arXiv subjects

Ralph Brinks

Publications and source records attributed to Ralph Brinks.

At least 19 recordsLinked to original sources

arXiv2025

Estimation of the incidence rate and mortality rate ratio for chronic conditions based on aggregated current status data

Recently, it has been shown that the transition rates of the illness-death model (IDM) for chronic conditions are related to the percentages of people in the states by a three-dimensional system of differential equations [Bri24]. The aim of this article is to introduce a method to estimate the age-specific incidence rate together with the mortality rate ratio from aggregated current status (ACS) data. By ACS data we mean counts of (non-necessarily different) people in the three states of the IDM at different points in time. ACS data stem from epidemiological studies where only current disease status and vital status data need to be collected without following-up people (as, for example, in cohort studies). As an application, we use the theory in a simulation study about diabetes in Germany with 600 study subjects at eleven repeated cross-sections each of which with 50% participation quote. Special focus is given to stochastic dependency of the sampled participants. We find a good agreement between the estimates and the input parameters used for the simulation.

stat.AP↗

arXiv2024

Application of the chemical master equation and its analytical solution to the illness-death model

The aim of this article is relating the chemical master equation (CME) to the illness-death model for chronic diseases. We show that a recently developed differential equation for the prevalence directly follows from the CME. As an application, we use the theory of the CME in a simulation study about diabetes in Germany from a previous publication. We find a good agreement between the theory and the simulations.

physics.bio-ph↗

arXiv2024

Maximum likelihood estimation for aggregate current status data: Simulation study using the illness-death model for chronic diseases with duration dependency

We use the illness-death model (IDM) for chronic conditions to derive a new analytical relation between the transition rates between the states of the IDM. The transition rates are the incidence rate (i) and the mortality rates of people without disease (m0) and with disease (m1). For the most generic case, the rates depend on age, calendar time and in case of m1 also on the duration of the disease. In this work, we show that the prevalence-odds can be expressed as a convolution-like product of the incidence rate and an exponentiated linear combination of i, m0 and m1. The analytical expression can be used as the basis for a maximum likelihood estimation (MLE) and associated large sample asymptotics. In a simulation study where a cross-sectional trial about a chronic condition is mimicked, we estimate the duration dependency of the mortality rate m1 based on aggregated current status data using the ML estimator. For this, the number of study participants and the number of diseased people in eleven age groups are considered. The ML estimator provides reasonable estimates for the parameters including their large sample confidence bounds.

stat.ME↗

arXiv2023

Illness-death model with renewal

The illness-death model for chronic conditions is combined with a renewal equation for the number of newborns taking into account possibly different fertility rates in the healthy and diseased parts of the population. The resulting boundary value problem consists of a system of partial differential equations with an integral boundary condition. As an application, the boundary value problem is applied to an example about type 2 diabetes.

q-bio.PE↗

arXiv2023

Stochastic differential equation for modelling health related quality of life

In this work we propose a stochastic differential equation (SDE) for modelling health related quality of life (HRQoL) over a lifespan. HRQoL is assumed to be bounded between 0 and 1, equivalent to death and perfect health, respectively. Drift and diffusion parameters of the SDE are chosen to mimic decreasing HRQoL over life and ensuring epidemiological meaningfulness. The Euler-Maruyama method is used to simulate trajectories of individuals in a population of n = 1000 people. Age of death of an individual is simulated as a stopping time with Weibull distribution conditioning the current value of HRQoL as time-varying covariate. The life expectancy and health adjusted life years are compared to the corresponding values for German women.

stat.ME↗

arXiv2023

Estimation of incidence from aggregated current status data

We use historical data about breathlessness in British coal miners and recent data about diabetes in Germany to illustrate a method for deriving the age-specific incidence from aggregated current status data, i.e. age-specific prevalence data. The method is centered on a differential equation, and special focus is put on maximum likelihood (ML) estimation of confidence intervals.

stat.ME↗

arXiv2023

Estimation of age-specific excess mortality of men and women with rheumatoid arthritis (RA) in Germany

A MCMC approach is used to estimate the age-specific mortality rate ratio for German men and women with RA. For constructing priors, we calculate a range of admissible values from prevalence and incidence data based on about 60 million people in Germany. Using these priors, MCMC mimics and compares estimated mortality to the findings of a recent register study from Denmark. It is estimated that the mortality rate ratio is highest in the young ages (4.0 and 3.5 for men and women aged 17.5 years, respectively) and declines towards higher ages (1.0 and 1.2 for men and women aged 92.5 years, respectively). The lengths of the credibility intervals decrease from younger towards older ages.

q-bio.OT↗

arXiv2022

Estimation of excess mortality in a chronic condition from current status data with disease duration: simulation study about need for long-term care

This article describes a method to estimate the mortality rate ratio R from current status data with duration in a chronic condition in case the general mortality of the overall population is known. Apart from the general mortality, the method requires four pieces of information from the study participants: age and time at the survey/interview, whether the chronic condition is present (current status) and if so, for how long the condition is present (duration). The method uses a differential equation that relates prevalence, incidence and mortality to estimate R of the people with the chronic condition compared to those without the condition. To demonstrate feasibility, a simulation based on the illness-death model (multi-state model) with transition rates motivated from long-term care is run. It is found that the method requires a large number of study participants (100000 or more) to estimate R with a reasonably low relative error. Despite the large sample size, the method can be useful in settings when cross-sectional information are easily available, e.g., in claims data, and national age-specific general mortality rates are accessible from vital statistics.

stat.ME↗

arXiv2022

Importance of diagnostic accuracy in big data: False-positive diagnoses of type 2 diabetes in health insurance claims data of 70 million Germans

Large data sets comprising diagnoses about chronic conditions are becoming increasingly available for research purposes. In Germany, it is planned that aggregated claims data including medical diagnoses from the statutory health insurance with roughly 70 million insurants will be published on a regular basis. Validity of the diagnoses in such big data sets can hardly be assessed. In case the data set comprises prevalence, incidence and mortality, it is possible to estimate the proportion of false positive diagnoses using mathematical relations from the illness-death model. We apply the method to age-specific aggregated claims data from 70 million Germans about type 2 diabetes in Germany stratified by sex and report the findings in terms of the ratio of false positive diagnoses of type 2 diabetes (FPR) in the data set. The age-specific FPR for men and women changes with age. In men, the FPR increases linearly from 1 to 3 per mil in the age 30 to 50. For ages between 50 to 80 years, FPR remains below 4 per mil. After 80 years of age, we have an increase to about 5 per mil. In women, we find a steep increase from age 30 to 60, the peak FPR is reached at about 12 per mil between 60 and 70 years of age. After age 70, the FPR of women drops tremendously. In all age-groups, the FPR is higher in women than in men. In terms of absolute numbers, we find that there are 217 thousand people with a false-positive diagnosis in the data set (95% confidence interval, CI: 204 to 229), the vast majority women (172 thousand, 95% CI: 162 to 180). Our work indicates that possible false positive (and negative) diagnoses should appropriately be dealt with in claims data, e.g., by inclusion of age- and sex-specific error terms in statistical models, to avoid potentially biased or wrong conclusions.

stat.ME↗

arXiv2020

Numerical considerations about the SIR epidemic model with infection age

We analyse the infection-age-dependent SIR model from a numerical point of view. First, we present an algorithm for calculating the solution the infection-age-structured SIR model without demography of the background host. Second, we examine how and under which conditions, the conventional SIR model (without infection-age) serves as a practical approximation to the infection-age SIR model. Special emphasis is given on the effective reproduction number.

q-bio.PE↗

arXiv2020

Estimation of the actual disease occurrence based on official case numbers during a COVID outbreak in Germany 2020

Since the beginning of March 2020, the cumulative numbers of cases of infection with the novel coronavirus SARS-CoV-2 in Germany have been reported on a daily basis. The reports originate from national laws, according to which positive test findings must be submitted to the Federal Health Authorities, the Robert Koch Institute, via the local health authorities. Since an enormous number of unreported cases can be expected, the question of how widespread the disease has been in the population cannot be answered based on these administrative reports. Using mathematical modeling, however, estimates can be made. These estimates indicate that the small numbers of diagnostic tests carried out at the beginning of the outbreak overlooked considerable parts of the infection. In order to cover the initial phase of future waves of the disease, wide-spread and comprehensive tests are recommended.

q-bio.PE↗

arXiv2019

Estimation of the excess mortality in chronic diseases from prevalence and incidence data

Aggregated health data such as claims data from health insurances become more and more available for research purposes. Estimates of excess mortality from prevalence and incidence of a chronic condition have only been possible for ages 50 years and older and have shown to be unstable in younger ages. The aim of this article is to explore the reasons why estimates of excess mortality for younger ages are prone to bias and what can be done to extend the age range to ages below 50 years.

q-bio.PE↗

arXiv2018

Morbidity, mortality and the illness-death model

In this article, we use the illness-death model to present a mathematical framework for studying the compression of morbidity (COM) hypothesis. It turns out that questions about COM are completely determined by the transition rates in the illness-death model and a closely related partial differential equation. By this, the COM hypothesis is analytically tractable. To demonstrate the usefulness of the mathematical framework, an example is given, which has been motivated by empirical findings from Germany.

q-bio.QM↗

arXiv2017

New ways of estimating excess mortality of chronic diseases: Insights from the illness-death model

Recently, we have shown that the age-specific prevalence of a disease can be related to the transition rates in the illness-death model via a partial differential equation (PDE). In case of a chronic disease, we show that the PDE can be used to estimate excess mortality from prevalence and incidence. Applicability of the new method is demonstrated in a simulation and claims data about diabetes in German men.

q-bio.PE↗

arXiv2016

On Keiding's Equation and its relation to differential equations about prevalence and incidence in chronic disease epidemiology

We study the relation between the age-specific prevalence, incidence and mortality in an illness-death model consisting of the three states Healthy, Ill, Dead. The dependency on three different time scales (age, calendar time, disease duration) is considered. It is shown that Keiding's equation published in 1991 is a generalisation of the solution of Brunet and Struchiner's partial differential equation from 1999. In a special case, we propose a particularly simple estimate of the incidence from prevalence data.

q-bio.PE↗

arXiv2016

How to assess case-finding in chronic diseases: Comparison of different indices

Recently, we have proposed a new illness-death model that comprises a state of undiagnosed chronic disease preceding the diagnosed disease. Based on this model, the question arises how case-finding can be assessed in the presence of mortality from all these states. We simulate two scenarios of different performance of case-finding and apply several indices to assess case-finding in both scenarios. One of the prevalence based indices leads to wrong conclusions. Some indices are partly insensitive to distinguish the quality of case-finding. The incidence based indices perform well. If possible, incidence based indices should be preferred.

q-bio.PE↗

arXiv2015

An identifiability problem in a state model for partly undetected chronic diseases

Recently, we proposed an state model (compartment model) to describe the progression of a chronic disease with an pre-clinical (undiagnosed) state before clinical diagnosis. It is an open question, if a sequence of cross-sectional studies with mortality follow-up is sufficient to estimate the true incidence rate of the disease, i.e. the incidence of the undiagnosed and diagnosed disease. In this note, we construct a counterexample and show that this cannot be achieved in general.

q-bio.PE↗

arXiv2014

State model for partly undetected non-communicable diseases (NCDs)

This article proposes an age-structured compartment model for irreversible diseases with a pre-clinical state of undiagnosed cases that precedes the diagnosis. The model is able to cope with mortality rates differing between the pre-clinical and the clinical state (differential mortality). Applicability is tested in a hypothetical disease with realistic incidence and mortality rates.

q-bio.PE↗