Costa Rica Field Trial: What the Network Taught Us

MMDM's 10th Anniversary Mission

The Costa Rica 2026 mission ran April 18–25 in Grano de Oro, a Cabécar indigenous community in the mountains of Costa Rica. Thirty-nine missionaries and 13 local Cabecar-Spanish translators — several of whom have served alongside MMDM for seven or eight years — served a record 1,121 patients in a single week. That included 151 patients seen during a one-day outreach to Bloriñak.

It was MMDM's 10th anniversary Costa Rica mission.

What "Field Deployment" Actually Means

The Costa Rica trial was paper-primary. Every patient was processed on paper in the morning. Station leads and doctors gathered in the evening to enter the day's data into OpenEMR together. The EMR was running and connected; it was not yet the primary workflow. That was intentional.

A shadow deployment is a lower-stakes environment for finding problems. The alternative — going EMR-primary without a trial — would have put data integrity and clinical workflow at risk simultaneously. The staged approach traded speed for confidence.

What Worked

Connectivity at each fixed station was reliable throughout the week. The server held under typical concurrent load. Registration, triage, vision, and dental maintained coverage across the full deployment. Android tablets connected and authenticated over the clinic network without issue.

The evening entry sessions revealed which forms needed adjustment. The medical lead's diagnosis-prompt gap was caught in session two. The vision lead adjusted the age cutoff for routing patients to the eye doctor midway through — a change that took seconds to make and would have been a multi-day turnaround if thresholds were hardcoded.

What Did Not Work

Signal dropped in the outdoor walking corridors between buildings. The medical station was barely in range — one provider was working from a position just outside reliable coverage. That is a physical AP placement problem, not a software problem, and the bill of materials for the next deployment now includes additional indoor APs and more cable.

The network switch ran out of ports. The clinic site has more stations than the initial equipment plan assumed. A second switch or a larger one is in the next procurement.

A network cable physically interfered with dental equipment in a way that created a tripping hazard. The cable had to be lifted overhead during the session. This is a cable-management and physical-setup problem. Pre-mission site walkthroughs will now include a cable routing plan.

The April 21 Reset

Midway through the deployment, the patient-creation flow broke. The diagnosis, the decision, and the rebuild are covered in a separate entry. The short version: the system was back up within the hour; the data from that morning was preserved on paper; no patient records were lost.

For Laredo

The Laredo deployment, November 4–10, 2026, adds complexity the Costa Rica site did not have: a street physically separates registration and triage from the clinical buildings, HIPAA jurisdiction applies, and a pharmacist will be on-site for dispensing and e-prescribing. The network plan will need to bridge that physical gap. The form set will need a Laredo-specific configuration. The privacy model will need to account for US patient data law in a way the Costa Rica trial did not.

The Costa Rica trial produced enough confidence to proceed. It also produced a specific punch list for what needs to change before November 4.