Table 2.
Latent class analysis model fit per number of latent classes for intervention end points.
| No. of classes | Log likelihood | Parameter number | BICa |
| 1 | –1631 | 22 | 3382 |
| 2 | –1510 | 45 | 3265 |
| 3b | –1451 | 68 | 3270b |
| 4 | –1424 | 91 | 3268 |
aBIC: Bayesian information criterion.
bThe number of classes for the intervention end point model. This decision was based on picking the model with the best interpretability and the lowest BIC.