Table 2 Models used for each time step of the proposed algorithm
Step in algorithm | Type of generalised linear model | Outcome | Model as specified in R-code |
|---|---|---|---|
1 | Logistic regression | Dead | Age + QCS(Age,100)*QCS(CCI,8) + QCS(Age,100)*QCS(DCI,14) + QCS(Age,100)*QCS(CCI,8)*QCS(DCI,14) |
3 | Logistic regression | Any.CCI.change | Age + DCI.fct + CCI.fct |
3 | Poisson regression | CCI.change | Age + DCI.fct + CCI.fct |
3 | Logistic regression | Any.CCI.change6 | Age + DCI.fct + CCI.fct.6p + CCI |
5 | Logistic regression | Any.DCI.change | Age + dci.fct + cci.fct + Age*cci.fct + Age*dci.fct |
5 | Logistic regression (restricted to men with DCI-change) | DCI.increase | Age + dci.fct + cci.fct + Age*cci.fct + Age*dci.fct |
6 | Gamma regression (restricted to men with DCI-increase) | DCI.change | Age + dci.fct + cci.fct + Age*cci.fct + Age*dci.fct |
7 | Gamma regression (restricted to men with DCI-decrease) | -DCI.change | Age + dci.fct + cci.fct + Age*cci.fct + Age*dci.fct |