CCTV design for hospitals: privacy zoning, retention windows and the camera count nobody calculates

Hospital surveillance is one of the briefs new ELV integrators most consistently misread. The instinct is to apply the commercial CCTV playbook — high camera density, broad coverage, generic analytics, 30-day retention — and stick a hospital sticker on the design. The result is a system that the clinical-engineering lead refuses to commission, and an integrator who has burned a relationship in week three.

The right starting point is the four constraints a hospital imposes that a commercial site does not: patient privacy zoning, infection-control cabling pathways, NABH-aligned retention windows, and clinical-engineering veto rights over every analytics rule. Once the design respects all four, the camera count, the storage sizing, the analytics schedule and the cabling pathway all fall into place. Get any one wrong and the whole plan is non-starter.

## Patient privacy zoning is the design floor, not a software setting

Cameras inside patient wards, treatment rooms and consultation rooms are not acceptable in almost every Indian hospital we engineer. The privacy expectation is part of the doctor-patient relationship and is enforced by the medical council's ethical framework as much as by any data-protection law. We design with cameras explicitly excluded from these zones — physical exclusion at the camera-placement drawing, not software masking — and we cover the surrounding corridors, lift lobbies, nurse stations, drug stores, pharmacy, OT entry/exit, OPD waiting halls and perimeter at the densities the operational brief justifies.

Where coverage of a ward is required for patient-safety reasons (high-dependency units, infectious-disease wards under quarantine), the brief is captured in writing, signed by the clinical-engineering lead and the medical superintendent, and the camera is specified with a hardware shutter that can be physically verified — not a firmware-only privacy mask that a future firmware update can disable. The discipline is to make the privacy promise visible to the patient and verifiable to the auditor.

## Camera count is lower than commercial, not higher

Most commercial integrators arrive at hospital briefs with a high-density camera plan and find themselves cutting cameras through the entire design phase. The right starting point is the opposite: 12–18 cameras per 1,500 sq m of clinical floor (against 18–28 for an equivalent commercial floor), with the headcount allocated to the corridors, lobbies and high-value zones (drug stores, pharmacy, OT corridors, records rooms, blood-bank entry) rather than spread thinly across the entire footprint.

The economics of a serious analytics layer also push toward fewer, better cameras: 4MP minimum at the corridor heads, varifocal at the lobby junctions, panoramic at the OT corridors, fixed at the drug-store entry. We specify per-zone, against the architect's drawings, with a coverage map the clinical-engineering team can review against their actual workflow. The result is a CCTV layer that surfaces the events the hospital actually needs to forensically reconstruct, not a wallpaper of low-resolution evidence that nobody has the storage to keep at usable quality.

## Retention windows are zone-specific, not building-wide

Commercial CCTV retention is typically a single number: 30 days, sometimes 60. Hospital retention is zone-specific because the audit and incident-investigation timelines vary by clinical context. We typically specify 90 days for clinical zones (NABH-aligned for incident-investigation timelines), 180 days for pharmacy and drug stores (controlled-substance audit windows), 30 days for admin and outpatient billing zones, and 365 days for records-room entry and blood-bank cold-chain.

Storage is sized by the formula retention × bitrate × camera count, summed across zones. A 200-bed teaching hospital we delivered ran to roughly 480 TB of usable storage at the formula, with 1.2× redundancy. The clinical-engineering lead saw the calculation before procurement; the storage was specified to it. We do not let storage become the silent compromise that turns a 4MP camera plan into a 1080p commissioning result.

## Cabling routes respect infection-control zoning, not just architectural drawings

The cabling pathway between every camera and the recording infrastructure has to respect the hospital's pressure-zoning, the cleaning protocols and the fire-and-smoke compartmentation. Routing CCTV cabling from a camera in a positive-pressure OT corridor through a negative-pressure isolation room is not a cabling decision — it is a clinical-compliance decision that can void NABH accreditation. We document the infection-control zoning before the cabling design begins, and where a cable crosses a compartmentation boundary, the penetration is sealed to NABH-and-NBC standard with a written penetration register.

This is the part of the brief that takes the most time and surprises the most integrators. The cabling drawing for a hospital CCTV deployment routinely runs 1.5–2× the page count of an equivalent commercial deployment, because the routes are constrained at every compartmentation boundary. We staff the design phase accordingly; the hospital is not a commercial tower with extra labels.

## Analytics rules are reviewed by clinical engineering, not configured by IT

The clinical-engineering team has a veto over every analytics rule we propose, and we hand the configuration over for review before commissioning. Loitering analytics in OT corridors are usually rejected (medical staff stand and confer there for legitimate clinical reasons); intrusion analytics on the drug-store door are usually accepted; line-crossing on the pharmacy entry is usually configured with a window-of-acceptance for inventory deliveries.

We script the analytics rules, walk them through with clinical engineering, modify them per the meeting, and only then push them to the VMS. The discipline is that the clinical-engineering team owns the operational picture, not the IT vendor. Hospitals that skip this review end up with analytics that misfire so often the operations team disables them — at which point the analytics tier was a wasted spend.

## Callout — what buyers most miss

**The clinical-engineering lead has a veto over CCTV that the facilities lead does not.** Hospital procurement typically routes CCTV through facilities and IT, who handle the contract and the network. The clinical-engineering lead — the person responsible for medical-device safety, infection control and patient-care continuity — is brought in for sign-off in week ten, often after specifications are locked. That is the wrong sequence: bring clinical engineering into the room from week one, or expect to discover their objections in week ten when they are also irreversible.

## Reference deployment context

The Tinsukia Medical College & Hospital deployment we handed over for NCC Limited in 2024 carried a coordinated IP CCTV layer engineered to all four of these constraints — privacy zoning enforced at the camera-placement drawing, retention sized per clinical zone, cabling routed against the infection-control annotation, and analytics reviewed by clinical engineering before commissioning. The system has been operating at the hospital for over a year under our active AMC, and the clinical-engineering team operates the VMS independently of our intervention. That is the test the design was built to pass.

## References

1. NABH 5th Edition Accreditation Standards for Hospitals — clauses on facility management, patient rights and infection control.

2. National Building Code of India 2016, Volume 2 — fire and life-safety provisions for hospital occupancy.

3. Tinsukia Medical College & Hospital — 200-bed teaching hospital, integrated ELV stack delivered for NCC Limited, 2024 commissioning.

4. Medical Council of India ethical framework — patient confidentiality and consent provisions.