The constraint

Hurricane Melissa had removed access to routine care across St. Elizabeth. CASSM's response was built on a hybrid model developed by Dr. Franz Collins, a stationary base clinic handling large inflows and emergencies, with mobile medical units deployed by bus to communities that were shut-in. The layout plan specified the clinical zones, the team structure across three units with cross-covering for wound care and emergency capabilities, and the resource categories the deployment would require.

Clinical services at the site were delivered by MobiCare Medical Centre under a CASSM-funded initiative, with volunteer practitioners working within that clinical structure. CASSM's role was to fund the deployment and commission the facility in which care would be provided. This note concerns the second of those, the resourcing decision, and not the clinical work itself.

Translating that plan into a commissioned site required reliable data points. The quantity of tents, tent sizes, how many beds, tables, and chairs, cooled by fans or air conditioning, partitioned or open, porta-potties, and hand wash stations. Each decision on a data point carried a cost, and at the point the decisions had to be made, CASSM did not know what funding would materialize.

The problem was therefore not to design an ideal facility. It was to design a facility that remained clinically adequate given funding uncertainty.

Building the model

Square footage and seating capacity for each tent type were provided by Mr. Gifford Boothe of KVG Hireage in Kingston, vendor data on what each configuration could actually hold, rather than estimates. Equipped with this information, CASSM built a spreadsheet that derived facility requirements from two inputs, a volunteer roster classified into twenty-nine designations as medical or non-medical, and an expected patient count entered for each area.

Personnel allocations flowed automatically from the roster into each workstation, equipment counts followed, and changes to the roster or the expected volume required a re-run of the whole configuration, which meant a proposed budget cut could be evaluated for what it would cost in capacity rather than argued about in the abstract.

The model produced two configurations. The first was what the deployment wanted, and the second was what it committed to after cost trimming:

• Beds — 7 reduced to 4
• Tables — 30 reduced to 13
• Chairs — 69 reduced to 57
• Portable toilets — 3 reduced to 2
• Tents — 8 in both configurations

The trimming was not uniform. Registration and Documentation grew from 10×10 to 15×15, while the Examination Area shrank from 20×20 to 10×10 (numbers are in feet), and capacity moved toward the front of the patient flow, on the reasoning that a bottleneck at intake stops the whole clinic, while a constrained examination area slows one stream.

Deciding what to cut

A model that quantifies trade-offs does not make the trade-offs. Reducing 30 tables to 13 and 7 beds to 4 required a basis for choosing, and the basis was not purely arithmetic.

The dominant uncertainty was not patient volume. It was capability. This was the first deployment in which this particular group of practitioners, volunteers, and logistics personnel worked together as a unit, and individual performance under field conditions could not be predicted from a roster. Provisioning therefore had to account for a team whose limits were unknown.

Against that, one capability was known directly rather than inferred. Dr. Collins's emergency practice at Kingston Public Hospital carries the kind of experience that does not transfer onto a spreadsheet. This translates to sustained decision-making under acute pressure with incomplete information. CASSM leadership had also seen him work at close range, during the final days of a parent's life. That is not evidence in any formal sense. It is the kind of judgment that decision-makers rely on constantly and rarely state, an assessment of a person formed by observation rather than by record, carrying real weight in a resourcing decision that had to be made without data.

Two questions ran in parallel. The first asked what demand was most likely, chronic conditions, principally diabetes and hypertension, interrupted medication access, and routine presentations deferred by the hurricane. The second asked what the worst plausible presentation would be, severe blunt force trauma, fractures, a life-threatening emergency arriving without warning. Neither question answers the other. Sizing purely to expected demand leaves the worst case uncovered, while sizing purely to the worst case exhausts a constrained budget on capacity that will probably sit idle. The configuration had to hold both.

What was not tradeable

Three provisions were treated as fixed rather than negotiable, and cuts were made around them:

• The Emergency Area — retained as fully enclosed with air conditioning and partitioning. If it were needed at all, it would need to be sterile and private.
• The Dressing Area — retained as fully enclosed on the same reasoning.
• A dedicated pharmacy station — retained as a discrete workstation rather than folded into another area.

Naming these as fixed constraints before trimming meant that the reductions occurred within a bounded space. The alternative, trimming proportionally across all areas, would have degraded the three stations least able to tolerate degradation.

Flexible staff, fixed space

The model deliberately assigned a baseline of zero personnel to the Examination and Emergency areas. Both were low-probability, high-consequence stations, and the deployment cannot leave them uncovered, but staffing them permanently parks practitioners in tents that may see nothing all day. Instead, a pool of six practitioners floated across the site, moving to those areas as need arose.

The reasoning was probabilistic. Presentations requiring severe wound care were expected to be a small fraction of total volume relative to chronic disease management, medication resupply, and routine consultation. Accepting a short delay in reaching a low-volume station was judged a better use of constrained capacity than continuous staffing.

Under a budget that could not accommodate any surplus capacity, CASSM covered demand uncertainty by floating staff. The model had no equivalent mechanism for physical space.

What the day actually looked like

The clinic operated from approximately 9 am to approximately 4 pm. The end time was not clinical. During the pre-deployment site visit conducted by Dr. Collins, Mrs. Gwen Miles, and Dr. Miles, the terrain was assessed and reported to the bus drivers, downed trees, debris, and standing water on the return route. The drivers set the departure time to get the convoy out in daylight. That decision sat correctly with the people who would have to drive it.

Within that window, the base clinic registered 209 patients, 172 of whom were seen by doctors on site, with a further 34 seen by mobile field teams. One hundred prescriptions were filled and ten dressings performed.

The deployed site fell short of even the trimmed configuration; however, the final numbers were not recorded.

Where the model was wrong

Pharmacy was allocated a 10×10 tent with three tables, the smallest clinical footprint on the site. It became the binding constraint. With hospitals and pharmacies destroyed across the parish, free medication matched to individual ailments drew a volume the space could not absorb, and the pharmacist could not leave the station to collect stock. Because the operating window was fixed by the return journey rather than by demand, there was no slack to absorb the delay and work not completed by 4 pm would not be completed at all. Floating staff could not solve this. A shortage of people can be moved, a shortage of square footage cannot.

The model sized a facility, not a day

Every output the spreadsheet produced was a quantity of physical resource derived from expected patient volume. Nothing in it represented time. The six-hour operating window, set by travel distance, road conditions, and available daylight rather than by the mission, was never included in the calculation, which meant throughput requirements were never tested against the hours actually available. For a post-disaster deployment this is not a minor omission and usable operating time is one of the first things a disaster takes away. It constrains what a given facility can deliver as directly as its footprint does.

What was not measured

The model was built to plan a deployment, not to be studied. No provision was made for capturing variance, and as a result three questions cannot now be answered, namely how far the delivered site fell short of the commissioned configuration, how the floating practitioners distributed themselves across the day, since they self-deployed on their own reading of need with no coordinating mechanism and no record kept, and how throughput at each station tracked against the model's assumptions.

Documenting field innovation was an afterthought on this deployment rather than part of its design. Recording that plainly is more useful than reconstructing figures after the fact.

What changes next time

Four additions would make this model materially stronger, and all of them come from what could not be answered above.

1. Usable operating hours as an explicit input, derived from travel time and daylight, with throughput calculated against it.
2. A space-contingency mechanism to match the staffing one. A reserve tent or expandable footprint, assignable on the day to whichever station is absorbing the surge.
3. A coordination point for the floating pool, established with the clinical provider, so that cross-cover is directed and recorded rather than emergent.
4. A variance record commissioned against delivered, projected against actual, captured during the deployment rather than reconstructed afterward.

The model has not been validated. It planned one deployment adequately under severe constraints, and the single day it was tested exposed its limits. That is a starting point for a reusable method, not the method itself.

Acknowledgements

The hybrid deployment model on which this work rests, the layout plan, team structure, and operational flow, was developed by Dr. Franz Collins.

Mrs. Gwen Miles coordinated the volunteer group and managed medication distribution, drawing on prior medical mission experience in Jamaica with the Jamaica Children's Heart Fund, which delivered open-heart surgery to children from infancy to thirteen years of age before its merger with Chain of Hope UK.

Mr. Gifford Boothe of KVG Hireage, Kingston, provided the tent specifications and capacity data on which the resourcing model was built.

See From the Field for the full account of the Hurricane Melissa response, and Connectivity as clinical infrastructure for the communications capability deployed at the same site.