Skip to main content

Table 3 Identified interactions across main models and sensitivity analyses

From: Identifying treatment effects of an informal caregiver education intervention to increase days in the community and decrease caregiver distress: a machine-learning secondary analysis of subgroup effects in the HI-FIVES randomized clinical trial

Outcomes

Models

Interaction effects

Type of analysis

Days not at home (count specification)

Generalized Poisson regression tree

Baseline Zarit Burden score (≤ 28, > 28)

Primary

Random forest with interactions

Baseline Zarit Burden score, CESD-10, nosos score and patient age

Sensitivity

Proportion of days not at home

mCART regression tree

None

Sensitivity

Any days not at home

Generalized linear regression tree (binomial)

None

Sensitivity

mCART classification tree

None

Sensitivity

CESD-10 score 12-month follow-up

Linear regression tree

Baseline CESD-10 score (≤ 8, > 8)

Primary

mCART regression

None

Sensitivity

Random forest with interactions

Baseline CESD-10 score and, Zarit Burden score

Sensitivity