For corporate venture teams evaluating digital health, the strongest diligence processes increasingly separate three questions that often get conflated:
- Does the technology work? (clinical validity)
- Can it be deployed and reimbursed? (regulatory and commercial viability)
- Can it scale profitably? (operational economics)
The best-performing CVCs I've seen build parallel workstreams rather than a single scorecard. Below is a framework that reflects common practice across healthcare investors.
| Dimension | Key Questions | Evidence to Seek | Red Flags |
|---|
| Clinical | Does it improve outcomes? | Peer-reviewed studies, RCTs where appropriate, real-world evidence, external validation | Internal-only studies, surrogate endpoints without justification |
| Regulatory | Can it legally operate and expand? | FDA strategy, quality systems, cybersecurity documentation, privacy compliance | Regulatory ambiguity, unclear intended use, changing product claims |
| Commercial | Will providers and payers buy it? | Contract renewals, implementation timelines, reimbursement strategy | Pilot-heavy business with little expansion |
| Technical | Can the platform scale? | Cloud architecture, uptime, integrations, security audits | Heavy manual workflows, poor interoperability |
| Financial | Is the business durable? | Gross margin trends, CAC/LTV, retention, implementation costs | High service costs disguised as software |
| Strategic | Is there defensible differentiation? | Proprietary datasets, workflow integration, network effects | Commodity monitoring hardware with little software moat |
1. Clinical evidence
This has become much more rigorous than it was 5–7 years ago.
Instead of asking "Do they have evidence?", ask four increasingly difficult questions.
A. Analytic validity
Does the device or algorithm measure what it claims?
Examples:
- Blood pressure accuracy
- ECG sensitivity/specificity
- Glucose prediction error
- Fall detection performance
Metrics
- Sensitivity
- Specificity
- PPV/NPV
- Bland-Altman analysis
- AUC
- Calibration
- Test-retest reliability
B. Clinical validity
Does the measurement correlate with disease?
Example:
Can atrial fibrillation detection actually identify AF?
Can heart failure algorithm predict decompensation?
C. Clinical utility
The highest-value question.
Does using the product actually change patient outcomes?
Examples:
- Reduced admissions
- Reduced readmissions
- HbA1c improvement
- Blood pressure control
- Medication adherence
- Mortality
- Quality of life
This is where many startups become much weaker.
D. Economic utility
Can it save money?
Metrics
- PMPM savings
- Reduced utilization
- Shorter LOS
- Fewer ED visits
- Provider productivity
- Nurse workload reduction
Health systems increasingly demand this evidence before enterprise expansion.
Evidence hierarchy
I generally weight evidence something like:
- Multi-center RCT
- Large prospective observational studies
- External validation
- Real-world evidence
- Retrospective analyses
- Internal pilot studies
- Case studies
- Testimonials
For RPM companies, large, high-quality real-world evidence can sometimes be more persuasive than a small RCT, especially when evaluating operational effectiveness.
2. Regulatory diligence
This area has become substantially more important as AI capabilities have expanded.
Questions include:
Product classification
- Medical device?
- Clinical decision support?
- Wellness product?
- Administrative software?
Misclassification is surprisingly common.
FDA pathway
Questions include:
- 510(k)?
- De Novo?
- PMA?
- Enforcement discretion?
- Software as Medical Device (SaMD)?
Evaluate:
- predicate strategy
- regulatory timeline
- post-market obligations
Quality systems
Look for evidence of:
- design controls
- CAPA
- complaint handling
- risk management
- document control
- supplier quality
A mature quality management system is a positive signal even before it is legally required.
AI governance
Increasingly important.
Evaluate:
- model drift monitoring
- dataset representativeness
- bias testing
- update process
- explainability
- human oversight
Privacy/security
Beyond compliance checklists:
- independent penetration testing
- encryption
- identity management
- audit logs
- incident response
- third-party risk
Healthcare buyers often scrutinize security as much as clinical performance.
3. Scalability assessment
The biggest failures in digital health often come from operations rather than technology.
Unit economics
Key metrics:
Gross margin
Implementation cost
Customer acquisition cost
Customer lifetime value
Contribution margin
Payback period
Net revenue retention
Clinical operations
Questions
How many patients per nurse?
How many alerts per day?
Alert fatigue?
Manual review percentage?
Clinical staffing ratios?
Escalation workflows?
Many RPM startups are actually services businesses disguised as software.
Automation ratio
A favorite diligence metric.
Measure:
Percentage of workflow requiring humans.
Examples:
Patient onboarding
Insurance verification
Device provisioning
Alert triage
Clinical documentation
Coding
Billing
The higher the automation, the more scalable the model.
Integration maturity
Assess:
- EHR integrations
- FHIR APIs
- HL7 support
- Identity management
- Single sign-on
- Workflow embedding
Implementation time is often a leading indicator of sales scalability.
Customer expansion
More informative than logo count.
Track:
- Net revenue retention
- Sites per customer
- Patients per customer
- Module expansion
- Renewal rate
- Upsell rate
Chronic disease management metrics
For diabetes:
- HbA1c reduction
- Time in range
- Hypoglycemia
- Medication adherence
For hypertension:
- BP control rate
- Mean systolic reduction
- Medication persistence
For heart failure:
- Admissions
- Readmissions
- Days at home
- Mortality
- Medication optimization
For COPD:
- Exacerbations
- Hospitalizations
- Rescue inhaler use
For CKD:
- eGFR decline
- Albuminuria
- Dialysis delay
Disease-specific outcomes should align with the intervention's intended mechanism.
Commercial diligence
One useful framework is to distinguish:
Product-market fit
- Net Promoter Score (NPS)
- Customer retention
- Expansion
- Reference customers
Reimbursement fit
- Covered CPT codes?
- Value-based care alignment?
- Self-insured employers?
- Direct employer?
- Health system budget?
Workflow fit
Ask clinicians:
"Would you still use this if reimbursement disappeared?"
This often reveals whether the product solves a genuine workflow problem or depends primarily on a temporary financial incentive.
A practical investment scorecard
Many CVCs use weighted scoring. One example:
| Category | Weight |
|---|
| Clinical evidence | 25% |
| Commercial traction | 20% |
| Regulatory readiness | 15% |
| Scalability | 15% |
| Economics | 10% |
| Team | 10% |
| Strategic fit | 5% |
Within each category, score both the current state (what has been demonstrated) and the execution risk (how difficult the next milestones are). That distinction often separates companies with similar traction but very different risk profiles.
Questions that frequently uncover hidden risk
- Which clinical outcomes have been independently replicated?
- How much of the clinical workflow is still manual?
- What percentage of revenue comes from pilots versus scaled deployments?
- What are the gross margins after accounting for clinical labor, device logistics, and support?
- How does performance vary across patient populations, care settings, and demographic groups?
- What assumptions underpin the reimbursement model, and how resilient is the business if those assumptions change?
- What implementation resources are required for each new enterprise customer, and how long is time-to-value?
- What proprietary data or workflow integration creates a durable competitive advantage beyond connected devices and dashboards?
One trend that's become increasingly important is moving beyond evaluating "software" to evaluating the care delivery model. Many remote patient monitoring and chronic disease management companies ultimately succeed or fail based less on their algorithms than on how effectively they integrate with clinical workflows, generate measurable outcomes, and maintain favorable unit economics as they grow. A rigorous diligence process should therefore assess the technology, the evidence, and the operating model as equally important components of enterprise value.
