relton@sheffield.ac.uk).

Bayesian Sensitivity Analysis of a Cardiac Cell Model Using a Gaussian Process Emulator.

Models of electrical activity in cardiac cells have become important research tools as they can provide a quantitative description of detailed and integrative physiology. However, cardiac cell models have many parameters, and how uncertainties in these parameters affect the model output is difficult to assess without undertaking large numbers of model runs. In this study we show that a surrogate statistical model of a cardiac cell model (the Luo-Rudy 1991 model) can be built using Gaussian process (GP) emulators. Using this approach we examined how eight outputs describing the action potential shape and action potential duration restitution depend on six inputs, which we selected to be the maximum conductances in the Luo-Rudy 1991 model. We found that the GP emulators could be fitted to a small number of model runs, and behaved as would be expected based on the underlying physiology that the model represents. We have shown that an emulator approach is a powerful tool for uncertainty and sensitivity analysis in cardiac cell models.

When Is a Model Good Enough? Deriving the Expected Value of Model Improvement via Specifying Internal Model Discrepancies ∗

A “law-driven” or “mechanistic” computer model is a representation of judgments about the functional relationship between one set of quantities (the model inputs) and another set of target quantities (the model outputs). We recognize that we can rarely define with certainty a “true” model for a particular problem. Building an “incorrect” model will result in an uncertain prediction error, which we denote “structural uncertainty.” Structural uncertainty can be quantified within a Bayesian framework via the specification of a series of internal discrepancy terms, each representing at a subfunction level within the model the difference between the subfunction output and the true value of the intermediate parameter implied by the subfunction. By using value of information analysis we can then determine the expected value of learning the discrepancy terms, which we loosely interpret as an upper bound on the “expected value of model improvement.” We illustrate the method using a case study model drawn from the ...

Healthcare providers’ views on the acceptability of financial incentives for breastfeeding: a qualitative study

BackgroundDespite a gradual increase in breastfeeding rates, overall in the UK there are wide variations, with a trend towards breastfeeding rates at 6–8 weeks remaining below 40% in less affluent areas. While financial incentives have been used with varying success to encourage positive health related behaviour change, there is little research on their use in encouraging breastfeeding. In this paper, we report on healthcare providers’ views around whether using financial incentives in areas with low breastfeeding rates would be acceptable in principle. This research was part of a larger project looking at the development and feasibility testing of a financial incentive scheme for breastfeeding in preparation for a cluster randomised controlled trial.MethodsFifty–three healthcare providers were interviewed about their views on financial incentives for breastfeeding. Participants were purposively sampled to include a wide range of experience and roles associated with supporting mothers with infant feeding. Semi-structured individual and group interviews were conducted. Data were analysed thematically drawing on the principles of Framework Analysis.ResultsThe key theme emerging from healthcare providers’ views on the acceptability of financial incentives for breastfeeding was their possible impact on ‘facilitating or impeding relationships’. Within this theme several additional aspects were discussed: the mother’s relationship with her healthcare provider and services, with her baby and her family, and with the wider community. In addition, a key priority for healthcare providers was that an incentive scheme should not impact negatively on their professional integrity and responsibility towards women.ConclusionHealthcare providers believe that financial incentives could have both positive and negative impacts on a mother’s relationship with her family, baby and healthcare provider. When designing a financial incentive scheme we must take care to minimise the potential negative impacts that have been highlighted, while at the same time recognising the potential positive impacts for women in areas where breastfeeding rates are low.

The UK Quality and Outcomes Framework pay-for-performance scheme and spirometry: rewarding quality or just quantity? A cross-sectional study in Rotherham, UK

BackgroundAccurate spirometry is important in the management of COPD. The UK Quality and Outcomes Framework pay-for-performance scheme for general practitioners includes spirometry related indicators within its COPD domain. It is not known whether high achievement against QOF spirometry indicators is associated with spirometry to BTS standards.MethodsData were obtained from the records of 3,217 patients randomly sampled from 5,649 patients with COPD in 38 general practices in Rotherham, UK. Severity of airflow obstruction was categorised by FEV1 (% predicted) according to NICE guidelines. This was compared with clinician recorded COPD severity. The proportion of patients whose spirometry met BTS standards was calculated in each practice using a random sub-sample of 761 patients. The Spearman rank correlation between practice level QOF spirometry achievement and performance against BTS spirometry standards was calculated.ResultsSpirometry as assessed by clinical records was to BTS standards in 31% of cases (range at practice level 0% to 74%). The categorisation of airflow obstruction according to the most recent spirometry results did not agree well with the clinical categorisation of COPD recorded in the notes (Cohens kappa = 0.34, 0.30 – 0.38). 12% of patients on COPD registers had FEV1 (% predicted) results recorded that did not support the diagnosis of COPD. There was no association between quality, as measured by adherence to BTS spirometry standards, and either QOF COPD9 achievement (Spearmans rho = -0.11), or QOF COPD10 achievement (rho = 0.01).ConclusionThe UK Quality and Outcomes Framework currently assesses the quantity, but not the quality of spirometry.