State agencies, Medicaid authorities, managed care organizations, county systems and human services leaders are under increasing pressure to commission, fund and oversee community-based care with better intelligence. Demand is rising across HCBS, LTSS, IDD, behavioral health, aging services, complex care and supportive housing. Workforce capacity is fragile. Provider networks are uneven. Quality signals often emerge too late. Outcomes evidence is inconsistent. Traditional oversight models are no longer enough.
This article forms part of the Innovation, Pilots & Emerging Models Knowledge Hub and connects with wider guidance on Using Data for Commissioning & Oversight, Data Collection & Data Quality and Dashboard Operating Rhythm & Performance. It explores how workforce, demand, quality and outcomes intelligence could transform Medicaid commissioning, HCBS oversight, MCO performance management and community-based service design.
Data-driven commissioning is not about more dashboards; it is about earlier insight, better decisions and stronger system accountability.
Why Medicaid and HCBS commissioning need better intelligence
Community-based care systems are becoming more complex. People are living longer with higher support needs. More individuals are being supported outside institutions, hospitals and residential settings. Behavioral health, IDD, aging, housing, substance use, crisis response and physical health needs increasingly overlap. At the same time, providers are managing workforce shortages, rising costs, administrative burden and greater expectations around value, quality and person-centered outcomes.
State agencies and funders often have large amounts of data, but not always the right intelligence at the right time. Claims data may show what was billed, but not whether support improved stability. Provider reports may show activity, but not whether workforce pressure is creating risk. Quality reviews may identify concerns, but only after patterns have been present for months.
A data-driven commissioning model would help leaders answer harder questions:
- Where is demand increasing fastest, and why?
- Which provider networks are under workforce stress?
- Where are people experiencing delayed access or service gaps?
- Which supports are preventing crisis, hospitalization or institutionalization?
- Which outcomes are improving, and which are not?
- Where is quality risk emerging before harm occurs?
- Which payment models encourage stability, prevention and better outcomes?
What a data-driven state agency could look like
A data-driven state agency or Medicaid system would connect intelligence across eligibility, assessment, utilization, provider capacity, workforce, claims, quality assurance, incident reporting, complaints, care coordination, outcomes and member experience. The goal would not be to centralize every decision or automate human judgment. The goal would be to create a clearer system picture so leaders can act earlier and commission more effectively.
The model would usually include four connected intelligence layers:
- Demand intelligence: who needs support, where demand is rising, what level of acuity is emerging and which pathways are under pressure.
- Workforce intelligence: whether provider networks have enough staff, skills, supervision and continuity to meet demand safely.
- Outcomes intelligence: whether funded services are improving stability, independence, quality of life, recovery, community participation or avoided crisis.
- Quality and risk intelligence: where provider, population or pathway risk is emerging before system failure occurs.
This connects directly with Quality Assurance, Oversight & Accountability, Value-Based Payment & Outcomes-Led Design and System Integration & Multi-Agency Working.
Demand intelligence: moving beyond utilization counts
Utilization data matters, but it does not tell the whole story. A system may know how many service units were authorized or delivered, but still lack clarity about unmet need, delayed access, intensity of support, service mismatch or why people cycle between crisis, hospitalization, emergency departments, homelessness or institutional care.
Better demand intelligence would examine:
- Eligibility and referral trends by population.
- Demand by county, region, zip code or service area.
- Assessment acuity and functional need.
- Waiting lists, delayed starts and provider acceptance rates.
- Repeat crisis use, emergency department use or hospitalization.
- Transitions from hospital, nursing facility, institution, foster care or justice settings.
- Unmet need for specialist supports.
- Demand patterns by age, disability, behavioral health need, rurality or social risk.
This is especially important across Home- and Community-Based Services (HCBS), LTSS Service Models & Care Pathways and IDD Service Models & Support Pathways.
Operational example 1: predicting HCBS access pressure before crisis
A state Medicaid agency notices that HCBS utilization appears stable overall, but county-level data shows growing delays in service starts for individuals with higher personal care needs, behavioral complexity and rural access barriers. Provider complaints also indicate that agencies are declining referrals because travel time, staff shortages and acuity are making packages unsustainable.
Without integrated intelligence, the issue might be viewed as general provider resistance or isolated capacity pressure. A data-driven approach links referral demand, provider acceptance rates, workforce vacancies, geography, acuity, authorization patterns and hospitalization risk.
The analysis shows that service gaps are concentrated in rural areas where travel time, low staff availability and higher acuity combine to make traditional fee-for-service arrangements unstable. The agency works with MCOs and providers to test revised rate assumptions, regional provider capacity incentives, remote monitoring for selected needs, and enhanced care coordination for high-risk members.
The agency then tracks whether referral acceptance improves, delayed starts reduce, hospital discharge delays fall and member outcomes stabilize. This turns fragmented operational pressure into targeted system redesign.
Required fields must include:
- Referral volume and reason.
- Provider acceptance and denial patterns.
- Workforce availability by geography.
- Member acuity and risk indicators.
- Hospitalization or crisis use linked to delayed access.
Cannot proceed without:
- Validated provider capacity data.
- Clear geographic demand mapping.
- Defined intervention thresholds.
Auditable validation must confirm:
- Intervention actions agreed.
- Access metrics reviewed after implementation.
- Member impact tracked beyond service start dates.
Workforce intelligence: the missing link in commissioning
Workforce is one of the biggest constraints in HCBS, LTSS, IDD and behavioral health systems. Yet commissioning and rate-setting often treat workforce capacity as if it exists once a provider is contracted. In reality, provider network capacity depends on recruitment, retention, supervision, training, skill mix, geography, scheduling and burnout risk.
A contracted provider may have theoretical capacity, but not enough direct support professionals, home care workers, behavioral health clinicians, supervisors, care coordinators or nurses to deliver stable support. This is why workforce intelligence is essential to system planning.
Useful workforce intelligence includes:
- Vacancy rates by provider, service type and region.
- Turnover and retention trends.
- DSP and direct care workforce availability.
- Use of overtime, temporary staffing or agency cover.
- Training and competency coverage.
- Supervisor span of control.
- Continuity for high-risk members.
- Workforce pressure linked to incidents, missed visits or service gaps.
This connects with Workforce Data & Capacity Planning, Workforce Retention Analytics & Insight and Competency-Based Workforce Planning.
Using workforce data to shape provider networks
Workforce intelligence should not be used simply to monitor providers after problems occur. It should inform provider network design, rate-setting, technical assistance, workforce investments and value-based models.
For example, workforce intelligence may show that:
- High turnover is concentrated in specific service lines or regions.
- Providers cannot accept complex members because competency requirements exceed available staff skills.
- Rural access gaps are driven more by workforce supply than provider willingness.
- Supervision capacity is too thin to support quality improvement.
- Training completion is high, but practice validation is weak.
- Burnout risk is linked to crisis-heavy caseloads or poor scheduling models.
These insights can support redesigned rates, provider capacity grants, workforce innovation pilots, DSP career ladders, competency-based training investments and targeted network development.
Outcomes intelligence: funding what improves lives
A data-driven commissioning model cannot focus only on access, utilization and compliance. The deeper question is whether services are improving people’s lives, preventing avoidable escalation and supporting long-term stability. This is especially important as states, Medicaid agencies and MCOs move toward value-based purchasing and outcomes-led contracting.
Outcomes intelligence should examine whether funded services support:
- Community living and reduced institutional reliance.
- Improved stability for people with complex needs.
- Reduced avoidable hospitalization, crisis use or emergency department reliance.
- Improved quality of life, autonomy and community participation.
- Better caregiver capacity and reduced family strain.
- Successful transitions from hospital, nursing facility, residential or institutional settings.
- Improved recovery, independence or functional outcomes.
This connects with Outcomes Frameworks & Indicators, Outcomes, Value & System Sustainability and Outcomes, Quality of Life & Impact. Strong outcomes intelligence blends quantitative measures with qualitative evidence from members, families, care teams, providers and community partners.
Operational example 2: using outcomes intelligence to redesign transitions
An MCO reviews data for members transitioning from hospital to community-based support. On paper, services are authorized quickly and care coordination contacts are completed. However, the MCO identifies high rates of repeat emergency department use within 30 days for members with behavioral health needs, unstable housing and limited caregiver support.
The MCO connects claims data, care coordination notes, provider capacity information, member-reported barriers, housing instability indicators and crisis service utilization. The analysis shows that many members receive authorized services but lack practical transition fidelity: follow-up appointments are missed, medication reconciliation is inconsistent, transportation barriers remain unresolved and providers are unclear who owns escalation after discharge.
The MCO redesigns the transition model. It introduces closed-loop referral tracking, stronger post-discharge risk stratification, enhanced care coordination for high-risk members, provider escalation playbooks and a dashboard showing repeat emergency use, follow-up completion, medication reconciliation and housing support actions.
The model is then monitored through outcomes, not just activity. The MCO tracks repeat ED use, avoidable readmissions, successful connection to community services, member experience and provider follow-up completion. This turns transition oversight from process compliance into measurable system improvement.
Required fields must include:
- Transition source and discharge risk level.
- Follow-up appointment status.
- Medication reconciliation status.
- Housing, transportation and caregiver risk factors.
- Repeat crisis or ED use within defined periods.
Cannot proceed without:
- Closed-loop referral confirmation.
- Named owner for post-discharge escalation.
- Risk-stratified follow-up schedule.
Auditable validation must confirm:
- Actions completed after discharge.
- Member outcomes reviewed at 30 and 90 days.
- Learning used to improve the transition pathway.
Quality and risk intelligence: seeing early warning signs
Quality oversight is strongest when it identifies early warning signs before people experience harm, service disruption or avoidable crisis. State agencies, counties and MCOs already hold significant quality signals, but these often sit across separate reporting structures.
Useful quality and risk signals include:
- Incident reports and near misses.
- Complaints and grievances.
- Protective services or safeguarding referrals.
- Provider corrective action plans.
- Licensing or accreditation concerns.
- Medication or clinical oversight incidents.
- Repeated crisis episodes.
- Missed visits, late visits or service interruptions.
- High staff turnover or supervision gaps.
- Provider financial or operational instability.
This connects with Audit, Monitoring & Assurance Playbooks, Assurance Dashboards & Metrics and Risk Management & Controls. The goal is to understand whether small signals are isolated events or evidence of wider system risk.
Provider risk profiles and network oversight
Data-driven commissioning should support fair, proportionate provider oversight. The purpose is not to create punitive surveillance. It is to identify where providers need support, where quality risks require intervention and where contingency planning is necessary.
A provider risk profile may include:
- Quality review results.
- Incident and complaint trends.
- Corrective action history.
- Workforce stability indicators.
- Service interruption patterns.
- Member outcomes and experience.
- Utilization anomalies.
- Financial or operational risk signals.
- Responsiveness to oversight actions.
Risk profiles should be reviewed through a clear governance process with opportunities for provider dialogue, validation and improvement planning.
Operational example 3: identifying provider instability before failure
A state oversight team sees an increase in incident reports from one HCBS provider. Separately, the contracting team receives delayed provider responses to corrective action requests. Workforce data also shows rising turnover and reduced supervision capacity. Each signal alone is concerning but not conclusive.
A connected intelligence model brings these signals together. The provider risk profile shows a pattern of increasing instability across quality, workforce and contract responsiveness. The oversight team schedules a targeted review, validates the data with the provider and identifies that rapid growth, supervisor vacancies and inadequate onboarding have weakened service quality.
Rather than waiting for serious service failure, the state agency agrees a focused remediation plan. The plan includes supervisor recruitment milestones, temporary admission controls, weekly reporting on high-risk members, targeted technical assistance, and evidence of improved staff onboarding and incident review.
The agency tracks whether incidents reduce, corrective actions are completed, workforce stability improves and member experience stabilizes. If the provider cannot improve, the agency has already identified high-risk members and contingency options.
Required fields must include:
- Incident trend by type and severity.
- Corrective action status.
- Workforce turnover and supervision capacity.
- Member risk profile.
- Provider response timelines.
Cannot proceed without:
- Provider validation conversation.
- Risk-rated remediation plan.
- Contingency plan for high-risk members.
Auditable validation must confirm:
- Actions completed on time.
- Quality indicators reviewed after intervention.
- Member safety and continuity protected.
Interoperability and data exchange
Data-driven commissioning depends on the ability to connect information across systems. Medicaid agencies, MCOs, counties, providers, hospitals, behavioral health organizations, housing partners and community-based organizations often hold different pieces of the same person or system story.
This is why Interoperability & Data Exchange Workflows, Health & Social Care Interoperability Frameworks and Closed-Loop Care Coordination & Data matter so much.
Important data exchange points include:
- Hospital discharge and HCBS service start.
- Behavioral health crisis response and community follow-up.
- Care coordination and provider task completion.
- Housing support and healthcare utilization.
- Protective services and provider risk management.
- Workforce capacity and network adequacy.
- Member outcomes and service authorizations.
Interoperability is not only a technical issue. It is an operational governance issue. Partners need shared definitions, consent workflows, minimum necessary access controls, data quality standards and clear responsibility for acting on information.
Data governance, privacy and trust
Data-driven commissioning must be built on trust. Medicaid and human services data often includes sensitive information about disability, health, behavioral health, substance use, housing instability, protective services, legal status, family circumstances and support needs. Better intelligence should improve equity, access and accountability — not create opaque decisions or inappropriate data use.
Strong data governance should include:
- Clear data ownership and stewardship.
- Defined data quality standards.
- Information-sharing agreements.
- Consent and authorization workflows where required.
- Minimum necessary access controls.
- Audit trails showing who used data and why.
- Transparency about how analytics inform decisions.
- Human review for high-impact decisions.
This connects with Data Governance & Information Accountability, Privacy-by-Design & Risk Mitigation Practices and Trust, Transparency & Ethical Data Use.
AI and predictive analytics: promise and caution
AI and predictive analytics could support future Medicaid and HCBS commissioning, but only if used carefully. Predictive tools may help identify emerging workforce gaps, crisis risk, provider instability, hospitalization risk, service access barriers or outcomes variation. However, they should support decision-making, not replace accountable human judgment.
Appropriate uses may include:
- Identifying likely provider network gaps.
- Flagging members at risk of repeat crisis or hospitalization.
- Forecasting workforce pressure by service type.
- Identifying where interventions are not producing expected outcomes.
- Detecting unusual patterns in incidents, claims or grievances.
- Supporting targeted quality review or technical assistance.
This links with AI & Automation in Care, Technology-Enabled Care and Scaling What Works. The key governance question is not whether analytics are powerful, but whether they are explainable, equitable, validated and used responsibly.
Equity intelligence: seeing who is not being reached
A data-driven system should improve equity. Demand and outcomes intelligence can reveal where people experience unequal access, delayed support, poorer outcomes or higher crisis use.
Equity intelligence may examine:
- Access by race, ethnicity, language, disability, geography and income.
- Rural and underserved community capacity.
- Digital exclusion and telehealth access barriers.
- Disparities in service authorization, provider availability or wait times.
- Differences in outcomes across populations.
- Whether culturally responsive providers are available.
- Whether people with complex needs are more likely to experience service denial or delay.
This connects with Data-Led Equity Planning, Rural & Underserved Communities and Health Equity & Disparities Impact.
Value-based payment and outcomes-led commissioning
Data-driven commissioning creates the foundation for value-based payment. Without reliable data, value-based models can become unfair, unclear or overly focused on narrow metrics. With good intelligence, funders can design payment models that reward stability, prevention, access, outcomes and quality.
Value-based commissioning may use intelligence around:
- Avoided hospitalization or institutionalization.
- Improved community tenure.
- Successful transitions and reduced bounce-back.
- Improved member-reported outcomes.
- Provider quality and continuity.
- Reduced crisis recurrence.
- Improved workforce stability.
- Reduction in avoidable high-cost utilization.
This connects with Outcome Commissioning & Pay for Performance, Cost vs Outcomes and System Capacity & Flow Impact.
Dashboard operating rhythm: insight only matters if someone acts
Dashboards do not transform systems on their own. A dashboard becomes useful only when there is a clear operating rhythm around it: who reviews the data, how often, what thresholds trigger action, who owns follow-up and how decisions are documented.
A commissioning intelligence dashboard may include:
- Demand and referral pressure.
- Provider network capacity.
- Workforce risk indicators.
- Waiting lists and delayed service starts.
- Quality and incident trends.
- Member outcomes and experience.
- High-cost utilization and avoidable use.
- Equity and access indicators.
- Corrective action and remediation status.
The operating rhythm should define:
- Weekly operational review.
- Monthly provider network review.
- Quarterly outcomes and value review.
- Escalation thresholds for urgent concern.
- Governance routes for unresolved system risks.
From pilots to system redesign
Many states and MCOs already run innovation pilots. The challenge is scaling what works. Data-driven commissioning helps leaders understand which pilots produce measurable outcomes, which populations benefit most, what operational conditions are required and whether the model is financially sustainable.
Useful pilot intelligence includes:
- Who the pilot served.
- Baseline risk and support needs.
- Service model fidelity.
- Workforce requirements.
- Cost and utilization impact.
- Member experience.
- Equity impact.
- Replicability across regions or populations.
This connects with Pilot Evaluation & Learning Loops, New Service Models and Integrated Funding Pilots.
Common pitfalls to avoid
- Building dashboards without decision rights or action thresholds.
- Collecting more data than teams can interpret.
- Using claims data as a proxy for outcomes without context.
- Separating workforce, demand, quality and outcomes intelligence.
- Using provider data punitively without validation or dialogue.
- Failing to include member and family experience.
- Ignoring equity, rurality and access barriers.
- Using predictive analytics without explainability or human oversight.
- Not linking intelligence to rates, procurement, network design and quality improvement.
How to evidence data-driven commissioning in strategies and procurement
State agencies, Medicaid authorities, MCOs and county systems can use data-driven commissioning evidence within strategic plans, waiver redesign, procurement specifications, provider manuals, quality strategies, value-based payment models and innovation roadmaps.
Useful evidence includes:
- Demand and capacity modelling by population and geography.
- Workforce risk dashboards.
- Provider network adequacy analysis.
- Outcomes frameworks and measure libraries.
- Quality and incident trend analysis.
- Equity and access review.
- Provider risk profiles and remediation pathways.
- Pilot evaluation and scaling evidence.
- Decision logs showing how intelligence informed system change.
This evidence can also strengthen procurement and contract operations. Specifications can require better outcomes reporting, stronger data quality, clearer provider capacity information, interoperable workflows and stronger quality improvement evidence.
What providers should understand
Providers should expect funders, MCOs and state agencies to ask more sophisticated questions about data, workforce, outcomes and quality. This does not mean every provider needs advanced analytics immediately. It does mean providers need reliable evidence and a clear operating rhythm around performance.
Providers should be ready to evidence:
- Workforce capacity and continuity.
- Staff competence and supervision.
- Service delivery reliability.
- Member outcomes and quality of life.
- Incident learning and quality improvement.
- Care coordination and referral follow-up.
- Equity, access and cultural responsiveness.
- How performance data changes practice.
This connects with “Evidence Packs” for Funders & Regulators, Translating Practice into Evidence and Contract Management & Provider Performance.
Leadership requirements for data-driven commissioning
Data-driven commissioning requires leadership, not only technical capability. Senior leaders must set expectations about how intelligence is used, how decisions are made, how risks are escalated and how improvement is tracked.
Leadership requirements include:
- Clear ownership of commissioning intelligence.
- Shared definitions across teams and partners.
- Governance forums that act on data.
- Investment in data quality and analytics capability.
- Provider engagement and trust-building.
- Ethical oversight of AI and predictive tools.
- Alignment between finance, quality, operations and outcomes.
This links with Executive Leadership & Strategic Oversight, System Leadership & Cross-Sector Governance and Organisational Culture & Learning Systems.
The future: commissioning as a learning system
The most advanced data-driven commissioning models will operate as learning systems. They will not only monitor what happened. They will identify where practice, funding, workforce, quality and outcomes need to change.
A learning system approach means:
- Data is reviewed regularly and acted upon.
- Member experience is treated as evidence.
- Providers are engaged in improvement rather than only compliance.
- Workforce pressure is understood as system risk.
- Payment models are adjusted when they create unintended consequences.
- Innovation pilots are evaluated and scaled based on evidence.
- Equity and access are monitored continuously.
In this model, commissioning becomes less reactive and more adaptive. The system learns from demand, outcomes, quality signals and workforce pressure, then adjusts service models accordingly.
Conclusion
The data-driven state agency is not defined by technology alone. It is defined by the ability to turn workforce, demand, quality and outcomes intelligence into better commissioning, funding, oversight and system design decisions.
Medicaid, HCBS, LTSS, IDD, behavioral health and human services systems will always require human judgment, partnership and ethical decision-making. Better data does not replace these. It strengthens them.
By connecting intelligence across providers, populations, pathways and outcomes, state agencies and funders can identify risk earlier, support provider networks more effectively, improve equity, commission better services and scale models that genuinely improve people’s lives.