Delhi’s harder problem is protecting ICU beds, not building them
Delhi holds the densest concentration of tertiary and quaternary capacity in the country, and almost none of it serves Delhi alone. Referral arrives continuously from western and central Uttar Pradesh, from Bihar, Haryana and Uttarakhand, at the end of a long journey — so patients present later and sicker than where the referral chain is short.
The consequence is occupancy. A unit in the high eighties behaves nothing like one at sixty-five percent: no slack to absorb a respiratory surge, discharge decisions made under bed pressure, and a real share of the most expensive beds in the hospital holding patients who no longer need Level III care because there is nowhere to step them down to.
So step-down capacity belongs in the brief from day one. For every ten ICU beds, plan six to ten high dependency beds. HDU plans at roughly ₹5-8 lakh per bed against ₹18-25 lakh for Level III — by a wide margin the cheapest critical care capacity a Delhi hospital can add, and the only reliable way to shorten ICU length of stay. Put it adjacent, on the same services spine, with two oxygen outlets and an essential-power circuit at every bay so a surge can convert bays upward.
Retrofit is the default, and the ceiling decides what is possible
Delhi has no spare land, so most critical care work here is conversion inside a building that never stops operating. Blocks from the 1970s and 1980s frequently sit at 3.0 to 3.3 metres slab to slab, and an ICU wants far more from that void than a ward did: ductwork for 10-15 air changes per hour, return path, gas copper, containment and steel for pendant loads.
Pendant loading is the item most often missed. A loaded single-arm pendant carrying ventilator, monitor and pump stack imposes a substantial dynamic moment at full extension and must be fixed to the slab through a dedicated frame, never the ceiling grid. At 3.0 metres slab to slab you finish near 2.55 to 2.7 metres, where a bridge pendant’s swing arc fights the examination light and IV track.
Grade the bays rather than specify uniformly: pendants where 360-degree access genuinely matters, bed head units elsewhere at 90 to 120 mm of wall depth and no slab fixing. Split ductwork into two shallow branches rather than one deep run, hold velocity low enough to keep bed-head noise under 45 dB(A), and phase the build in bays of four to six behind sealed hoardings under negative pressure.
Winter particulate is an operating line, and the ICU pays more than the theatre
Filter loading follows the air volume pushed through the media, and an ICU air handler runs continuously — 8,760 hours a year against the two to three thousand a theatre AHU accumulates on day lists. The same ambient dust deposits roughly three to four times the mass in an ICU train, and the late-October-to-January NCR season is layered on top of that.
Build the train properly: coarse pre-filter, MERV 14 or F9 fine filter, terminal HEPA H13/H14 at isolation and protective bays, dedicated pre-filtration on the fresh-air unit, and differential pressure gauges across every stage so changes are triggered by measurement rather than calendar. Then be disciplined about outdoor air. ASHRAE 170 asks two outdoor changes within the six total minimum, and every litre beyond that is dust and cooling load bought at full price — verify with CO2 monitoring rather than assuming more fresh air is always better.
Sizing plant from both ends of a very wide year
Delhi runs past 45°C in May and June and down to single-digit January nights, and the ICU is occupied through all of it. Air-cooled condensers derate exactly at the summer peak, so plant must be selected against the derated figure or it will not hold 21-24°C when it matters. In January sensible load collapses while the unit still owes 10-15 air changes per hour, and without reheat and variable-speed supply the space overshoots cold and drifts outside the 30-60 percent humidity band ASHRAE 170 sets. Fix that strategy at design stage; retrofitting it into a running ICU is close to impossible.
Isolation, zoning and the rest of the bedside engineering
Plan one airborne infection isolation room per eight beds rather than the usual ten, at 12 air changes per hour and about 2.5 Pa negative, with anteroom, sealed penetrations, self-closing doors and a permanently visible pressure monitor. Site them at the entry end so an infectious admission never travels past open bays, and damper two to four adjacent bays for switchable exhaust so a cohort area can be created during a surge — trivial at construction, near-impossible afterwards.
The rest follows ISCCM norms: 125-150 sq ft per open bay, 250-300 sq ft gross per bed, 2.4-2.5 m bed centres, nurse station sightlines within 12 metres, one-directional clean-to-dirty flow.
Approvals, life safety and scheme economics
Interior fit-out inside a sanctioned plate is largely a Delhi Fire Service NOC exercise, plus DPCC consent where plant or DG capacity changes and a registration update under the Delhi Nursing Homes Registration Act. Touch built-up area or use and DDA or MCD sanction enters the programme — the long pole. NBC 2016 compartmentation and a workable horizontal evacuation route for non-ambulant ventilated patients belong on the layout: cross-corridor doors wide enough to move an occupied bed, and smoke management that does not destroy the isolation pressure regime.
Economics sit underneath all of it. CGHS rates and Ayushman Bharat package volumes cap revenue per ICU day across much of the market, while PM-ABHIM funding drives critical care block construction with defined documentation and inspection expectations. Where revenue per bed-day is fixed, the only levers left are running cost and length of stay — the argument for honest filtration budgeting and for the step-down beds.
How we deliver a Delhi ICU from Pune
We have one office, in Pune, roughly 1,400 km away. No Delhi office, no NCR branch, no standing North India team. The distance is real, and managed by structure rather than denied.
- Site survey within 5 to 7 working days, by the engineer who will own the design.
- A site team deployed to Delhi for the continuous installation window, accommodated locally, under a named project manager accountable through validation and handover — not intermittent visits.
- Consolidated material consignments sequenced against the programme and checked against the bill of materials before despatch — a missing gasket set costs days in transit, not hours.
- An enlarged critical-spares handover — pre-filter and fine filter sets, gas terminal spares, pendant brakes and hoses, nurse call modules, pressure sensors and control panel components.
- Documented biomedical training on filter changes, pressure interpretation, isolation verification, nurse call diagnostics and gas alarm response, backed by written AMC windows: same-working-day remote diagnostics and an engineer on site within 48-72 hours for a unit-down fault. Better a window we can hold than a same-day promise from 1,400 km away.
Settle three things before briefing anyone: measured slab-to-slab height in the space you intend to convert, your true level of care and HDU ratio, and whether approvals touch DDA or MCD sanction. For full technical scope, level definitions and the validation protocol, see our ICU setup service page.