Indirect estimation of a discrete-state discrete-time model using secondary data analysis of regression data.
TL;DR: This paper presents an approach that allows the use of published regression data in a multi- state model when the published study may have ignored intermediary states in the multi-state model, called the Lemonade Method.
Abstract: Multi-state models of chronic disease are becoming increasingly important in medical research to describe the progression of complicated diseases. However, studies seldom observe health outcomes over long time periods. Therefore, current clinical research focuses on the secondary data analysis of the published literature to estimate a single transition probability within the entire model. Unfortunately, there are many difficulties when using secondary data, especially since the states and transitions of published studies may not be consistent with the proposed multi-state model. Early approaches to reconciling published studies with the theoretical framework of a multi-state model have been limited to data available as cumulative counts of progression. This paper presents an approach that allows the use of published regression data in a multi-state model when the published study may have ignored intermediary states in the multi-state model. Colloquially, we call this approach the Lemonade Method since when study data give you lemons, make lemonade. The approach uses maximum likelihood estimation. An example is provided for the progression of heart disease in people with diabetes.
Citations
More filters
TL;DR: A taxonomy based on possible scenarios faced by the analyst when dealing with the available evidence is developed that can help modelers identify the most appropriate methods to use when synthesizing the available data for a given model parameter.
46 citations
TL;DR: A new computer tool designed for chronic disease modeling is described and the modeling capabilities of this tool were used to model the Michigan model for diabetes.
34 citations
TL;DR: A likelihood approach to correctly model the design of clinical studies under the conditions where 1) the theoretical model may include an instantaneous state of distinct interest to the researchers, and 2) the study design may be such that study data can not be used to estimate a single parameter in the theoreticalmodel of interest.
16 citations
Dissertation•
01 Jan 2012
TL;DR: In this paper, a taxonomy of the methodological and analytical issues in the use and synthesis of evidence for cost effectiveness modelling is presented, with guidance on appropriate synthesis methodologies to use and identifying areas where further methodological contributions are needed.
Abstract: Health care economic evaluations assess the costs and consequences of competing interventions, programmes or services. Such assessments use a decision model, with
parameters informed by available evidence. Evidence, however, is rarely derived from a single source, in which case researchers are expected to combine information on multiple sources. This thesis contributes to the methodological debate on the use of evidence, particularly, the use of individual level data (IPD), for cost effectiveness analysis.
This thesis defines a taxonomy which summarises the methodological and analytical issues in the use and synthesis of evidence for cost effectiveness modelling. For
alternative parameter types (e.g. relative effectiveness, costs) the taxonomy offers guidance on appropriate synthesis methodologies to use and identifies areas where
further methodological contributions are needed. The thesis also explores methods of synthesis of IPD and develops novel frameworks which allow both IPD and AD to be jointly modelled, specifically in estimating relative effectiveness. The use of IPD from studies is found desirable, particularly when the estimation of subgroup effects is of interest.
An applied decision model of the cost effectiveness of smoke alarm equipment in households with pre-school children is developed within this thesis. This application
offers a means to evaluate the impact of using IPD on the cost effectiveness outcomes, compared to the use of AD. The thesis examines the advantages of having access to IPD when quantifying decision uncertainty. Additionally, it discusses the use of IPD in estimating the value of further research. Specifically, a framework is used which allows considering population subgroups. It is argued that the use of IPD allows a more suitable characterisation of decision uncertainty, appropriately allowing for subgroup value of information analysis.
8 citations
27 Nov 2013
TL;DR: In this article, reference disease models predict progression of disease within given populations, utilizing publically available clinical data and risk equations, to give a birds-eye view of clinical trials by allowing multiple trials to be systematically compared simultaneously via parallel processing/High Performance Computing which allows competition among alternative equations/hypothesis combinations; cross validation; and, then ranks results according to fitness via a fitness engine.
Abstract: A method wherein reference disease models predict progression of disease within given populations, utilizing publically available clinical data and risk equations, to give a birds-eye view of clinical trials by allowing multiple trials to be systematically compared simultaneously via parallel processing/High Performance Computing which allows competition among alternative equations/hypothesis combinations; cross validation; and, then ranks results according to fitness via a fitness engine.
7 citations
References
More filters
TL;DR: It is suggested that diabetic patients without previous myocardial infarction have as high a risk of myocardia infarctions as nondiabetic patients with previous my Cardiac Arrest.
Abstract: Background Type 2 (non-insulin-dependent) diabetes is associated with a marked increase in the risk of coronary heart disease. It has been debated whether patients with diabetes who have not had myocardial infarctions should be treated as aggressively for cardiovascular risk factors as patients who have had myocardial infarctions. Methods To address this issue, we compared the seven-year incidence of myocardial infarction (fatal and nonfatal) among 1373 nondiabetic subjects with the incidence among 1059 diabetic subjects, all from a Finnish population-based study. Results The seven-year incidence rates of myocardial infarction in nondiabetic subjects with and without prior myocardial infarction at base line were 18.8 percent and 3.5 percent, respectively (P<0.001). The seven-year incidence rates of myocardial infarction in diabetic subjects with and without prior myocardial infarction at base line were 45.0 percent and 20.2 percent, respectively (P<0.001). The hazard ratio for death from coronary heart di...
6,359 citations
Journal Article•
4,506 citations
TL;DR: This tutorial on advanced statistical methods for meta-analysis can be seen as a sequel to the recent Tutorial in Biostatistics on meta- analysis by Normand, which focused on elementary methods.
Abstract: This tutorial on advanced statistical methods for meta-analysis can be seen as a sequel to the recent Tutorial in Biostatistics on meta-analysis by Normand, which focused on elementary methods. Within the framework of the general linear mixed model using approximate likelihood, we discuss methods to analyse univariate as well as bivariate treatment effects in meta-analyses as well as meta-regression methods. Several extensions of the models are discussed, like exact likelihood, non-normal mixtures and multiple endpoints. We end with a discussion about the use of Bayesian methods in meta-analysis. All methods are illustrated by a meta-analysis concerning the efficacy of BCG vaccine against tuberculosis. All analyses that use approximate likelihood can be carried out by standard software. We demonstrate how the models can be fitted using SAS Proc Mixed.
1,417 citations
"Indirect estimation of a discrete-s..." refers background in this paper
...likelihood function of unknown parameters k(k) is Normal(k̂(k),R(k)) [17, 18]....
[...]
TL;DR: The model is diabetes-specific and incorporates glycaemia, systolic blood pressure and lipid levels as risk factors, in addition to age, sex, ethnic group, smoking status and time since diagnosis of diabetes, which provides the estimates ofCHD risk required by current guidelines for the primary prevention of CHD in Type II diabetes.
Abstract: A definitive model for predicting absolute risk of coronary heart disease (CHD) in male and female people with Type II diabetes is not yet available. This paper provides an equation for estimating the risk of new CHD events in people with Type II diabetes, based on data from 4540 U.K. Prospective Diabetes Study male and female patients. Unlike previously published risk equations, the model is diabetes-specific and incorporates glycaemia, systolic blood pressure and lipid levels as risk factors, in addition to age, sex, ethnic group, smoking status and time since diagnosis of diabetes. All variables included in the final model were statistically significant (P<0.001, except smoking for which P=0.0013) in likelihood ratio testing. This model provides the estimates of CHD risk required by current guidelines for the primary prevention of CHD in Type II diabetes.
1,188 citations
"Indirect estimation of a discrete-s..." refers methods in this paper
...We desired to use the method in [12] to combine these data into a single model of diabetes progression; however, one important study, the UKPDS [11], provided a risk equation rather than the cumulative counts....
[...]
...For example, the well-known UKPDS group [4, 10, 11] has developed the UKPDS Risk Engine Model for coronary heart disease and published their risk equation [11]....
[...]
...The incident counts were estimated by using the expected value of the UKPDS risk engine formula for men and for women separately by using the mean substitution of averages defined in Table 1 in [11] and rounding the obtained number....
[...]
...0318 Haffner, 1998 [22] 3 B2: 0 to 2 4540 See UKPDS 10 Stevens, 2001 risk equation (UKPDS 56) [11] 4 C: 1 to 2 569 61 2 0....
[...]
...To conduct this least-squares fit, we used the marginal distributions for covariates as published in Table 1 of [11] to generate a random population of 4540 people representing the UKPDS sample....
[...]
TL;DR: Hospitalization for unstable angina or non–Q-wave myocardial infarction predicts a high 2-year morbidity and mortality; this is especially evident for patients with diabetes.
Abstract: Background—Although unstable coronary artery disease is the most common reason for admission to a coronary care unit, the long-term prognosis of patients with this diagnosis is unknown. This is particularly true for patients with diabetes mellitus, who are known to have a high morbidity and mortality after an acute myocardial infarction. Methods and Results—Prospectively collected data from 6 different countries in the Organization to Assess Strategies for Ischemic Syndromes (OASIS) registry were analyzed to determine the 2-year prognosis of diabetic and nondiabetic patients who were hospitalized with unstable angina or non–Q-wave myocardial infarction. Overall, 1718 of 8013 registry patients (21%) had diabetes. Diabetic patients had a higher rate of coronary bypass surgery than nondiabetic patients (23% versus 20%, P<0.001) but had similar rates of catheterization and angioplasty. Diabetes independently predicted mortality (relative risk [RR], 1.57; 95% CI, 1.38 to 1.81; P<0.001), as well as cardiovascul...
765 citations