O slideshow foi denunciado.
Utilizamos seu perfil e dados de atividades no LinkedIn para personalizar e exibir anúncios mais relevantes. Altere suas preferências de anúncios quando desejar.

Modelling causal pathways in health services part 1, Prof Richard Lilford

487 visualizações

Publicada em

The first of a two-part talk from Richard Lilford and Sam Watson on modelling causal pathways in health services for the CLAHRC West Midlands Scientific Advisory Group meeting, 9th June 2015, Birmingham, UK

Publicada em: Saúde e medicina
  • Entre para ver os comentários

  • Seja a primeira pessoa a gostar disto

Modelling causal pathways in health services part 1, Prof Richard Lilford

  1. 1. Modelling causal pathways in health services S. Watson and R. Lilford University of Warwick CLAHRC WM Scientific Advisory Group – June 2015
  2. 2. What is the problem? Causal effects of generic service interventions Multiple data of different types To inform decision models
  3. 3. Brown et al. Qual Saf Health Care. 2008;17:178-81. Brown & Lilford. BMJ. 2008;337:a2764. Policy Targeted service process Clinical process Patient Outcome Generic service process Classifying Health Interventions
  4. 4. Generic Processes Death Adverse events Patient satisfaction Brown et al. Qual Saf Health Care. 2008;17:178-81. Lilford et al. BMJ. 2010;341:c4413. Targeted Service Processes Clinical Processes QoL End-Points
  5. 5. Generic intervention Mediating variable Errors AEs QoL Δ1 Δ2 Δ3 +Δ = +Qualitative +
  6. 6. How can we make use of all the observations in a multi-level, multi-method study? Bayesian Modelling Lilford & Braunholtz. BMJ. 1996; 313: 603-7. Lilford, et al. BMJ. 2010; 341: c4413. Yao et al. BMJ Qual Saf. 2012; 21: i29-38. Hemming et al. PLoS One. 2012; 7(6): e38306. Lilford et al. BMC Health Serv Res. 2014; 14: 314.
  7. 7. Method 1: Mental Integration Alone Systematic review Theoretical knowledge Multi-level / multi- method observation Bias
  8. 8. Bayesian elicitation for intervention to reduce adverse events after discharge from hospital Relative risk reduction preventable adverse events – priors from 24 experts Pooled ‘prior’ for risk reduction of adverse events Yao et al. BMJ Qual Saf 2012; 21: i29-38. Hemming et al. PLoS ONE. 2012; 7(6):e38306.
  9. 9. Generic intervention Mediating variable Errors AEs Δ1 Δ2 ++ + Method 2: Bayesian Causal Network Analysis
  10. 10. Generic intervention Mediating variable Errors AEs Method 3:Intermediate methods Qualitative + ++
  11. 11. Factor Bias Start with meta- regression data Method 1 Method 2 Update mathematically (Turner & Spiegelhalter) Elicit distribution for bias Update mathematically

×