Triage

Firefighter and full ALS paramedic with the City of Alexandria Fire Department. Before, and during that time, I was a critical care medic with an interfacility transport service. Think: when helicopters can’t fly between hospitals. I was on a specialty transport team. ICU, NICU, PICU. Patients who needed extra help or a specialist somewhere else.

EMT, ACLS, ITLS, NFPA 1001. HAZMAT Operations. I sat on a medical protocol review committee that wrote the field protocols other medics worked to.

People assume the thing that carries over is composure. I’m calm under pressure. But composure is easy to fake. You get back to the station and scream into a pillow. Take it out in the gym. The tenured guys, the ones already burnt out, never gave advice. They had one answer for all of it. “I know I know…here, drink this”

Prioritization ranks a list you control. Triage is what you do when the list is longer than the capacity, the information is thin, and waiting costs more than choosing.

The field version sorts patients into three groups. Who is treated now. Who waits. Who cannot be saved with what is on scene. The third group is the one nobody wants to name, and it is why the system works. Spend forty minutes on the patient you cannot save and three survivable patients die.

Most roadmaps do exactly that.

Decide on what you have#

You never get the full picture on scene. You have seconds to gather, and read the room. Scan for clues. Delegate. Send the crew to grab information. You get vitals, a story from a bystander that is half wrong, and a clock. You act on the best read available and correct as more comes in. Constantly shaping. Iterating. Reevaluating. Situations change. The patient changes. Priorities now shift. You pivot. Adjust. Reassess. Repeat.

Executives stand still and call it diligence. One more discovery round. One more stakeholder review. That’s their job. Wait for the data. More often the delay costs more than the wrong call would have.

Nobody prices the delay.

Price the reversal#

Nobody deliberates about an airway. Protect the airway. You deliberate before something you cannot undo. The cost of reversal sets the care a decision deserves.

Most product decisions are cheap to reverse and get escalated as though they are not. Naming. Layout. Rollout order. The expensive ones go quickly, decided by whoever was in the room. Data model. Integration contract. What you put in writing to a customer.

Match the deliberation to the reversal cost. That alone returns weeks a year.

Protocols#

Field protocols do not replace judgment. They make the known cases automatic so attention is free for the case that does not fit.

“Protocols are black and white, but all our patients are grey.”

— my old EMT mentor

Writing them taught me what a good one looks like. Specific enough to follow under load. Explicit about when to break it. Short. A protocol nobody can recall at three in the morning is a liability. Remove the ambiguity. Provide clarity.

An engineering function needs the same. How a release goes out. What happens when production breaks. Who is called, in what order. Decide those once. Otherwise the function spends its judgment on solved problems and has none left for the live one.

Handover#

Almost every bad outcome I saw happened at a boundary. Crew to crew. Field to hospital. The person who knows to the person who now owns it.

Software has the same boundaries and treats them lightly. Engineering to support. Product to delivery. Delivery to market. The person who resigns to the person who inherits. A structured handover reads as bureaucracy until the first missing detail costs a week.

Write it while you still know the thing. Nobody writes it after. The person who inherits it does not know what to ask. We don’t know what we don’t know.

What does not carry over#

Nobody dies when a release slips. I am wary of war rooms and firefighting language in software. It inflates ordinary work and tires people for nothing. The stakes are not that high.

The pace does not carry over either. In the field you run flat out for a time, then you stop. Software does not stop. A function held at emergency intensity does not get fast. It gets tired, and tired functions make the errors triage exists to prevent. Burnout. Software keeps me up at night. But not in the same way as the feel of his ribs fracturing under compressions. The lifeless body of a child who never got a chance. Scars of wounds that never heal right.

The method carries over. The urgency mostly should not. That’s where the composure comes from. The stakes are not life or death in software.

The third group#

Most functions can rank a backlog. Few will name the third group.

It is not the nice to have. That one waits, and everyone knows it waits. The third group is the ask with no solution behind it. The feature you have no data to justify building. The thing somebody called a fire that is not a fire worth putting out.

Unnamed, they stay on the board. They get discussed every week. They take attention from work that could have shipped, and they take it quietly.

Somebody is already deciding that work will not happen. They are deciding it by never getting to it.

Name it. Then go and do the rest.