EDORA Learn β Pipelines
Referral and Intake Screening (Pipeline Entry)
Pipeline 01
Transparency note: intake data often come from multiple agency systems with variable field coverage. Timeliness and completeness checks are essential before comparison.
Overview
Referral and intake mark the entry into the juvenile justice process. Each jurisdiction maintains its own pathways, but the logic is similar: determine whether the youth should proceed to formal petition, be diverted to an alternative program, or be screened for detention. Screening tools and intake interviews combine structured data (offense type, prior record, risk level) with clinical or situational context. These decisions have high leverage, influencing later outcomes and case mix.
What We Track
Referral Sources
- Source recorded: law enforcement, schools, prosecutors, other agencies, community/caregiver
- Composition tracked over time (monthly/quarterly trend & seasonality)
Screening Tools (Risk & Behavioral Health)
- Risk/needs tools (e.g., YLS/CMI, SAVRY) applied and scored; coverage rate
- Behavioral health screens (e.g., MAYSI-2, CRAFFT) completed; coverage rate
- Version control & calibration documented (instrument, scoring rules, change dates)
Decision Logic
- Algorithmic/policy rules mapped to outcomes (diversion, release, detention screen, petition)
- Overrides recorded with justification and approver
Timeliness & Same-Day Screening
- Elapsed time from referral receipt β decision captured (median/IQR)
- % with same-day screening flagged
Cross-System Handoffs
- Redirects to education, behavioral health, or child welfare logged before court filing
- Receipt confirmed by receiving unit (timestamp, contact)
Typical Flow
- Referral received β intake logs a new case with demographics & alleged offense
- Law-enforcement citations/arrests
- School-based or prosecutor referrals
- Community or caregiver reports
- Initial screening β risk/needs + BH triage; confirm prior history
- Risk tools (e.g., YLS/CMI, SAVRY) applied & scored
- BH screens (e.g., MAYSI-2, CRAFFT) completed
- Family, education, and prior-contact checks performed
- Decision node β determine diversion, release, detention review, or petition
- Diversion β meets low-risk criteria; refer to community program
- Release β low risk with caregiver supervision plan
- Detention screening β higher risk; proceed to detention instrument
- Petition β ineligible for diversion; prepare filing
- Documentation β record all screening results & decisions
- Upload tool scores/forms; note overrides with rationale
- Time-stamp entries for audit tracking
- Data transmission β forward the referral packet to next responsible unit
- Court β petition & adjudication scheduling
- Diversion provider β program intake & monitoring
- Detention center β admission & 24β48h hearing
Fields
Field | Type | Required | Codeset | Description |
---|---|---|---|---|
pipeline_place_id | uuid | β | β | Unique identifier for this pipeline place row. |
pipeline_stage_id | enum | β | stages.yml#stage_key(8)
| One of the 8 canonical stages. |
pipeline_place_key | enum | β | pipeline_places.yml#place_key(45)
| Canonical key for this place (maps to route/slug). |
youth_id | string | β | β | Salted/hashed or state UID. No plaintext PII. |
staff_id | string | β | Staff primary actor for this place, if applicable. | |
occurred_datetime | datetime | β | β | Timestamp when this place occurred or was recorded. |
jurisdiction_code | string | β | β | County/parish/circuit or standardized local code. |
location_site_id | string | β | Site/facility/office identifier. | |
actor_role_code | enum | roles.yml#actor_role(7)
| Role primarily responsible for this place. | |
legal_case_id | string | β | Docket/case identifier if already created. | |
source_system | string | β | β | Origin system name. |
source_file | string | β | Source batch/file id if flatfile. | |
extract_run_id | string | β | ETL run id for lineage. | |
series_break_flag | boolean | β | Comparability break applies to this row. | |
series_break_reason | enum | series_breaks.yml#reason(4)
| Reason for break when flagged. | |
referral_source_code | enum | β | referral_sources.yml#source(10)
| Origin of the referral (law enforcement, school, prosecutor, caregiver, other agency). :contentReference[oaicite:0]{index=0} |
referral_received_datetime | datetime | β | β | Timestamp when referral packet or contact was first received at intake. :contentReference[oaicite:1]{index=1} |
intake_decision_datetime | datetime | β | Timestamp of the decision at this node (e.g., diversion offer, release, detention screen, petition prep). :contentReference[oaicite:2]{index=2} | |
risk_tool_name | enum | instruments.yml#risk_tool_name(2)
| Validated risk/needs instrument used (e.g., YLS/CMI, SAVRY). :contentReference[oaicite:3]{index=3} | |
risk_tool_version | string | β | Version identifier or effective date for the risk instrument. :contentReference[oaicite:4]{index=4} | |
risk_total_score | integer | β | Total score produced by the risk/needs instrument (raw or scaled). :contentReference[oaicite:5]{index=5} | |
bh_screen_name | enum | instruments.yml#bh_screen_nameβ οΈ using instruments.yml#risk_tool_name(2)
| Behavioral health screen used (e.g., MAYSI-2, CRAFFT). :contentReference[oaicite:6]{index=6} | |
bh_screen_version | string | β | Version identifier or effective date for the BH screen. :contentReference[oaicite:7]{index=7} | |
bh_immediate_flag | boolean | β | Immediate safety concern flagged by BH screening (triage). :contentReference[oaicite:8]{index=8} | |
intake_decision_code | enum | β | intake_decisions.yml#decision(4)
| Outcome at intake (diversion, release, detention_screen, petition). :contentReference[oaicite:9]{index=9} |
override_flag | boolean | β | True if decision overrode tool or policy guidance. :contentReference[oaicite:10]{index=10} | |
override_reason_code | enum | overrides.yml#reason(4)
| Reason code for override; approver captured in audit trail. :contentReference[oaicite:11]{index=11} | |
redirected_system_code | enum | cross_handoff_targets.yml#system(3)
| Early redirect target when handoff occurs before filing (education, behavioral_health, child_welfare). :contentReference[oaicite:12]{index=12} | |
receiving_unit_receipt_datetime | datetime | β | Timestamp of receipt confirmation by receiving unit (education/BH/CW). :contentReference[oaicite:13]{index=13} | |
Download CSVwhat_we_track.csv |
Data & Methods
Intake indicators use referral-based denominators to measure same-day decision rates and cross-agency handoffs. Risk-tool entries are validated against version logs, and overrides are coded as binary events with documented rationale. Linking intake data to education, mental-health, and child-welfare systems is essential for tracing pre-court diversions or non-admitted cases. See Data Linkage & Integration. When statutes or screening instruments change, series breaks are annotated per Series Breaks & Definition Changes. Small-sample counts follow Suppression & Small-n conventions.
Analysts should review data for completeness (β₯ 95 % of intakes with timestamp and tool version) and verify that referral-source codes align with jurisdictional definitions before comparison across sites.