NSAG / Framework Foundation

Framework Foundation · Version 1.0

The NSAG Framework:
What the tiers mean.

Every tier, every dimension, every module — operationalized with specific, observable, documentable governance criteria. This is the intellectual foundation of the 15 educational self-assessments that accompany the framework.

Tier Definitions All 15 Modules Module Relationships

Scope & limits

What this is

A structured model of what good governance covers across six dimensions, published so that an institution — or an individual inside one — can see where its handling of those dimensions is explicit and where it is improvised. The self-assessments that once accompanied the modules are retired; what follows is the reasoning they were built on, which is the part that was always doing the work.

What this is not

Not a clinical instrument, a validated psychometric index, an accreditation, or a measure of neural activity. The framework is grounded in physiological and environmental-design research; that research motivates the six dimensions, it does not certify a score. No score is produced here. The written-evidence test in each tier definition below is the bar, and it is deliberately answerable only with documents.

Part I

Global Tier Definitions

Tiers apply uniformly across all 15 modules. An institution's tier reflects the state of its governance infrastructure — not its intentions, values, or aspirations.

Pioneering Emerging Developing Early Stage

Tier	What it means	The governance test
PIONEERING	The institution has fully implemented documented governance standards for this module domain. Standards are written and formally adopted by institutional leadership; tracked through operational metrics reviewed at least quarterly; accountability is assigned to a named role; standards survive leadership transitions because they are embedded in institutional policy that outlasts any individual.	Can the institution produce written evidence of implementation, measurement, and leadership accountability — today, without advance notice?
EMERGING	The institution has assessed its governance gap, adopted formal standards for at least four of six dimensions, and is actively implementing the remaining two. At least one governance change has been documented and measured. Leadership has formally acknowledged the domain as a governance responsibility.	Has the institution documented at least one governance change in this domain and measured its effect?
DEVELOPING	The institution has conducted some form of governance assessment in this domain and has begun drafting standards or making informal changes in at least two of six dimensions. Changes are in motion but are not yet formally adopted, tracked, or tied to accountability structures. Implementation depends on individual champions rather than institutional policy.	Would the governance change survive the departure of the person currently driving it?
EARLY STAGE	The institution has no formal governance framework for this domain. Effects in this domain are produced by the institution without measurement, governance standards, or accountability. This is a structural description. Most institutions begin here in most domains.	Does the institution know what it is currently producing in this domain — in measurable, documentable terms?

Part II

Module-by-Module Operationalization

Each module below contains: (1) why the module exists with its primary evidence base, (2) the six dimensions and why each was selected, and (3) specific, observable governance criteria for each of the four tiers.

Trauma-Informed Legal Space

Legal & Justice

▼

Up to 90% of justice-involved youth report exposure to at least one traumatic event (Dierkhising et al., 2013). Legal institutions are structurally activating — adversarial processes, authority with power over outcomes, public exposure — and most have no governance framework for the physiological effects they produce.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Physical Environment Design	Physical space produces measurable physiological activation states (Ulrich, 1984; SAMHSA, 2014). Acoustic separation, lighting, and seating are modifiable governance levers.
2	Procedural Design & Plain Language	Reading-level failures produce compounded threat activation. SAMHSA's principle of Trustworthiness and Transparency operationalizes as plain language.
3	Interpreter & Translation Access	Requiring clients to self-identify language need places the burden on the person least able to advocate in a high-stakes moment.
4	Staff Training & Trauma-Informed Behavior	McKenna & Holtfreter (2021) document that trained staff de-escalation reduces re-traumatization even before structural changes are in place.
5	Access & Accommodation Infrastructure	Neurodivergence and physical accessibility failures produce compounded activation in populations already overrepresented in legal contexts.
6	Feedback & Accountability Mechanisms	Without a mechanism structurally separate from the case relationship, the institution cannot identify re-traumatization patterns.

Tier Criteria

PIONEERING

Physical environment assessed against trauma-informed standards; modifications documented; assessment updated every 2 years
All client-facing documents at or below 6th-grade reading level; reading level audit documented within past 3 years
Proactive language identification protocol: language need identified before client self-identification required; interpreter access documented within 24 hours
100% of client-facing staff trained in trauma-informed practice within past 2 years; training includes co-regulation, de-escalation, and dissociation recognition
Formal accommodation process does not require self-advocacy in adversarial context; neurodivergence accommodations documented as standard intake step
Client feedback mechanism structurally independent of case relationship; reviewed quarterly by leadership; documented responses to findings

EMERGING

Physical environment assessment completed; at least 3 of 6 modifications implemented
Plain language audit completed; ≥75% of documents at or below 8th-grade reading level
Interpreter protocol in place for top 3 languages in service population
≥60% of client-facing staff trained within past 3 years
Accommodation process documented; not yet independent of case relationship
Feedback mechanism exists; review frequency not regular

DEVELOPING

Assessment initiated; 1–2 modifications made informally
Plain language review underway; no formal target adopted
Interpreter access available but not proactively identified
Training completed by some staff; not systematically required
Accommodation process informal; staff-dependent
Feedback mechanism informal or absent

EARLY STAGE

No physical environment assessment
No plain language standard; documents written for legal professionals
Interpreter access reactive; client must request
No trauma-informed training requirement
No formal accommodation process
No independent client feedback mechanism

Open the M1 module scope →

AI Legal Navigation

Legal & Justice

▼

Algorithmic risk assessment tools cover approximately two-thirds of the US population in pretrial and sentencing decisions (Berk et al., 2024). Dressel & Farid (2018) demonstrated that COMPAS performs at the level of untrained human prediction while producing racial disparities in false-positive rates.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	AI Tool Transparency & Explainability	Due process requires that individuals can understand and contest the basis of decisions affecting them. Algorithmic opacity violates this at scale.
2	Bias Audit & Demographic Impact	Dressel & Farid (2018) and Chouldechova (2017) establish that racial disparities are structural and systematic. Independent audit is required for the specific population served.
3	Human Oversight & Override Protocols	An override protocol that requires extensive written justification effectively coerces deference to the algorithm. The governance value of an override depends on whether it can be exercised without penalty.
4	Plain-Language Communication to Affected Parties	People subject to AI-influenced decisions have a due process right to know an algorithm influenced their case.
5	Accountability & Error Correction	Without a remediation pathway, documented errors produce harm without institutional consequence.
6	Community Input & Governance Structures	The populations most affected by algorithmic risk assessment must have formal participation in adoption and discontinuation decisions.

Tier Criteria

PIONEERING

Plain-language documentation of every AI tool influencing legal decisions; publicly available; updated within 12 months of any algorithm change
Independent demographic impact assessment for each tool; specific to the jurisdiction's population; results reviewed by leadership; updated annually
Override protocol: no documentation burden beyond standard case documentation; no adverse consequences for overriding; override rate tracked and reviewed
Every individual notified in plain language when AI influenced a decision in their case; notification before decision takes effect
Error classification system; remediation includes review of prior affected cases; remediation timeline enforced
Formal community advisory body with authority to recommend discontinuation; composed of people directly affected by algorithmic risk assessment

EMERGING

Documentation of major AI tools available; plain-language summary for ≥50% of tools
Demographic impact assessment for primary risk tool; reviewed by leadership at least once
Override protocol documented; burden reduction measures taken
Notification in place for primary risk assessment tool
Error classification system exists; remediation informal
Community input mechanism exists; formal authority not yet established

DEVELOPING

Some internal documentation; not plain language; not public
Bias awareness acknowledged; no formal audit
Override technically possible; no formal protocol
Notification informal; staff-dependent
Error tracking informal; no remediation process
Community input informal; no structured mechanism

EARLY STAGE

No plain-language documentation
No bias audit; relying on vendor validation
Override possible in principle; no protocol; no tracking
No notification policy
No error classification or remediation
No community input

Open the M2 module scope →

Psychedelic Harm Reduction Governance

Healthcare & Clinical

▼

FDA Breakthrough Therapy designation for psilocybin (2018, 2019), the NEJM psilocybin trial (Carhart-Harris et al., 2021), and the JAMA Psychiatry RCT (Davis et al., 2021) establish clinical legitimacy. Johnson, Richards & Griffiths (2008) establish that adverse events are primarily determined by governance factors — set, setting, supervision — not pharmacological properties.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Participant Screening Protocols	Contraindication screening is the primary risk-reduction tool established by Johnson et al. (2008).
2	Preparation & Informed Consent	Informed consent for psychedelic sessions requires consent for specific phenomenological experiences the participant cannot anticipate from general descriptions.
3	Physical & Interpersonal Setting Standards	Johnson et al. (2008) identify session environment as a primary adverse-event determinant. Set and setting are not metaphors — they are safety standards.
4	Facilitator Training & Scope of Practice	Facilitator conduct during a session is a primary safety lever. Scope of practice documentation protects both participant and institution.
5	Adverse Event Classification & Response	Without classification, challenging experiences, psychological crises, and medical emergencies receive the same response — producing harm in each direction.
6	Integration Support	The post-session integration period is when consolidation of therapeutic gains and working through difficult material occurs. Absence of integration support predicts negative post-session outcomes.

Tier Criteria

PIONEERING

Written contraindication criteria covering psychotic-disorder history, cardiovascular conditions, and active SUD; reviewed annually by licensed clinician
Preparation protocol: minimum 2 sessions before psychedelic session; written informed consent specifying compound, dose range, phenomenological range, and right to discontinue
Physical session environment standards documented: minimum acoustic dampening, blackout capability, reclining furniture, safety equipment accessible within 30 seconds; inspected annually
Minimum facilitator training hours documented; credentials verified; scope of practice agreement signed; supervision structure in place
Three-tier adverse event classification: challenging experience, psychological crisis, medical emergency; response protocols for each; simulation-trained annually
Minimum 3 integration sessions within 30 days post-session; referral pathway to licensed therapist documented; integration completion tracked

EMERGING

Contraindication criteria documented; not recently updated
Preparation protocol in place; consent covers major elements
Setting standards documented for primary session space
Training hours documented; supervision informal
Two-tier adverse event classification (support vs. emergency)
At least 1 integration session offered; referral pathway documented

DEVELOPING

Contraindication criteria informal; not institutional policy
Preparation conversations occur; no structured protocol
Setting standards informal; facilitator-dependent
Training hours recorded; no scope of practice agreement
Adverse event response: 'call 911 if needed'; no classification
Integration encouraged; not structured or tracked

EARLY STAGE

No written contraindication criteria
No preparation protocol; consent is general medical consent
No documented setting standards
No training requirements or scope of practice documentation
No adverse event classification; no response protocol beyond emergency services
No integration requirement or support structure

Open the M3 module scope →

Cannabis Public Health Infrastructure

Healthcare & Clinical

▼

Boehnke, Litinas & Clauw (2016) documented a 64% decrease in opioid use among medical cannabis patients. VanDolah et al. (2019) document CYP450 drug interaction pathways. Sajdeya et al. (2021) identify six structural barriers to documentation that exist entirely within institutional design.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Cannabis Screening Standards	Sajdeya et al. (2021) identify absence of standardized screening as the primary structural barrier; without a standard instrument, documentation is ad hoc.
2	Provider Training	Clinician knowledge gaps are the second structural barrier. CYP450 interaction risk (VanDolah et al., 2019) requires pharmacological competence.
3	EHR Documentation Architecture	Beiler et al. (2024) demonstrate that documentation completeness is infrastructure-dependent: structured fields produce 93.6% completion vs. 30.8% without.
4	Disclosure Safety Infrastructure	Patient fear of employment/insurance/punitive consequences is a structural barrier. Institutional design either mitigates or amplifies that fear.
5	Regulatory Compliance	Federal Schedule I status creates genuine legal ambiguity; documented legal guidance protects both patient and institution.
6	Quality Improvement Metrics	Without tracking documentation rates as a QI metric, there is no institutional visibility into the gap and no organizational pressure to close it.

Tier Criteria

PIONEERING

Validated cannabis screening instrument (CUDIT-R or equivalent) at every patient contact; completion tracked; target ≥85%
All providers trained in cannabis pharmacology within 2 years; training includes CYP450 interaction protocol; CME credit offered
Cannabis-specific structured EHR data fields in intake and update workflows; completion rate tracked; target ≥85%
Written intake policy stating documentation cannot be shared adversarially; policy communicated in writing at intake
Documented legal guidance reviewed by counsel within past 2 years on documentation obligations and protections
Cannabis documentation rate tracked as a named QI metric; reviewed quarterly; improvement target set annually

EMERGING

Cannabis screening question in intake; not validated instrument; tracked
Provider training available; completed by ≥60% of clinical staff
At least one cannabis-specific EHR field in intake workflow
Disclosure safety policy exists; not systematically communicated
Legal guidance obtained; not recently reviewed
Documentation rate tracked; not yet formal QI metric

DEVELOPING

Cannabis question in some intake forms; not systematic
Provider training informal; available on request
EHR documentation in free-text only
Disclosure safety addressed informally by individual providers
Legal guidance not formally obtained
No documentation rate tracking

EARLY STAGE

No cannabis screening in standard intake
No provider training on cannabis pharmacology
No cannabis-specific EHR fields
No disclosure safety policy
No legal guidance on documentation
No QI tracking

Open the M4 module scope →

Biophilic Civic Infrastructure

Civic & Built Environment

▼

Ulrich (1984) established in a nine-year matched cohort that window view alone produced measurable differences in surgical recovery outcomes. Ulrich (1991) provided the psychophysiological mechanism (Stress Recovery Theory). Al Khatib et al. (2024) confirmed across a 2010–2023 systematic review that biophilic design reduces hospitalization time, mortality, pain, and staff stress.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Physical Environment Standards	Ulrich (1984, 1991) and Al Khatib et al. (2024) establish that natural light, nature views, and biophilic elements produce measurable physiological and clinical outcomes.
2	Outdoor Access Infrastructure	Razani et al. (2018) demonstrate in a randomized trial that structured nature access (park prescriptions) is a governable health intervention.
3	Built Environment Assessment	Without a formal assessment, the institution cannot identify what it is currently producing or measure whether it has changed.
4	Community Green Space Standards	Sarkar et al. (2018) establish that the equity burden of green space deprivation falls on lower-income populations; civic institutions have a governance obligation.
5	Nature-Based Programming	Institutionalizing nature access through programming creates the park-prescription pipeline demonstrated in Razani et al. (2018).
6	Equity Assessment	Without an explicit equity assessment, biophilic improvements benefit primarily those with the most institutional access.

Tier Criteria

PIONEERING

Documented minimum standards for natural light, nature views, and biophilic elements in all client-facing spaces; assessed annually
Formal outdoor access provision: scheduled outdoor breaks for staff; patient/client outdoor access built into protocols; documented in encounter records where clinically relevant
Evidence-based built environment assessment completed within 3 years; documented against published standards; findings reported to leadership
Documented institutional position on green space in surrounding community; institution participates in or supports local green space advocacy
At least one nature-based programming component formally offered; participation tracked
Equity assessment of biophilic infrastructure access completed; documented disparities addressed in improvement plan; affected communities named explicitly

EMERGING

Natural light and nature view standards documented for primary spaces
Outdoor access available but not formally structured
Built environment assessment completed; not recently updated
Community green space position informal; some participation
Nature-based programming offered informally
Equity assessment begun; not yet completed

DEVELOPING

Some biophilic elements present; no documented standards
Outdoor access available but not provided; staff-dependent
No formal assessment; informal awareness
No documented community green space position
No formal nature-based programming
No equity assessment

EARLY STAGE

No documented biophilic standards
No structured outdoor access
No built environment assessment
No community green space engagement
No nature-based programming
No equity assessment

Open the M5 module scope →

Ethical Civic Sponsorship

Legal & Justice

▼

Fooks et al. (2013) documented how corporations use CSR frameworks to neutralize regulatory threats and build civic relationships. Fabbri et al. (2018) found in a systematic review of 36 articles that industry sponsorship consistently drives research agendas away from public-health-relevant questions. Lundh et al. (2017) confirmed in a Cochrane review that industry-sponsored studies are more likely to produce favorable outcomes and conclusions.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Sponsorship Ethics Policy	Fooks et al. (2013) demonstrate that CSR relationships produce structural governance incentives regardless of individual intent; a written policy creates institutional resistance.
2	Mission Alignment Review	Fabbri et al. (2018) demonstrate that research agendas shift under sponsorship; a formal review detects this before it occurs.
3	Exclusion Criteria	Some industries' documented public health harms are so established that mission alignment review will routinely produce the same result; exclusion criteria make that result a governance standard.
4	Transparency Standards	Public disclosure is the accountability mechanism that makes governance real; without it, there is no community check on sponsorship decisions.
5	Conflict of Interest Management	Sponsorship creates financial relationships that distort programming, research, and advocacy without any individual acting in bad faith.
6	Review & Accountability Mechanism	Without an independent review body with authority to decline or terminate, policy exists without enforcement.

Tier Criteria

PIONEERING

Written sponsorship ethics policy formally adopted by board; specifies acceptable and unacceptable categories; reviewed every 3 years
Formal mission alignment review for each prospective sponsor; includes review of regulatory history, public health impact, and community relationships; documented and signed
Explicit written exclusion criteria for industries with documented public health harms; publicly available
All sponsorship relationships publicly disclosed on institutional website within 30 days of agreement; disclosure includes amount, duration, and scope
Written conflict of interest policy; affected staff recuse from decisions where sponsor has an interest; recusals documented
Independent review body with documented authority to recommend against acceptance; at least one rejection or termination on record

EMERGING

Written sponsorship ethics policy in place; not reviewed in past 3 years
Mission alignment review occurs; not formally documented
Exclusion criteria informal; case-by-case leadership judgment
Disclosure on website; not always timely or complete
Conflict of interest policy exists; recusal informal
Leadership review in place; no independent body

DEVELOPING

Sponsorship decisions by leadership without formal policy
Mission alignment considered informally
Some exclusions applied; no written criteria
Disclosure informal; not systematic
Conflict of interest awareness; no formal management
No independent review

EARLY STAGE

No written sponsorship ethics policy
No mission alignment review
No exclusion criteria
No disclosure policy
No conflict of interest management
No independent review

Open the M6 module scope →

Conscious Cities

Civic & Built Environment

▼

Evans (2003) established built environment as a mental health determinant across residential density, housing quality, neighborhood disorder, and noise. Sarkar et al. (2018) demonstrated in 94,879 UK Biobank participants that residential greenness reduces MDD odds with the largest protective effects in lower-income populations. Münzel et al. (2014) documented the cardiovascular and neuroendocrine mechanism for noise-induced harm.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Mental Health Impact Assessment	Evans (2003) establishes built environment as a mental health determinant; without a formal assessment requirement, planning decisions continue to be made without governing their primary human consequence.
2	Green Space Standards	Sarkar et al. (2018) establish green space as a measurable MDD risk modifier; documented minimum standards make the equity obligation governable.
3	Noise & Environmental Stressor Management	Münzel et al. (2014) establish noise as a physiological stressor with cardiovascular and neuroendocrine consequences; governance requires monitoring and standards.
4	Housing Quality Standards	Evans (2003) establishes housing quality as a mental health determinant independent of socioeconomic factors.
5	Community Participation in Design	The populations most affected by built environment decisions are least represented in the processes that make them.
6	Equity Impact Assessment	Sarkar et al. (2018) establish that green space deprivation burdens lower-income populations disproportionately; without an explicit equity assessment, governance perpetuates the disparity.

Tier Criteria

PIONEERING

Mental health impact assessment required for all built environment decisions above a defined threshold; methodology documented; findings reported to decision-making body before vote
Documented minimum green space access standard; compliance tracked; deficiencies documented and addressed in capital planning
Noise monitoring in primary service area; data reviewed quarterly; documented threshold requiring governance response; at least one governance response documented
Housing quality standards documented for population served; inspection data reviewed; partnership with housing authority for enforcement
Community participation mechanism in place; composition documented; historically underrepresented populations specifically included; documentation of how input influenced decisions
Equity impact assessment required for all major built environment decisions; specific populations named; findings incorporated into decision documentation

EMERGING

Mental health impact considered in major decisions; not formally required
Green space standard for major developments; not citywide
Noise monitoring in primary areas; threshold and response informal
Housing quality engagement with housing authority; no formal standards
Community input mechanism exists; composition not specified
Equity assessment informal; not required

DEVELOPING

Mental health awareness in planning; no formal assessment
Green space planning considered; no standard
Noise complaints tracked; no proactive monitoring
Housing quality addressed case-by-case
Community input through public comment only
Equity awareness; no formal assessment

EARLY STAGE

No mental health impact assessment
No green space standard
No noise monitoring or governance standard
No housing quality governance
No formal community participation mechanism
No equity impact assessment

Open the M7 module scope →

Burnout Recovery Infrastructure

Healthcare & Clinical

▼

Maslach, Schaufeli & Leiter (2001) established burnout as a mismatch across six organizational domains. Kiratipaisarl et al. (2024) confirmed in a 2024 meta-analysis that organizational interventions produce significantly greater and more durable burnout reductions than individual interventions. Thomas Craig et al. (2021) found in a systematic review that EHR-targeting organizational interventions improved burnout in 68% of eligible studies.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Workload Assessment & Governance	Maslach et al. (2001) identify workload as the primary burnout determinant; without measurement and governance standards, workload increases occur without institutional visibility.
2	Autonomy & Control Standards	Maslach et al. (2001) identify control as the second burnout determinant; organizational governance of decision-making latitude is the structural response.
3	Reward & Recognition Infrastructure	Maslach et al. (2001) identify reward as the third domain; reward governance is distinct from compensation — it includes recognition systems that are consistent, transparent, and independent of individual personalities.
4	Community & Social Support Standards	Maslach et al. (2001) identify community as the fourth domain; governance requires structural provision built into the organization.
5	Fairness & Equity Standards	Maslach et al. (2001) identify fairness as the fifth domain; perceived inequity in workload distribution and recognition is a direct burnout driver.
6	Values Alignment Assessment	Maslach et al. (2001) identify values as the sixth domain; values misalignment — being required to act against professional or personal ethics — is a burnout driver no wellness program addresses.

Tier Criteria

PIONEERING

Workload governance standard documented: maximum caseload/patient load/contact hours by role; workload reported through formal channel quarterly; reports reviewed by leadership; documented response required within 30 days
Autonomy governance: decision-making scope specified by role; autonomy reductions require formal documentation and review; staff survey on perceived control conducted annually
Recognition system: criteria transparent; not supervisor-dependent; peer recognition mechanism in place; recognition frequency tracked
Peer support program formally offered; structured (not just encouraged); utilization tracked; program reviewed annually
Fairness governance: workload distribution equity reviewed quarterly; grievance mechanism not requiring supervisor as first contact; grievance outcomes tracked
Values alignment assessment conducted annually: specific survey on values misalignment; results reviewed by leadership; documented response to specific conflicts raised

EMERGING

Workload standard documented; not recently reviewed; formal reporting channel exists but infrequently used
Autonomy scope documented; not routinely reviewed
Recognition system in place; not systematic; supervisor-dependent
Peer support available; not formally structured
Workload equity reviewed annually; grievance mechanism exists
Values alignment discussed in annual review; not formally assessed

DEVELOPING

Workload standards informal; staff report concerns informally
Autonomy expectations implicit; not documented
Recognition occurs; no standard or tracking
Peer support encouraged; no formal structure
Fairness concerns addressed case-by-case
Values alignment not formally addressed

EARLY STAGE

No workload governance standard
No autonomy governance
No formal recognition system
No peer support program
No formal fairness mechanism
No values alignment assessment

Open the M8 module scope →

Cannabis Healthcare Visibility

Healthcare & Clinical

▼

Lapham et al. (2022) found in JAMA Network Open that 35.1% of primary care patients reported implicit medical cannabis use while only 4.8% had EHR documentation — a verified 7:1 gap within the same patient population. Tavabi et al. (2023) found documentation rates below 2% across 370,087 patients with significant racial and socioeconomic disparities independent of use prevalence.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Cannabis Screening Standards	Lapham et al. (2022) document the 7:1 documentation gap; without a standardized instrument, there is no systematic way to close it.
2	Provider Training Standards	Sajdeya et al. (2021) identify clinician knowledge gaps as a primary structural barrier; VanDolah et al. (2019) document the CYP450 interaction risk requiring pharmacological competence.
3	EHR Documentation Standards	Beiler et al. (2024) demonstrate that documentation completeness is infrastructure-dependent: structured fields produce 93.6% completion vs. 30.8% without.
4	Disclosure Safety Infrastructure	Patient fear of employment/insurance/punitive consequences produces the disclosure avoidance driving the gap; the institution's design either mitigates or amplifies that fear.
5	Regulatory Compliance & Legal Review	Federal Schedule I status creates genuine legal ambiguity requiring documented legal guidance for both patient and institution.
6	Quality Improvement Integration	Without tracking documentation rates as a QI metric, there is no institutional visibility into the gap and no organizational pressure to close it.

Tier Criteria

PIONEERING

Validated cannabis use screening instrument at every patient contact; completion tracked; ≥85% target; instrument updated within 5 years
All clinical providers trained in cannabis pharmacology within 2 years; covers CYP450 interactions, contraindications, disclosure conversation standards; completion tracked by role
Standardized cannabis-specific structured EHR data fields in intake and update workflows; completion rate ≥85%; data reviewed quarterly for quality
Written intake policy: cannabis disclosure cannot be used adversarially; communicated verbally and in writing at intake; staff trained on communication; reviewed annually by counsel
Documented legal guidance from qualified counsel on documentation obligations and protections; updated within 3 years or upon any relevant legal change
Cannabis documentation rate tracked as named QI metric; reviewed quarterly by clinical leadership; annual improvement target set and reported

EMERGING

Cannabis screening question in intake; not validated instrument; tracked; ≥60% completion
Provider training available and completed by majority of clinical staff; not universally required
At least one cannabis-specific EHR field in intake workflow
Disclosure safety policy documented; communicated inconsistently
Legal guidance obtained informally; not recently updated
Documentation rate tracked but not formal QI metric

DEVELOPING

Cannabis question in some intake forms; not systematic across the institution
Provider training informally available; no completion tracking
EHR cannabis documentation in free-text only; no structured fields
Disclosure safety addressed informally by individual providers
Legal guidance not formally obtained
No documentation rate tracking

EARLY STAGE

No cannabis screening in standard intake
No provider training on cannabis pharmacology
No cannabis-specific EHR fields
No disclosure safety policy
No legal guidance on documentation
No QI tracking

Open the M9 module scope →

M10

Healthcare Built Environment

Healthcare & Clinical

▼

Ulrich et al. (2008) synthesized approximately 600 empirical studies on healthcare built environment effects. Al Khatib et al. (2024) confirmed in a 2024 systematic review that biophilic design reduces hospitalization time, mortality, pain, and staff stress. Münzel et al. (2014) established the cardiovascular and neuroendocrine mechanism for noise-induced harm.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Evidence-Based Environmental Assessment	Ulrich et al. (2008) demonstrate that built environment has documented clinical effects; without assessment against evidence-based standards, the institution cannot govern what it produces.
2	Acoustic Standards	Ulrich et al. (2008) establish noise above 45 dBA as a documented patient and staff physiological stressor; Münzel et al. (2014) establish the cardiovascular and neuroendocrine mechanism.
3	Lighting Standards	Ulrich et al. (2008) document natural light effects on length of stay and staff sick days; lighting is a modifiable governance lever.
4	Single-Patient Room Standards	Ulrich et al. (2008) document 10–50% infection rate reductions with single-patient rooms; this is the strongest governance lever in the evidence base.
5	Wayfinding & Spatial Orientation	Spatial disorientation produces anxiety, missed appointments, and delayed care-seeking — measurable clinical outcomes of a governable design variable.
6	Biophilic Design Standards	Al Khatib et al. (2024) confirm that biophilic design reduces hospitalization time, mortality, pain, and staff stress — clinical outcomes governable through design standards.

Tier Criteria

PIONEERING

Full evidence-based environmental assessment completed within 3 years using a validated tool; findings reported to leadership; improvement plan documented
Continuous noise monitoring in all patient care areas; documented target ≤45 dBA in patient rooms during sleep hours; exceedance events tracked; governance response documented for persistent exceedance
Daylight analysis completed; minimum daylight access standard documented for patient care areas; compliance tracked
Single-patient room standard documented; protocol for shared-room situations; shared-room assignment rate tracked; reduction plan in place if rate exceeds target
Wayfinding assessment completed; patient-reported difficulty tracked; modification plan documented; improvement assessed post-modification
Minimum biophilic design standard documented for patient-facing environments; natural elements inventoried; compliance assessed annually

EMERGING

Environmental assessment completed within 5 years; findings partially addressed
Noise monitoring in primary patient care areas; target documented; response informal
Daylight standard documented; not fully implemented
Single-patient room standard documented; protocol for shared rooms
Wayfinding assessment completed; modifications made
Biophilic elements present; no formal standard

DEVELOPING

Environmental assessment informal; some awareness of evidence-based design
Noise complaints tracked; no proactive monitoring; no formal target
Lighting awareness; no standard
Single-patient rooms preferred; no governance standard
Wayfinding issues recognized; modifications informal
Biophilic elements present; plants in lobby; no standard

EARLY STAGE

No evidence-based environmental assessment
No noise monitoring or standard
No lighting standard
No single-patient room governance
No wayfinding assessment or standard
No biophilic design standard

Open the M10 module scope →

M11

Medical Technology & Evidence Standards

Healthcare & Clinical

▼

Obermeyer et al. (2019) found in Science that an algorithm used for approximately 200 million patients produced systematic racial bias; remedying it would raise Black patients receiving extra help from 17.7% to 46.5%. The ASTP/ONC Data Brief No. 80 (Chang et al., 2025) documents that 71% of hospitals use EHR-integrated predictive AI while fewer than half evaluate all models for demographic bias.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	AI Tool Inventory & Transparency	Chang et al. (2025) document that 71% of hospitals use EHR-integrated predictive AI; governance requires knowing what tools exist and what they do.
2	Independent Evidence Evaluation	Obermeyer et al. (2019) demonstrate that vendor validation and FDA clearance did not detect systematic racial bias at scale; independent evaluation is required.
3	Demographic Impact Assessment	The bias in the Obermeyer et al. (2019) case was detectable through demographic stratification of outcomes data; this dimension governs that detection process.
4	Human Oversight & Override Protocols	Chang et al. (2025) document that most hospitals lack clear AI accountability ownership; override protocols maintain human decision-making authority.
5	Patient Disclosure Standards	Patients have a governance right to know when algorithms influence their care.
6	Post-Adoption Monitoring & Suspension Standards	The Obermeyer et al. (2019) bias persisted at scale without detection; post-adoption monitoring with defined thresholds is the detection mechanism.

Tier Criteria

PIONEERING

Complete inventory of all AI tools influencing clinical decisions; publicly available; updated within 12 months of any change; plain-language description of each tool's function
Internal evaluation standard documented, independent of vendor claims and regulatory clearance; conducted before adoption; includes methodology, evidence base, and population applicability review
Demographic impact assessment required before adoption and annually post-adoption; stratified by race, ethnicity, sex, age, and socioeconomic status for the specific patient population; results reviewed by leadership
Override protocol: no documentation burden beyond standard clinical documentation; no adverse consequences for overriding; override rate tracked and reviewed quarterly
All patients notified in plain language when an AI tool was used in their care; notification delivered before decision takes effect; right to request human-only review documented
Monitoring plan for each AI tool; outcome metrics stratified by demographic group; defined disparity threshold requiring governance response; suspension criteria documented; at least one tool reviewed annually for continued use

EMERGING

AI tool inventory maintained internally; not yet publicly available
Internal evaluation standard in development; applied to primary tools
Demographic impact assessment for primary clinical AI tools; results reviewed
Override protocol documented; burden reduction in progress
Patient notification for primary risk tools
Monitoring plan for primary tools; suspension criteria informal

DEVELOPING

Major AI tools identified internally; no formal inventory
Vendor validation relied on primarily; some informal internal review
Demographic impact awareness; no formal assessment conducted
Override possible; no formal protocol
Notification informal; tool-dependent
Monitoring informal; outcome data not stratified

EARLY STAGE

No AI tool inventory
No internal evaluation standard; relying on vendor and regulatory clearance
No demographic impact assessment
No override protocol
No patient notification standard
No monitoring plan

Open the M11 module scope →

M12

Nervous-System-Aware Education Systems

Civic & Built Environment

▼

Immordino-Yang & Damasio (2007) established that physiological threat states suppress prefrontal cortical access required for learning. Barrett et al. (2015) found in the HEAD Project (153 classrooms, 3,766 pupils) that seven built environment parameters explain 16% of the variation in pupils' academic progress. Cantor et al. (2019) established that educational environment decisions during critical developmental periods literally shape neural architecture.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Physical Environment Standards	Barrett et al. (2015) establish that seven built environment parameters explain 16% of learning progress variation; without documented standards, this 16% is ungoverned.
2	Psychological Safety Infrastructure	Immordino-Yang & Damasio (2007) establish that threat states suppress prefrontal access required for learning; psychological safety is a neurobiological prerequisite.
3	Individualization & Flexibility Standards	Barrett et al. (2015) identify individualization as one of the seven learning-environment parameters; flexibility accommodates diverse nervous system profiles.
4	Transition & Schedule Governance	Predictability reduces threat activation in individuals with trauma histories (Felitti et al., 1998); schedule governance is a trauma-informed educational infrastructure decision.
5	Trauma Awareness in Educational Design	The ACE Study (Felitti et al., 1998) documents dose-response relationship between adverse childhood experiences and adult outcomes; institutional design either activates or buffers ACE sequelae.
6	Quality Assessment	Without systematic tracking of environment quality against evidence-based standards, improvements cannot be identified, sustained, or attributed.

Tier Criteria

PIONEERING

Documented standards for lighting, acoustics, temperature, and air quality in all primary learning spaces; compliance assessed annually; standards aligned with Barrett et al. (2015) HEAD Project parameters
Formal psychological safety infrastructure: social environment standards documented; regular student psychological safety surveys with institutional response requirement; adult learning safety standards separate from student standards
Flexibility standards: ≥30% of primary learning spaces have flexible furniture; 100% have individualization provision (personal storage, display space, environmental adjustment capacity)
Transition governance: maximum class transitions per day documented; protocols standardized; schedule changes communicated ≥48 hours in advance except emergencies; predictability standard reviewed annually
Trauma awareness training for ≥80% of instructional and administrative staff within past 2 years; includes nervous system activation recognition and de-escalation; trauma-informed procedural modifications documented for testing, discipline, and attendance
Learning environment quality assessment completed annually against evidence-based standards; results reported to leadership and governing board; improvement plan documented

EMERGING

Physical environment standards for lighting and acoustics; temperature and air quality informal
Psychological safety surveys conducted; institutional response inconsistent
Flexibility standards for some spaces; not systemwide
Transition protocols in place; schedule governance informal
Trauma awareness training completed by ≥60% of staff; modifications informal
Environment quality assessed annually; not against published standards

DEVELOPING

Environmental awareness; no documented standards
Psychological safety valued; no formal infrastructure
Some flexible spaces; no standard
Schedule consistency valued; no formal governance
Trauma awareness training available; not required
Quality assessment informal; anecdotal

EARLY STAGE

No documented environmental standards
No psychological safety infrastructure beyond disciplinary policies
No flexibility or individualization standard
No transition or schedule governance
No trauma awareness requirement or training
No quality assessment

Open the M12 module scope →

M13

Traditional & Complementary Medicine Governance

Healthcare & Clinical

▼

Steel et al. (2025) in the WHO meeting report establish that the evidence exists; the primary global barrier to TCIM integration is the absence of governance infrastructure. Tabish (2008) documents that RCT frameworks cannot adequately evaluate individualized, relationship-based, meaning-embedded modalities. Kawakami et al. (2022) document lāʻau lapaʻau as a living practice among Native Hawaiian elders with specific governance implications for Hawaiian health systems.

Six Governance Dimensions

#	Dimension	Why This Dimension
1	Evidence Integration Framework	Steel et al. (2025) and Tabish (2008) establish that RCT frameworks cannot adequately evaluate traditional medicine; without a multi-epistemic framework, adoption decisions are made without governance.
2	Cultural Sovereignty & IP Protections	Indigenous healing practices belong to the communities that developed them; institutional incorporation without formal community agreement is cultural appropriation with governance implications.
3	Practitioner Credentialing Standards	Unverified practitioner scope of practice creates patient safety and institutional liability exposure.
4	Patient Safety Infrastructure	Herb-drug interactions, contraindications, and modality-specific risks require pre-treatment screening — the clinical safety governance layer.
5	Regulatory Compliance	Licensure requirements for CAM practitioners vary by jurisdiction; documented legal review protects both practitioners and institution.
6	Integration with Conventional Care	Without formal communication protocols between CAM and conventional care teams, integration is nominal and patient safety gaps persist.

Tier Criteria

PIONEERING

Evidence integration framework documented: explicit criteria for evaluating CAM evidence engaging traditional use evidence, observational data, and experiential knowledge alongside biomedical evidence; reviewed annually by interdisciplinary body including TCIM practitioners and community representatives
Written community agreement for each Indigenous healing practice offered: covers attribution, compensation structure, scope of use, and right of community to withdraw consent; community partner named; agreement reviewed every 3 years
Credential verification documented for all CAM practitioners: licensure, scope of practice agreement signed, training verification, supervision structure; credentials reviewed annually
Pre-treatment safety screening documented for each modality: herb-drug interaction screening, contraindication assessment; updated within 3 years; documented by licensed clinician
Legal guidance from qualified counsel on licensure and liability for each CAM modality offered; updated within 3 years or upon regulatory change
Communication protocol between CAM practitioners and conventional care team: patient consent obtained for communication; treatment records shared within documented timeframe; contraindication concerns escalated through documented pathway

EMERGING

Evidence integration framework drafted; not yet formally adopted; some criteria documented
Community agreement in place for primary Indigenous practice; not extended to all
Credential verification completed; scope of practice informal
Pre-treatment screening for primary modalities; not all covered
Legal guidance obtained; not recently updated
Communication protocol in place; not consistently followed

DEVELOPING

Evidence integration considered informally; no formal framework
Community relationships informal; no written agreement
Credentials reviewed informally; no formal verification process
Pre-treatment screening informal; practitioner-dependent
Legal guidance informal; relying on common practice
Communication informal; patient-dependent

EARLY STAGE

No evidence integration framework
No community agreements for Indigenous healing practices offered
No formal credential verification
No pre-treatment safety screening protocol
No legal guidance on CAM modality governance
No formal communication protocol

Open the M13 module scope →

Part III

Module Relationships & Dependency Map

Not all modules are equal. Some are prerequisites. Some amplify others. Understanding the relationships shapes how institutions sequence governance work.

M8 IS THE PREREQUISITE MODULE

A burned-out staff cannot implement, sustain, or give institutional attention to governance changes in any other module. If burnout is severe, other module investments will stall. Assess M8 before any other module.

M9 AND M4 ARE THE SAME DOMAIN AT DIFFERENT SCALES

M4 is the public health infrastructure layer — population-level cannabis documentation governance. M9 is the clinical encounter layer. They share evidence bases and amplify each other. M4 governance creates the conditions in which M9 governance can be sustained.

M2 AND M11 SHARE ALGORITHMIC ACCOUNTABILITY INFRASTRUCTURE

Both require demographic impact assessment, human override protocols, and post-adoption monitoring. Institutions with M2 governance in place can adapt that infrastructure for M11 with less foundational work. Assess and address these together.

M5, M7, AND M10 SHARE THE BUILT ENVIRONMENT EVIDENCE BASE

M5 is biophilic infrastructure in civic settings. M7 is urban built environment governance. M10 is evidence-based healthcare design. All three apply Ulrich's foundational evidence in different institutional contexts. Treat them as a suite.

M1 AMPLIFIES M3

Trauma-informed legal space creates the safety conditions in which harm reduction governance can function. Harm reduction services offered in legally activating environments are systematically undermined by the governance failure in M1.

M12 IS FOUNDATIONAL FOR M1

The nervous system states that educational institutions govern in learners are the states that legal institutions then receive. ACE-informed educational governance reduces the severity of the legal governance problem downstream.

Priority Sequencing for Institutions New to NSAG

Start

M8 Burnout

Institutional readiness prerequisite — staff wellness is the precondition for implementing any other governance change

Then

Highest safety-stakes module

M9 for healthcare · M2 for legal · M12 for educational — the module with the most immediate patient/client/learner safety implications

Then

Emerging risk module

M11 for institutions adopting AI · M13 for institutions expanding CAM · M3 for institutions entering psychedelic therapy

Then

Built environment modules

M5, M7, M10 — these require capital investment and longer implementation timelines; begin planning while relational and procedural governance is established

Part IV

Feedback Architecture

NSAG is a living framework. The 15 modules, six dimensions per module, and tier definitions in this document are Version 1.0. The feedback mechanism ensures the framework improves based on institutional deployment experience.

Trigger Conditions for Dimension Review

New Primary Evidence

A peer-reviewed primary source published after framework adoption that materially changes the evidence base for a dimension triggers formal review.

Deployment Failure

A documented institutional deployment in which a dimension fails to identify a governance gap that subsequently produces harm triggers formal review of that dimension.

Content Validity Review

A formal content validity review conducted every 3 years in which domain experts identify a dimension as miscalibrated triggers formal review.

Institutional Feedback Pathway

Institutions completing advisory engagements are invited to submit a 90-day governance report documenting what they changed and what changed as a result. Governance reports are reviewed by NSAG and incorporated into the framework update process. Anonymized findings are published annually in the NSAG Brief as the empirical record of institutional outcomes. This is the feedback loop that converts NSAG from a static framework into an evidence-generating practice.

Dimension Derivation Standard

The six dimensions in each module were selected by applying five criteria: (1) evidence existence — the dimension has a documented evidence base linking it to measurable institutional outcomes; (2) governance addressability — the dimension reflects an institutional design choice that can be changed through governance action; (3) measurability — the dimension can be assessed through observable, documentable institutional characteristics; (4) independence — the dimension captures something meaningfully distinct from the other five; (5) actionability — when scored as EARLY STAGE, the dimension points toward a specific governance change rather than a general aspiration.

Read the tier definitions against your own institution.

The fifteen self-serve module assessments are retired and no tier report is generated. What remains is more useful and entirely public: the tier criteria above are written so that a person holding documents — policies, floor plans, procurement files, incident logs — can place an institution on them without any tool at all.

Read the module scopes See Your Experience →

The NSAG Framework:What the tiers mean.

What this is

What this is not

Global Tier Definitions

Module-by-Module Operationalization

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Six Governance Dimensions

Tier Criteria

Module Relationships & Dependency Map

M8 IS THE PREREQUISITE MODULE

M9 AND M4 ARE THE SAME DOMAIN AT DIFFERENT SCALES

M2 AND M11 SHARE ALGORITHMIC ACCOUNTABILITY INFRASTRUCTURE

M5, M7, AND M10 SHARE THE BUILT ENVIRONMENT EVIDENCE BASE

M1 AMPLIFIES M3

M12 IS FOUNDATIONAL FOR M1

Priority Sequencing for Institutions New to NSAG

Feedback Architecture

Trigger Conditions for Dimension Review

New Primary Evidence

Deployment Failure

Content Validity Review

Institutional Feedback Pathway

Dimension Derivation Standard

Read the tier definitions against your own institution.

The NSAG Framework:
What the tiers mean.