Preliminary DETECT R01 analysis

[mbcann01]

Overview

We met on 2024-03-27 to discuss the analysis plan for the DETECT R01 main outcomes.

2024-10-18: When I was at a meeting in Houston, an audience member asked a couple of good questions:

• What is the underlying rate of positive screens in the entire sample vs. the sample in the confusion matrix? There appear to be relatively few positive screens in that data.
• Dig into the 30 cases that were LEAD panel positive by DETECT screen negative. What did the DETECT tool seem to miss?

Links

• SharePoint folder for the manuscript
• Grant/Analysis plan

Task

• Create an overall confusion matrix using LEAD as the gold standard.
• Create Table 1
• Create confusion matrices for each abuse type using LEAD as the gold standard.
• Create confusion matrices using LEAD as the gold standard and each DETECT screening item as the exposure.
• Meet with Doug to discuss multiple imputations.
• Merge APS data for the APS analysis (Merge APS investigation outcomes into DETECT data #39) (Morri)