Mwangaza’s simulation room was nothing special. It had one tired mannequin, a training monitor, a trolley with demonstration medicines, and a whiteboard carrying the ghosts of old lessons. That morning Farah wrote across it in large letters: `ESCALATION LADDER — TEST, NOT TRUST.`
Malik brought a draft checklist built from Kevin’s case, Mama Sera’s night, and the family forum. Achieng carried forms. Kamau and two junior doctors would run the simulations.
“First scenario,” Farah said. “Patient with chest pain. Initial ECG not decisive. Vitals borderline. Treat it like a real shift.”
Kamau moved through the checklist: stabilize, repeat ECG, open the receiving line, define a trigger, assign an owner. Everything went smoothly. The timer showed twelve minutes until the referral pathway was open.
“Pass,” one junior doctor said.
Malik was not satisfied. “Too easy. Change the data.”
Farah handed over another card. “Blood pressure improves. Pain decreases. Receiving hospital says there is no bed. What now?”
“Continue monitoring and update in thirty minutes,” Kamau said.
Achieng asked, “What if the nurse says the patient looks different, but the numbers have not crossed the trigger?”
Kamau looked down at the checklist. “The rule doesn’t say.”
“Exactly,” Malik said.
They added a `clinical concern override`, but Kamau objected. “Then everybody can call a senior because of a feeling.”
“I would rather a senior be called for a concern someone can describe,” Achieng said, “than have a nurse stay quiet because the number has not reached the threshold.”
Farah wrote a condition beside the override: it had to include a concrete observation and trigger mandatory senior review, not automatic transfer.
The second scenario was a child in respiratory distress. The checklist contained an oxygen threshold, respiratory rate, and response-to-treatment criteria. Wanjiru, one of the junior doctors, followed every step. On the monitor the simulated child looked “stable,” but Farah introduced new information: increasing exhaustion and weaker chest movement.
“The threshold has not been crossed,” Wanjiru said.
“And what do your eyes tell you?” Achieng asked.
“This child is tiring.”
“So?”
“Override. Senior call now.”
The timer stopped.
“A good checklist needs a door that lets you say, ‘This is not normal,’” Malik said.
Kamau wrote in the margin: `numbers support judgment; do not replace it.`
The third scenario was harder. An older patient with infection. Blood pressure drifting downward. Bahari had no bed. The protocol called for repeated monitoring and a shared plan. The team did everything properly. Then Farah added the twist: the phone bridge was down and the other ambulance was away.
“We use backup contacts,” one junior said.
“None of them answer.”
“Then continue local care.”
“Until when?” Malik asked.
Nobody answered.
That was when they saw the weakness in their escalation ladder: it still depended on the receiving hospital responding. They had not created a trigger that automatically started a search for an alternative facility after a defined period of silence.
“If we set a fixed number of minutes,” Kamau said, “we may send people farther away without good reason.”
“If we set no time at all,” Farah replied, “we rebuild the old `await callback` problem.”
They argued until they reached a compromise: a `no-response escalation` rule. In a critical case, if the receiving hospital did not respond within the defined period, the owner had to escalate to a senior and begin an alternative-facility search in parallel. It did not mean the ambulance left without a destination. It meant silence could not count as a plan.
“That is Kevin’s lesson in one sentence,” Achieng said.
“And Mama Sera’s lesson is that no bed does not end coordination,” Malik replied.
They ran the scenario again. This time the junior doctor used the override earlier, an owner was named, and the alternative search began while the phone bridge was down. The timer fell.
Wekesa arrived halfway through. “I’m told the protocol is three pages now.”
“The draft is three pages,” Malik said. “The frontline card will be one page.”
“And liability? Once you publish triggers, a family can say a rule was crossed by two minutes.”
“That is why we separate a reassessment trigger from automatic transfer,” Farah said.
Wekesa read the draft. “It is more defensible. Training will be expensive.”
“Cheaper than a protocol nobody uses,” Achieng replied.
They ran one last scenario without Malik intervening. Wanjiru led the team. The patient data started normally, then changed in a pattern not covered by the numerical threshold. Wanjiru stopped, studied the numbers, and said, “Clinical concern override. Senior review now.”
Farah stopped the timer.
“Good,” Malik said.
Kamau, however, had turned over the scenario sheet. He frowned.
“Wait. Under our first version, this patient would have remained under observation another twenty minutes.”
“That is why we simulate before launch,” Achieng said.
Kamau picked up the old card they had marked `PASS` after the first scenario.
The simulation timer ended.
He looked at the team and said, “Under our first rule, this patient would have died while still marked PASS.”
---