Before I worked on surveillance systems and dashboards, I worked in the operating room of a district hospital in Bassila, Benin, as a general surgeon for a rural population with little access to specialist care. That is about as far from a KPI framework as clinical work gets. It is also where I learned what “infrastructure” means in public health, mostly by working without much of it.
Scarcity makes priorities obvious
A well-funded hospital can hide a lot of inefficiency because there is slack in the system. A district hospital has none. Every missing piece shows up right away: in delays, in outcomes, in which operations you can safely do at all.
You learn quickly which steps are essential and which are just habit. You prioritize hard because you can’t do everything. I’ve carried that into every data and program job since. Before adding anything, I ask whether it is needed or just familiar.
You notice infrastructure when it’s missing
Public health infrastructure means data systems, surveillance protocols, supply chains, and referral networks. It mostly gets attention when it fails. Nobody praises an immunization registry that works. Everybody notices when a coverage gap goes undetected for months.
In Bassila I saw how much the system decided outcomes before any clinical decision was made. A good surgeon can’t make up for a supply chain that doesn’t deliver what the operation needs. A good treatment protocol can’t help a patient who has no reliable way to get from a rural clinic to a hospital that can treat them.
I ran into the same thing later in surveillance work, at a much larger scale. The system around an intervention usually matters as much as the intervention.
Trust is infrastructure too
Equipment isn’t the only constraint. Relationships between providers matter just as much. Referrals work when doctors in different facilities know and trust each other, and that same trust is what gets a rural clinic to send its data on to regional and national systems in the first place.
Programs that ignore this human side tend to perform worse than their technical design would predict. A well-written surveillance protocol won’t help if the clinic on the ground doesn’t trust the system enough to report into it.
What I carry forward
It’s tempting to treat public health infrastructure as a technical problem: better software, better dashboards, better pipelines. Those help. But the main thing I took from a resource-limited operating room is that infrastructure comes down to reliability. Can the people who depend on the system count on it to work every time they need it?
That’s the test I apply to anything I help build now. I care less about how sophisticated it is than about whether the people using it can rely on it.