Continuous Quality Improvement (CQI) & Policy Feedback

Pipeline 10

A pipeline that cannot learn will repeat its mistakes. CQI turns dashboards and audits into training, memos, and better outcomes.

Overview

This node captures how jurisdictions review stage metrics, investigate disparities, document fixes, and publish transparent updates when policy shifts create series breaks.

What We Track

Benchmarks & Rate Ratios (Equity-Ready Analytics)

Benchmarks defined per stage (e.g., ≤10 days to school re-enrollment; ≤48h detention hearing)
Parity bands / rate ratios set (reference = 1.00) across race/ethnicity, gender, language, disability, rural/urban
Disaggregation rules documented (population vs. prior-stage denominators; cohort windows fixed)

Stage Reviews (Where We Look)

Detention: eligibility/overrides, ALOS/median, ATD utilization, FTA rates after release
Hearings: timeliness, continuance reasons, counsel presence, language access coverage
Supervision: contact standards met, response grid adherence, service dosage vs. plan
Reentry: first contact ≤7d, school re-enroll ≤10d, early stability (30/90/180)

Root-Cause & Action Logs (Closing the Gap)

Structured analysis used (5 Whys / Fishbone / A3); hypothesis recorded prior to action
Owner, due date, resources assigned; RACI noted for cross-agency tasks
Pre/post metric chosen (e.g., median days, completion rate, disparity index) for effect validation

Training & Policy Memos (Turning Insight into Practice)

Change documented: policy memo, SOP update, or script/template revision (version + effective date)
Training roster & completion tracked; refresher cadence set (e.g., annual/biannual)
Series break flagged in dashboards if definitions/tools changed

Transparency & Governance (Showing Our Work)

Dashboards with confidence bands, small-n flags, and notes on comparability
Public-facing summaries & stakeholder brief cadence (monthly/quarterly)
Data governance minutes archived; action status (“on track / at risk / blocked”) visible

Feedback Loop

Review stage metrics & disparities
- Compare to benchmarks and parity bands; confirm data quality (coverage, series breaks)
Investigate drivers; form hypotheses
- Run 5 Whys/Fishbone; inspect flow bottlenecks (timeliness, capacity, eligibility rules)
Draft & implement actions
- Publish policy/training updates; configure checklists/UI nudges; mark series breaks as needed
Re-measure & validate impact
- Pre/post comparison or PDSA cycle; adjust benchmarks if sustained improvement observed
Report out & iterate
- Share progress with stakeholders; log residual gaps; queue next cycle

Fields

Field	Type	Required	Codeset	Description
pipeline_place_id	uuid	✅	—	Unique identifier for this CQI review/action record.
pipeline_stage_id	enum	✅	stages.yml#stage_key(8) intake detention adjudication disposition supervision commitment_and_placement reentry closure	One of the canonical stages; defaulted here to reentry_aftercare.
pipeline_place_key	enum	✅	pipeline_places.yml#place_key(45) intake pre_petition_diversion_and_deflection diversion youth_assessment_tools prosecutorial_screening status_offenses adjudication case_planning family_engagement case_timeliness language_access restorative_justice detention_screening community_atds electronic_monitoring court_appearance_and_fta initial_hearing pretrial_supervision probation intensive_supervision …and 25 more	Canonical key for this place (maps to route/slug).
occurred_datetime	datetime	✅	—	Anchor timestamp (e.g., review meeting date or action decision time).
jurisdiction_code	string	✅	—	County/parish/circuit or standardized local code.
source_system	string	✅	—	Origin system name.
source_file	string		—	Source batch/file id if flatfile.
extract_run_id	string		—	ETL run id for lineage.
series_break_flag	boolean		—	Comparability break applies due to policy/process/tool change.
series_break_reason	enum		series_breaks.yml#reason(4) definition_change vendor_change coverage_change policy_change	Reason for series break.
review_domain_code	enum	✅	review_domains.yml#domain(10) intake detention hearings supervision reentry placement violation_response education workforce cross_agency	Which stage/focus area is under review (detention, hearings, supervision, reentry, etc.).
metric_code	enum	✅	metric_catalog.yml#metric(10) reenroll_within_10d detention_hearing_within_48h first_contact_within_7d atd_utilization_rate response_grid_adherence service_dosage_alignment attendance_30d_stability ui_earnings_any fta_rate_post_release other	The specific stage metric being reviewed (e.g., reenroll_within_10d, detention_hearing_within_48h).
benchmark_definition_code	enum	✅	benchmark_definitions.yml#benchmark(6) target_le_10d_school target_le_48h_hearing target_le_7d_contact target_cte_first_appt_le_7d locally_defined statutory_standard	Benchmark selected/defined for this metric.
parity_reference_group_code	enum		disparity_dimensions.yml#group⚠️ using disparity_dimensions.yml#dimension(5) race_ethnicity gender language disability rural_urban	Reference group for rate ratios (e.g., overall, White, etc.).
disaggregation_rule_code	enum		disaggregation_rules.yml#rule(4) population_denominator prior_stage_denominator fixed_exit_cohort rolling_window	Population vs. prior-stage denominators; fixed cohort window rule.
disparity_dimension_codes	array<string>		disparity_dimensions.yml#dimension(5) race_ethnicity gender language disability rural_urban	Dimensions examined (race_ethnicity, gender, language, disability, rural_urban); semicolon-delimited.
analysis_method_code	enum		root_cause_methods.yml#method(4) five_whys fishbone a3 pdsa	Structured method used (five_whys, fishbone, a3, pdsa).
hypothesis_document_id	string		—	Document or note reference recording the hypothesis before action.
action_type_code	enum		action_types.yml#type(5) policy_memo sop_update script_revision checklist_nudge training	Policy_memo, SOP_update, script_revision, checklist_nudge, training.
action_owner_staff_id	string		—	Responsible owner for this action.
action_due_date	date		—	Due date for action completion.
raci_roles	array<string>		raci_roles.yml#role(4) responsible accountable consulted informed	RACI roles assigned for cross-agency tasks; semicolon-delimited.
raci_assignee_staff_ids	array<string>		—	Staff ids aligned positionally to raci_roles; semicolon-delimited.
action_status_code	enum		action_statuses.yml#status(4) on_track at_risk blocked complete	on_track, at_risk, blocked, complete.
change_document_id	string		—	ID/link for policy memo/SOP/script/versioned artifact.
change_version	string		—	Version or effective label for the change.
change_effective_date	date		—	Effective date of the policy/SOP/script change.
training_required_flag	boolean		—	Whether a training was required due to this change.
training_completion_cadence_code	enum		training_cadences.yml#cadence(4) annual biannual quarterly ad_hoc	Annual, biannual, quarterly, ad_hoc.
training_total_required_count	integer		—	Count of staff required to complete training.
training_completed_count	integer		—	Count of staff who completed training.
pre_post_metric_code	enum		metric_catalog.yml#metric(10) reenroll_within_10d detention_hearing_within_48h first_contact_within_7d atd_utilization_rate response_grid_adherence service_dosage_alignment attendance_30d_stability ui_earnings_any fta_rate_post_release other	The metric chosen for pre/post comparison.
pre_period_start_date	date		—	Start of “pre” observation period.
pre_period_end_date	date		—	End of “pre” period (inclusive/exclusive per local rule).
post_period_start_date	date		—	Start of “post” observation period.
post_period_end_date	date		—	End of “post” observation period.
dashboard_publication_flag	boolean		—	True if dashboard was (or will be) publicly published.
confidence_band_method_code	enum		confidence_band_methods.yml#method(3) binomial_ci bootstrap bayesian	Method used to produce dashboard bands (binomial_ci, bootstrap, bayesian).
small_n_threshold_value	integer		—	Minimum cell size before suppression.
publication_cadence_code	enum		publication_cadences.yml#cadence(3) monthly quarterly annual	Monthly, quarterly, annual.
governance_minutes_document_id	string		—	Reference to governance minutes or decision log.
data_quality_issue_codes	array<string>		data_quality_issue_types.yml#issue(7) left_censoring right_censoring match_failure definitional_drift missing_fields series_break small_n	Coverage/gap notes (left/right censoring, match failures, definitional drift); semicolon-delimited.
data_steward_staff_id	string		—	Steward responsible for governance documentation.
Download CSVwhat_we_track.csv

EDORA • Learn