+
Result files (first order: no variance)
++
+ - Estimated transition probabilities over 1 (stepm) months: biaspar/PIJ_biaspar.txt
+ - Estimated back transition probabilities over 1 (stepm) months: biaspar/PIJB_biaspar.txt
+ - Period (forward) prevalence in each health state: biaspar/PL_biaspar.txt
+ - Backward prevalence in each health state: biaspar/PLB_biaspar.txt
+ - (a) Life expectancies by health status at initial age, ei. (b) health expectancies by health status at initial age, eij . If one or more covariates are included, specific tables for each value of the covariate are output in sequences within the same file (estepm= 1 months): biaspar/E_biaspar.txt
+ - Prevalence projections by age and states: biaspar/F_biaspar.txt
+
- Graphs
+
-
+
- Logit model (yours is: logit(pij)=log(pij/pii)= aij+ bij age+) as a function of age: biaspar/PE_biaspar_1-1-1.svg
+- Pij or conditional probabilities to be observed in state j being in state i, 1 (stepm) months before: biaspar/PE_biaspar_1-2-1.svg
+- Iij or Conditional probabilities to be observed in state j being in state i 1 (stepm) months before but expressed in per year i.e. quasi incidences if stepm is small and probabilities too, incidence (rates) are the limit when h tends to zero of the ratio of the probability hPij divided by h: hPij/h : biaspar/PE_biaspar_1-3-1.svg
+- Survival functions in state 1. And probability to be observed in state 1 being in state (1 to 2) at different ages. biaspar/LIJ_biaspar_1-1-1.svg
+- Survival functions in state 2. And probability to be observed in state 2 being in state (1 to 2) at different ages. biaspar/LIJ_biaspar_2-1-1.svg
+- Survival functions in state 1 and in any other live state (total). And probability to be observed in various states (up to 2) being in state 1 at different ages. biaspar/LIJT_biaspar_1-1-1.svg
+- Survival functions in state 2 and in any other live state (total). And probability to be observed in various states (up to 2) being in state 2 at different ages. biaspar/LIJT_biaspar_2-1-1.svg
+- Convergence to period (stable) prevalence in state 1. Or probability for a person being in state (1 to 2) at different ages, to be in state 1 some years after. biaspar/P_biaspar_1-1-1.svg
+- Convergence to period (stable) prevalence in state 2. Or probability for a person being in state (1 to 2) at different ages, to be in state 2 some years after. biaspar/P_biaspar_2-1-1.svg
+- Projection of cross-sectional prevalence (estimated with cases observed from 1984.0 to 1988.4 and mobil_average=0), from year 1989.0 up to year 2000.0 tending to period (stable) forward prevalence in state 1. Or probability to be in state 1 being in an observed weighted state (from 1 to 2). biaspar/PROJ_biaspar_1-1-1.svg
+- Projection of cross-sectional prevalence (estimated with cases observed from 1984.0 to 1988.4 and mobil_average=0), from year 1989.0 up to year 2000.0 tending to period (stable) forward prevalence in state 2. Or probability to be in state 2 being in an observed weighted state (from 1 to 2). biaspar/PROJ_biaspar_2-1-1.svg
+- Life expectancy by health state (1) at initial age and its decomposition into health expectancies in each alive state (1 to 2) (or area under each survival functions): biaspar/EXP_biaspar_1-1-1.svg
+- Life expectancy by health state (2) at initial age and its decomposition into health expectancies in each alive state (1 to 2) (or area under each survival functions): biaspar/EXP_biaspar_2-1-1.svg
Result files (second order: variances)
+ - Parameter file with estimated parameters and covariance matrix: rbiaspar.imach- 95% confidence intervals and Wald tests of the estimated parameters are in the log file if optimization has been done (mle != 0).
But because parameters are usually highly correlated (a higher incidence of disability and a higher incidence of recovery can give very close observed transition) it might be very useful to look not only at linear confidence intervals estimated from the variances but at the covariance matrix. And instead of looking at the estimated coefficients (parameters) of the logistic regression, it might be more meaningful to visualize the covariance matrix of the one-step probabilities. See page 'Matrix of variance-covariance of one-step probabilities' below. + - Standard deviation of one-step probabilities: biaspar/PROB_biaspar.txt
+ - Variance-covariance of one-step probabilities: biaspar/PROBCOV_biaspar.txt
+ - Correlation matrix of one-step probabilities: biaspar/PROBCOR_biaspar.txt
+ - Variances and covariances of health expectancies by age and initial health status (cov(eij,ekl)(estepm= 1 months): biaspar/CVE_biaspar.txt
+
+ - Variances and covariances of health expectancies by age. Status (i) based health expectancies (in state j), eij are weighted by the forward (period) prevalences in each state i (if popbased=1, an additional computation is done using the cross-sectional prevalences, i.e population based) (estepm=1 months): biaspar/V_biaspar.txt
+ - Total life expectancy and total health expectancies to be spent in each health state e.j with their standard errors (if popbased=1, an additional computation is done using the cross-sectional prevalences, i.e population based) (estepm=1 months): biaspar/T_biaspar.txt
+ - Standard deviation of forward (period) prevalences: biaspar/VPL_biaspar.txt
+
- Graphs
+
- Observed (cross-sectional with mov_average=0) and period (incidence based) prevalence (with 95% confidence interval) in state (1): biaspar/V_biaspar_1-1-1.svg
+ 
+
- Observed (cross-sectional with mov_average=0) and period (incidence based) prevalence (with 95% confidence interval) in state (2): biaspar/V_biaspar_2-1-1.svg
+ 
+
- Total life expectancy by age and health expectancies in states (1) and (2). If popbased=1 the smooth (due to the model) true period expectancies (those weighted with period prevalences are also drawn in addition to the population based expectancies computed using observed and cahotic prevalences: biaspar/E_biaspar_1-1.svg
+
Computing and drawing one step probabilities with their confidence intervals
Matrix of variance-covariance of one-step probabilities (drawings)
this page is important in order to visualize confidence intervals and especially correlation between disability and recovery, or more generally, way in and way back. File biaspar-cov.htmComputing probabilities of dying over estepm months as a weighted average (i.e global mortality independent of initial healh state)
-STABLBASED_
+ +
File (multiple files are possible if covariates are present): biaspar/PRMORPREV-1-STABLBASED_biaspar.txt + +
Probability is computed over estepm=1 months.
+ +
Computing probabilities of dying over estepm months as a weighted average (i.e global mortality independent of initial healh state)
-POPULBASED-NOMOBIL_
+ +
File (multiple files are possible if covariates are present): biaspar/PRMORPREV-1-POPULBASED-NOMOBIL_biaspar.txt + +
Probability is computed over estepm=1 months.
+
Local time at start Thu May 20 15:42:13 2021 +
Local time at end Thu May 20 15:43:12 2021 +
+ \ No newline at end of file