Limits...
Modeling test and treatment strategies for presymptomatic Alzheimer disease.

Burke JF, Langa KM, Hayward RA, Albin RL - PLoS ONE (2014)

Bottom Line: Net population benefit was estimated in aggregated QALYs.In the base-case scenario, treatment effects were uniformly positive, and net benefits increased with increasing age at screening.Highly efficacious presymptomatic screen and treat strategies for AD are likely to produce substantial aggregate population benefits that are likely greater than the benefits of aspirin in primary prevention of moderate risk cardiovascular disease (28 QALYS per 1000 patients treated), even in the context of an imperfect treatment delivery environment.

View Article: PubMed Central - PubMed

Affiliation: Dept. of Neurology, University of Michigan, Ann Arbor, Michigan, United States of America; Robert Wood Johnson Clinical Scholars Program, University of Michigan, Ann Arbor, Michigan, United States of America; Center for Clinical Management Research, VAAAHS, Ann Arbor, Michigan, United States of America.

ABSTRACT

Objectives: In this study, we developed a model of presymptomatic treatment of Alzheimer disease (AD) after a screening diagnostic evaluation and explored the circumstances required for an AD prevention treatment to produce aggregate net population benefit.

Methods: Monte Carlo simulation methods were used to estimate outcomes in a simulated population derived from data on AD incidence and mortality. A wide variety of treatment parameters were explored. Net population benefit was estimated in aggregated QALYs. Sensitivity analyses were performed by individually varying the primary parameters.

Findings: In the base-case scenario, treatment effects were uniformly positive, and net benefits increased with increasing age at screening. A highly efficacious treatment (i.e. relative risk 0.6) modeled in the base-case is estimated to save 20 QALYs per 1000 patients screened and 221 QALYs per 1000 patients treated.

Conclusions: Highly efficacious presymptomatic screen and treat strategies for AD are likely to produce substantial aggregate population benefits that are likely greater than the benefits of aspirin in primary prevention of moderate risk cardiovascular disease (28 QALYS per 1000 patients treated), even in the context of an imperfect treatment delivery environment.

Show MeSH

Related in: MedlinePlus

State change (Markov) model.How patients were selected for either the treatment or the untreated state change models and how state changes were assigned annually. Severity and living at home were hierarchically structured – once an individual arrived at the lowest level of the hierarchy (severe AD, living in a nursing home) an individual stayed in that state in subsequent years.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4256252&req=5

pone-0114339-g001: State change (Markov) model.How patients were selected for either the treatment or the untreated state change models and how state changes were assigned annually. Severity and living at home were hierarchically structured – once an individual arrived at the lowest level of the hierarchy (severe AD, living in a nursing home) an individual stayed in that state in subsequent years.

Mentions: Our approach assumes a single pathway to AD that is detectable by the screening strategy. This assumption is consistent with the prevailing amyloid cascade hypothesis that suggests a relatively stereotyped sequence of changes secondary to primary abnormalities in APP metabolism. [3] We designed a Monte Carlo simulation framework that allowed us to estimate the impacts of varying important parameters: age of treatment initiation, screening tool accuracy, overall treatment efficacy, variation in treatment efficacy over time, magnitude and probability of treatment-related harm, and probability of treatment discontinuation. Within this framework, we focused on the trade-offs of the magnitudes of treatment effect and treatment-related harm to estimate the parameters needed to achieve aggregate net population benefit. All simulation parameters are summarized in Table 1 and the structure of the underlying Markov model is outlined in Figure 1. All analyses were performed and all simulation code was written in Stata (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP.). Stata do-files will be made available upon request (James Burke; jamesbur@umich.edu).


Modeling test and treatment strategies for presymptomatic Alzheimer disease.

Burke JF, Langa KM, Hayward RA, Albin RL - PLoS ONE (2014)

State change (Markov) model.How patients were selected for either the treatment or the untreated state change models and how state changes were assigned annually. Severity and living at home were hierarchically structured – once an individual arrived at the lowest level of the hierarchy (severe AD, living in a nursing home) an individual stayed in that state in subsequent years.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4256252&req=5

pone-0114339-g001: State change (Markov) model.How patients were selected for either the treatment or the untreated state change models and how state changes were assigned annually. Severity and living at home were hierarchically structured – once an individual arrived at the lowest level of the hierarchy (severe AD, living in a nursing home) an individual stayed in that state in subsequent years.
Mentions: Our approach assumes a single pathway to AD that is detectable by the screening strategy. This assumption is consistent with the prevailing amyloid cascade hypothesis that suggests a relatively stereotyped sequence of changes secondary to primary abnormalities in APP metabolism. [3] We designed a Monte Carlo simulation framework that allowed us to estimate the impacts of varying important parameters: age of treatment initiation, screening tool accuracy, overall treatment efficacy, variation in treatment efficacy over time, magnitude and probability of treatment-related harm, and probability of treatment discontinuation. Within this framework, we focused on the trade-offs of the magnitudes of treatment effect and treatment-related harm to estimate the parameters needed to achieve aggregate net population benefit. All simulation parameters are summarized in Table 1 and the structure of the underlying Markov model is outlined in Figure 1. All analyses were performed and all simulation code was written in Stata (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP.). Stata do-files will be made available upon request (James Burke; jamesbur@umich.edu).

Bottom Line: Net population benefit was estimated in aggregated QALYs.In the base-case scenario, treatment effects were uniformly positive, and net benefits increased with increasing age at screening.Highly efficacious presymptomatic screen and treat strategies for AD are likely to produce substantial aggregate population benefits that are likely greater than the benefits of aspirin in primary prevention of moderate risk cardiovascular disease (28 QALYS per 1000 patients treated), even in the context of an imperfect treatment delivery environment.

View Article: PubMed Central - PubMed

Affiliation: Dept. of Neurology, University of Michigan, Ann Arbor, Michigan, United States of America; Robert Wood Johnson Clinical Scholars Program, University of Michigan, Ann Arbor, Michigan, United States of America; Center for Clinical Management Research, VAAAHS, Ann Arbor, Michigan, United States of America.

ABSTRACT

Objectives: In this study, we developed a model of presymptomatic treatment of Alzheimer disease (AD) after a screening diagnostic evaluation and explored the circumstances required for an AD prevention treatment to produce aggregate net population benefit.

Methods: Monte Carlo simulation methods were used to estimate outcomes in a simulated population derived from data on AD incidence and mortality. A wide variety of treatment parameters were explored. Net population benefit was estimated in aggregated QALYs. Sensitivity analyses were performed by individually varying the primary parameters.

Findings: In the base-case scenario, treatment effects were uniformly positive, and net benefits increased with increasing age at screening. A highly efficacious treatment (i.e. relative risk 0.6) modeled in the base-case is estimated to save 20 QALYs per 1000 patients screened and 221 QALYs per 1000 patients treated.

Conclusions: Highly efficacious presymptomatic screen and treat strategies for AD are likely to produce substantial aggregate population benefits that are likely greater than the benefits of aspirin in primary prevention of moderate risk cardiovascular disease (28 QALYS per 1000 patients treated), even in the context of an imperfect treatment delivery environment.

Show MeSH
Related in: MedlinePlus