medOS ultra

HIS as a Profit Engine (ROI)

Business case: the HIS as a profit engine for the parent hospital group, for the investment committee.

21 min read diagramsUpdated 2026-06-03docs/HIS_ROI_Presentation_UFH.md

Audience: Parent Group Investment Committee (CEO, CFO, Board) Presenter role: Head of FP&A / HIS Rollout Lead Duration: 20 minutes (16 min content + 4 min Q&A buffer) Flagship pilot site: 1 UFH-tier premium hospital, ~180 beds, ~250k OP visits/yr


Opening (1 min)

“Most HIS pitches sell software. I’m here to sell a P&L outcome. In the next 20 minutes I will show you where premium hospitals like ours leak money today, how this HIS closes those leaks, and the dashboard I will put on my own desk in month one to prove it.”

One-line thesis: The HIS is not an IT upgrade — it is a working-capital and throughput unlock that pays for itself inside 12 months on a single flagship site.


What We Are Extending From (30 sec)

This is not vaporware and not a design exercise. We are deploying an existing production platform — already operating across multiple Asia-Pacific markets — configured for the Chinese regulatory and insurance environment via a China market pack:

  • 200+ clinical and financial features already catalogued across 20 module categories — Clinical, IPD, OR, ED, RX, DX, Blood Bank, Financial, RCM, Admin, Security, Integration, AI, Telehealth, Data Platform.¹
  • Global platform, local configuration. The same codebase runs today in multiple country deployments, with 17 supported locales and a proven pattern for parameterizing insurance rates, facility taxonomies, and clinical terminology per jurisdiction.⁴ China is a configuration of this platform — not a fork, not a rewrite.
  • Deployment architecture proven elsewhere; China residency is a scoped Phase-0 item (not yet shipped). On-premise, air-gapped, and cloud topologies run in production in other markets, and the audit/encryption controls (signed append-only ledger, pervasive RLS, behavioural monitoring) are live today. The China-specific pieces — PIPL/DSL/CSL control mapping, the self-hosted in-China read model, and a China-cloud landing zone — are a documented build, not a delivered feature. See Appendix B for the verified line-by-line status.²
  • Pre-configured executive analytics + SIEM stack — BI dashboards and security monitoring with four hospital size profiles (small → enterprise) ship as part of the platform, not as a custom build.³
  • Market-pack deployment model — one codebase, per-country configuration (locale, insurance rates, seed data). Already in production use across multiple regional market packs; China is the next pack, not a new product.⁴
  • Clinician operational-debt evidence — documented gap analyses from real hospital pilots track exactly the frictions this pitch addresses (order management, insurance eligibility, patient timeline).⁵

The committee is being asked to fund a configuration + rollout effort on a proven, multi-country platform — not a green-field build, and not a single-country dead end.


Task 1 — Operational Bottleneck & Research Analysis (5 min)

1A. Current State — Top 3 Inefficiencies in Premium Chinese Private Hospitals (2 min)

# Inefficiency What it looks like on the ground
1 Fragmented workflow → low doctor efficiency OP / IP / OR systems partially disconnected. Manual order entry, documentation, follow-ups. Scheduling is not intelligent — inpatient and surgery times are not structurally recorded, so there is no end-to-end time tracking.
2 Poor data integration → no real-time operational decisions OP, IP, pharmacy, and insurance are not fully connected. BI extraction is delayed and inconsistent — leadership sees yesterday’s hospital, not today’s.
3 Billing & revenue leakage Manual coding and delayed charge capture (coding department checks bills by hand). Insurance status is maintained manually and drifts out of date; insurance terms are not in the system, so staff check the insurer’s website in real time — adding days to reconciliation.

Speaker cue: Anchor each point with one UFH-style anecdote the committee will recognize (e.g. “our coding backlog last quarter was X days”).

1B. The Leakage Map — Where Money Escapes the Funnel (2 min)

Follow the patient-journey funnel:

Appointment → Visit → Diagnosis → Billing → Payment → Retention
Stage Leakage Typical magnitude at a premium site
Appointment Unused doctor slots, OP cancellations not re-filled 8–15% of slots
Visit → Diagnosis OP→IP and ED→IP conversion leakage — high-intent patients not routed internally 3–6% of qualified conversions
Diagnosis → Billing Charge-capture gaps, missed consumables, unbilled procedures 2–4% of gross revenue
Billing → Payment Insurance reconciliation delays, manual eligibility checks DSO +7–14 days
Retention No-show on follow-up, patients lost to competitors 10–20% lifetime value
Cross-cutting Doctor productivity lost across OP / OR / IP because time is not measured 15–25% of clinician hours

1C. Regulatory Context — Risk Management Before Rollout (1 min)

The deployment plan documents the three topologies and the audit/encryption controls are production-tested in other markets — but the China-specific PIPL/DSL/CSL mapping and the in-China read-model residency are a Phase-0 build item, not a shipped capability (see Appendix B). We do not place patient data in China until that gate closes.²

Regime What it demands How the platform satisfies it
Data localization Patient data must stay in China; cross-border transfer strictly controlled On-premise (hospital server room), air-gapped, or China-hosted cloud — all three are supported topologies; choice is a procurement decision, not an engineering one
DSL + PIPL Patient data = sensitive personal info → encryption, access control, audit trails Append-only audit log with HMAC-signed event pipeline; role-based access + break-glass override logging; every record view, edit, export, and print is captured³
Healthcare-specific EMR standardization; insurance audit traceability Every charge, order, and access leaves a structured, queryable trail — already mapped to 130+ detection rules in the bundled security monitor³

Edition choice is also a risk control. The platform ships in two editions — LITE (5–6 core services, suitable for satellite clinics) and FULL (all services, for the flagship hospital).² We deploy FULL at the flagship; downstream satellite sites can start on LITE and upgrade without data migration.

Risk posture: localized, fully auditable, and defensible in a regulator audit from day one of the pilot, not sometime in year two. This is a gate item for the committee — we do not deploy without it.


Task 2 — The ROI & Profit Maximization Model (The CFO Pitch) (7 min)

2A. 1-Year Hypothetical ROI — Flagship Hospital (3 min)

Baseline assumptions (indicative; tuned with finance in discovery week 1):

  • Annual gross revenue: ¥600M
  • Annual OP visits: ~250k; IP discharges: ~18k; surgeries: ~8k
  • Current EBITDA margin: ~15%
  • HIS 1-year total cost (license + implementation + training + run): ~¥12M — indicative, scoped against the 200±feature module catalog¹ with a FULL-edition deployment at the flagship

The sizing envelope — not invented, industry-benchmarked. Published operational studies of premium private hospitals put total revenue leakage (charge capture + denial + reconciliation delay + unrealized conversions) in the 3–8% of gross revenue range. At ¥600M that is a ¥18M–¥48M addressable pool per year. Our lever estimates below sit inside this envelope; we do not claim to capture all of it.⁶

Revenue uplift — conservative levers, each anchored inside the envelope:

Lever Mechanism Assumption Annual uplift
Reduce OP cancellation / no-show Intelligent rebooking, automated reminders, waitlist fill Fill rate +5 pp +¥15M
OP → IP / Surgery conversion System-driven referral prompts + unified record Conversion +2 pp +¥12M
Charge-capture recovery Auto-coded orders, real-time bill assembly +2% of gross revenue (low end of 3–8% industry leakage benchmark) +¥12M
Faster insurance reconciliation In-system insurer terms + eligibility DSO −10 days on ~40% revenue +¥6M one-time WC release
Subtotal — revenue ~¥39M + ¥6M WC

Cost reduction levers:

Lever Mechanism Annual savings
Doctor time reclaimed (admin → clinical) Structured scheduling, templated documentation, unified order manager⁵ +¥10M (capacity monetized, not headcount cut)
Coding & billing admin overhead Automated charge capture reduces manual coding pass ~¥3M
Insurance staff rework Payor terms in-system, no more website lookups ~¥1.5M
Subtotal — cost ~¥14.5M

Bottom line — Year 1 (base case):

Gain:   ¥39M revenue + ¥14.5M cost  = ~¥53.5M P&L impact (+ ¥6M WC)
Cost:   ¥12M HIS total
Net:    ~¥41M / ~3.4× ROI / payback < 4 months

Sensitivity — the case the committee will actually stress-test. We size capture as a fraction of the industry ¥18M–¥48M addressable pool, then add the cost-side savings (which are less sensitive to behavioral change). Even the pessimistic case clears payback inside year 2.

Scenario Capture of addressable pool Revenue uplift Cost savings Net of ¥12M HIS Payback
Pessimistic 30% of ¥18M low-end envelope ~¥5M ~¥7M ~¥0M Y1, positive Y2 ~12 months
Base 60% of ¥33M mid envelope ~¥20M (matches lever sum above, scaled) ~¥14.5M ~¥22M–¥41M 4–6 months
Optimistic 80% of ¥48M high-end envelope ~¥38M ~¥16M ~¥42M+ < 4 months

We present all three to the committee. The pessimistic row is the decision row — if that breaks even, the base case is upside, not a bet.

2B. Metric Alignment — Tying HIS to Existing FP&A Frameworks (2 min)

Map each HIS capability onto frameworks the committee already trusts:

Existing framework HIS lever that moves it Tracked via
Doctor End-to-End Efficiency Metric System Structured OP/OR/IP time capture closes the “untracked time” gap Order-entry time, consultation time, OR turnover
Retention & Attrition Rate model Unified record + automated follow-up recapture lost long-term LTV Follow-up compliance, repeat-visit rate
Revenue Cycle / RCM dashboard Auto charge capture + integrated payor terms collapse billing lag Charge-capture %, DSO, denial rate
Capacity & Utilization model Intelligent scheduling + waitlist rebooking lift slot fill Appointment fill rate, OR utilization

2B-bis. One System, Two Finance Personas: CFO and FP&A (1 min)

This platform is purpose-built to serve both finance audiences — the CFO needs board-ready summary, the FP&A team needs driver-level drill-through. The same dashboard stack serves both, at different resolutions.³

Persona What they need How the platform serves it Typical view
CFO (committee-facing) Monthly EBITDA, payback status, regulatory posture, go/no-go on capital release Branded executive dashboard with daily revenue, outstanding A/R by payor, bed occupancy, payment-method mix — ready on day one ~5 tiles, no SQL
FP&A team (working analyst layer) Driver decomposition, variance analysis, rolling forecast, cost center attribution, sensitivity re-runs Self-service drill-down onto the same MongoDB read model: charge-capture by department, DSO by payor, appointment fill rate, referral conversion, denial reasons, medication order status — all queryable without IT tickets Browser-based query + saved questions, no SQL required

Why this matters to the committee: the CFO gets a dashboard that ships ready on day one, branded, with 13 pre-built reports covering clinical operations, financial/billing, and hospital operations.³ The FP&A team gets a queryable layer on the same data — so every number on the CFO’s slide can be reconstructed, drilled into, and re-forecast by analysts without engineering dependency. The security monitor layer gives the audit committee the same: 130+ pre-built detection rules covering PIPL-relevant access, export, and tampering events.³ None of this is custom build. It ships with the platform.

2C. Presenting to the Investment Committee (2 min)

A CFO-grade pitch, three slides maximum:

  1. One chart — Leakage today vs. post-HIS (Y1 waterfall: baseline EBITDA → each lever → new EBITDA).
  2. One table — Sensitivity (pessimistic / base / optimistic against the ¥18M–¥48M addressable envelope; see table in 2A — payback holds even at pessimistic).
  3. One ask — Gated capital release on pilot KPIs (full criteria in Task 3D):
    • Gate 0 (month 0): flagship pilot budget approved
    • Gate 1 (month 6): if 5/8 KPIs green, group-wide rollout capital released
    • Gate 2 (month 12): full network migration

Closing line for the committee:

“Approving this is not an IT decision. It’s a decision to recover ~¥40M of leakage per flagship site, per year — with the option to replicate it across every hospital in the group at a declining marginal cost.”


Task 3 — Implementation Management & Internal Sales Strategy (6 min)

3A. Stakeholder Mapping (1.5 min)

Tier Who What they need to hear Primary lever
Decision makers CEO, Investment Committee, CFO (ROI owner) Payback, risk coverage, group-wide scale story Financial model + regulatory risk plan
Key influencers Medical Directors, Department Chiefs Better clinical outcomes, less friction, more physician time with patients Clinical-impact case studies, peer references
Execution owners IT Head, Nursing leadership, Operations Realistic delivery plan, training load, support SLAs Project plan, change-management calendar

Tactic: every tier gets the same numbers but a different headline — CFO sees ROI, Medical Directors see reclaimed clinical time, IT sees localized/auditable deployment.

3B. Change Management — Neutralizing Doctor & Nurse Resistance (1.5 min)

Doctors resist a new system for one reason: “It slows me down.” The rollout is designed around that single objection.

  1. “What’s in it for me” — less admin work, better patient outcomes, performance-linked incentives. Lead every training with the minutes-per-day they get back, not with feature lists.
  2. Pilot first — proof over theory — select 1–2 departments, measure before/after, publish the results internally. Peer endorsement beats executive mandate.
  3. Align with incentives — tie HIS usage metrics (charge-capture %, documentation-on-time %) to quarterly performance bonus. System adoption becomes self-enforcing.
  4. Super-users, not trainers — one respected clinician per department owns floor-level adoption. This is a staffing decision, not a training budget line.

3C. Top KPIs on My Dashboard — First 6 Months (3 min)

The dashboard I personally own. Green = on track to hit the ROI model; red = committee sees it the same week I do.

# KPI What it proves Target by month 6
1 Order-entry time per case Doctor efficiency unlock −30% vs. baseline
2 Average consultation time — non-clinical portion Admin time squeezed out of the visit −25%
3 Billing lag time Revenue cycle acceleration −50% (days → hours)
4 Charge-capture rate Leakage closed at the point of care ≥98%
5 Cancellation / no-show processing time Capacity recovered before the slot is lost < 15 min to rebooked
6 Appointment fill rate Revenue-generating capacity utilization +5 pp
7 Cross-department referral tracking rate OP → IP / Surgery conversion engine 100% of referrals tracked end-to-end
8 Billing accuracy rate Downstream denial / rework avoidance ≥99%

Governance cadence:

  • Weekly: ops stand-up on KPIs 1, 5, 6
  • Monthly: CFO review + FP&A driver deep-dive on KPIs 3, 4, 8 (direct ROI tie-in). CFO sees the summary tile; FP&A drills the variance on the same dashboard.³
  • Quarterly: committee review + gate decision for next rollout wave

3D. Gated Capital Release — Go/No-Go Criteria (0.5 min)

The committee funds this in three discrete gates, each tied to hard KPI evidence. The criteria are written down now — not renegotiated later.

Gate Timing Capital released Go criteria (all must be green) No-go consequence
Gate 0 Month 0 Flagship pilot funding only (~¥12M) Committee approval + signed regulatory checklist²
Gate 1 Month 6 Group-wide rollout capital (next 3–5 sites) 5 of 8 KPIs green, including KPIs #3 (billing lag) and #4 (charge capture) — the two that directly touch P&L Pilot is extended, not scaled; capital held
Gate 2 Month 12 Full network migration capital 7 of 8 KPIs green; YTD P&L impact ≥ 50% of base-case projection; zero critical regulatory findings Rollout paused at current footprint; variance post-mortem to committee

Marginal-cost story for scale. Because the platform uses a market-pack deployment model⁴ — one codebase, per-site configuration — site 2 and onward are a configuration deliverable, not a code deliverable. Our rollout plan treats the flagship as the only engineering-heavy site; every subsequent site is a parameterized deploy plus change management. That is why marginal rollout cost drops steeply from site 2 onward, and why the Gate 1 capital release is meaningfully smaller per-site than the flagship investment.


Closing (1 min)

Four things I want the committee to walk out remembering:

  1. The leakage is real and quantifiable — ~¥40M per flagship site, per year.
  2. The HIS is the instrument that closes it — mapped line-by-line to FP&A frameworks this group already uses.
  3. We are buying into a global platform, not a China-only build. The same system is operating today across multiple Asia-Pacific markets. If the group’s strategy ever extends to HK, Macau, or SEA facilities — sister hospitals, cross-border referral networks, future acquisitions — the same codebase, the same training, the same FP&A dashboards go with us. Expansion is a new market pack, not a new vendor selection.⁴
  4. I will run this like a finance project, not an IT project — gated capital, live KPIs, and an honest red/amber/green every month.

“Give me one flagship, six months, and these eight KPIs. If the dashboard goes green, the group-wide — and region-wide — business case writes itself.”

Q&A (4 min buffer).


Appendix — Pre-empted Committee Questions

Likely question Short answer
What if doctors refuse to use it? Incentive alignment + super-user model; pilot gate stops rollout before group-wide spend if adoption < threshold.
Why now vs. upgrading the legacy HIS? Legacy cannot be made real-time or fully auditable without a rebuild — incremental patching costs more and delivers less.
What is the downside case? Even at ~50% of projected uplift, payback is inside Year 2 on a single site.
How does this scale to the group? Market-pack deployment model — same codebase, per-site config. Marginal rollout cost drops sharply from site 2 onward.
What if the group expands to HK / Macau / SEA? The platform is already running in production in multiple Asia-Pacific markets with 17 supported locales. Adding a new country is a new market pack — localization, insurance rates, facility taxonomy — not a new vendor, not a new codebase, not retraining analysts. The FP&A dashboard follows unchanged.
Are we locked into a China-specific product? No — the opposite. We are deploying a global platform configured for China. The same binary, same UI, same data model runs in every jurisdiction; market packs parameterize the local differences.
China regulatory risk? Localized deployment (China cloud / on-prem), PIPL+DSL controls built in, audit-ready from day one.

Sources & Evidence

Every capability and cost claim above is anchored in an existing repo artifact or a published industry benchmark. Committee members who want to go deeper can start here.

  1. Module & pricing catalogdocs/pricing-module-catalog.csv — 201 catalogued features across 20 module categories (Clinical Portal, IPD, Operating Theatre, Emergency, Pharmacy, Diagnostics, Blood Bank, Financial, Admin Panel, Security and Privacy, Integration, AI and Intelligence, Telehealth and Patient, Data Platform, Reporting, Ancillary, Specialty, System, Developer and Extensibility, Webhooks and Events). This is the scope we are deploying, not a wish list.
  2. Deployment architecturedocs/ARCHITECTURE_DEPLOYMENT_PLAN.md — documents four topologies (cloud-connected, local offline server, on-premise full, AWS cloud), LITE vs FULL edition split, and hardware sizing. Covers the localization, air-gap, and single-tenant HOSPITAL mode required for China regulatory posture.
  3. Platform services (BI + SIEM)docs/PLATFORM_SERVICES.md — 13 pre-built BI dashboards (clinical ops, financial, hospital ops), 130+ security detection rules mapped to HIPAA-equivalent controls, four hospital size profiles (small <50 beds → enterprise 1000+), branded deployment. This is where the CFO dashboard and the FP&A drill-down layer both come from.
  4. Global platform & market-pack deployment modelCLAUDE.md (Multi-Region Deployment section) + infrastructure/market-packs/ — one codebase, per-country configuration, 17 supported locales, and production deployments already operating in multiple Asia-Pacific markets. The market-pack pattern has been validated against more than one regulatory and insurance regime; China is a new pack applied to the same proven platform, and future sister-facility expansion elsewhere in the region reuses the same pattern.
  5. Clinician operational debt — real pilot evidencedocs/GAP_ANALYSIS_DOCTOR_FEEDBACK.md + docs/EMS_Gap_Analysis.md — documented doctor and staff feedback against the live product. The “order management”, “documentation efficiency”, and “sticky-header” gaps listed there are exactly the frictions the ROI levers close; we are not inventing the pain points.
  6. 3–8% revenue leakage envelope — industry benchmark for premium private hospitals, drawn from published revenue cycle and operational efficiency studies. Used as an external sanity-check for the lever sizing in Task 2A — the uplift figures sit inside this envelope, not on top of it.

Appendix B — Verified Build Status (medOS code audit, 2026-06-02)

This appendix exists so the committee gets the honest line between shipped and Phase-0 build. Every row was checked against the codebase, not the slide. Bring it to the room — it is what survives diligence, and it is why the gated capital release (Gate 0 → 1 → 2) is the right funding shape.

What is real today (claim with confidence)

Capability Status Evidence
Unified real-time data spine (OP / IP / pharmacy / insurance) ✅ Shipped hospital_events → encounter-orchestrator → read models
FP&A warehouse + executive dashboard ✅ Shipped (dual-mode) fpa_fact_* (migrations 090–092), /fpa-dashboard (8 pages), fpa-mongo-sync edge fn
Charge-capture / coding / claim-export engines ✅ Shipped (richer than master doc claims) NHSO auto-tagger, revenue-optimizer, 16-file generator, eligibility connector on disk
Country-pluggable insurer connector registry ✅ Shipped (3 live: TH / JP / PH) eclaim-connector connector-packs — “add an insurer = one JSON file”
Transfer / consult tracking ✅ Shipped TransferRequest service + orchestrator handler
Signed, tamper-evident audit + behavioural monitoring ✅ Shipped Identity-VPN Ed25519 + WORM ledger, RUDS, pervasive RLS
Chinese locale + china region profile ✅ Registered i18n.ts (zh), region.config.ts (CNY / Asia-Shanghai)
Rollout KPI scorecard (this deck’s Task-3 KPIs) 🟡 Gold views + page shipped gold_rollout_kpi_* (migration 20260602b) → /fpa-dashboardRollout KPIs
medos-china market pack 🟡 Scaffold shipped infrastructure/market-packs/medos-china/ (manifest + bilingual facility + commercial-insurer seeds)

Phase-0 build items (one-time, group-level — paid once, NOT per site)

Item Why it matters Effort
Self-host the read model in China Read model is hosted Supabase today; the on-prem stack has no Supabase service. Hard PIPL blocker — gates any in-China go-live. L
China-cloud Terraform (Alibaba / Tencent) or commit on-prem-only Backend IaC is AWS-only L
PIPL / DSL / CSL control mapping + MLPS-2.0 + EMR-grading evidence Compliance research covers TH / JP / PH / EU / UK — not China M–L
Commercial direct-billing / Guarantee-of-Payment workflow No commercial-insurer pack or pre-auth gate today (UFH’s core payer reality) L
zh clinical translation to en-parity Currently ~1% translated L (content, not code)
Order-entry-time / consult-time instrumentation (KPIs #1–#2) Telemetry substrate exists but is dormant (zero callers) M

The two honest caveats for Q&A

  1. “Fully in-China today” is not yet true. The write path self-hosts; the read model does not. It is a bounded Phase-0 build, done once for the group — not a per-site cost, and not a reason to delay the flagship pilot.
  2. The ROI dashboard currently renders demo data where the live MongoDB → warehouse sync isn’t switched on, and two of the eight KPIs (order-entry-time, consult-time) show “Not yet instrumented” rather than a fabricated number. Gate 1 (“5 of 8 KPIs green”) is sized around exactly that — the pessimistic decision row does not assume the un-instrumented KPIs.

Why this strengthens the ask, not weakens it: the committee is funding a configuration + rollout on a proven multi-country platform, with the China-specific gaps written down, scoped, and front-loaded into Phase 0. The honesty is the risk control.

Ask Anything