EDORA Learn — Pipelines
Detention Decisions & Alternatives (Screening and Placement)
Transparency note: detention cohorts can be small in some jurisdictions. Policy memos or tool changes (e.g., revised risk thresholds) create series breaks; we pool low-n periods and annotate definition changes in metadata.
Overview
Detention decisions occur at intake and the first 24–48 hour hearing. Jurisdictions apply legal criteria and, often, a standardized risk screen to determine whether youth are held, released, or assigned to an alternative to detention (ATD) like evening reporting or electronic monitoring. This page outlines how decisions are made and recorded, how admissions and length of stay (LOS) are measured, and how equity and timeliness are tracked.
What We Track
- Screening & Criteria: detention risk instrument used, legal eligibility criteria, exclusion lists, and cut points with version dates.
- Overrides: departures from tool-indicated decisions with documented reasons (public safety, failure to appear history, victim concerns).
- Admissions & Releases: admission counts, time from admission to initial hearing, release pathways (to home, ATD, shelter/respite), and conditions.
- Length of Stay (LOS): median and 90th percentile LOS by legal status (pre-adjudication, post, warrants/holds) and by release pathway.
- Equity & Capacity: bed utilization, crowding days, rate ratios by race/ethnicity and geography, and access to ATDs as substitutes for secure beds.
Simple Flow
- Intake decision using legal criteria and risk screen; record any override.
- Admission or release: admit to detention, or release (with/without ATD conditions).
- Initial hearing within 24–48 hours; review detention decision and set next steps.
- Ongoing status checks (review hearings, case updates); consider step-down to ATD or release.
- Release with documented pathway and conditions; update LOS and outcome fields.
Data & Methods
Detention metrics use admission-based denominators and event timestamps (admission, hearing, release). LOS is reported with medians and 90th percentiles to handle skew. Risk screens and cut points are version-tagged; series breaksare marked when instruments or thresholds change. Equity monitoring uses rate ratios by race/ethnicity and geography following Cross-Jurisdiction Comparability. Exposure and timeliness conventions follow Denominators & Exposure Time and Time Series & Seasonality. Small cells are handled per Suppression & Small-n.