Non-communicable diseases — cardiovascular disease, diabetes, cancer, and chronic respiratory illness — account for the large majority of deaths in Jamaica, with published estimates placing the NCD share of mortality at roughly three-quarters or more. These are, to a substantial degree, preventable deaths driven by a small number of modifiable risk factors: diet, physical inactivity, tobacco and alcohol use, and undiagnosed or uncontrolled hypertension and blood sugar.
Health policy typically starts at the top — a budget, a set of programmes — and hopes the deaths follow. This paper argues for the reverse discipline: start from the mortality figure and work backward, link by link, to the specific, fundable interventions that a given death depended on. A death from an NCD is the end of a causal chain; each link back from it is a place where the chain could have been broken, and the earliest breakable links are the cheapest.
This paper sets out a backward-induction model from NCD mortality to intervention, identifies where Jamaica's current NCD response starts too late in the chain, and offers four recommendations to the Ministry of Health and Wellness for a mortality-anchored NCD strategy.
A national NCD strategy usually begins with an appropriation and a list of activities, and its success is measured by activities delivered — clinics opened, screenings performed, campaigns run. But activities are inputs, and the only output that matters is a death that did not happen. A strategy that is not anchored to the mortality figure cannot tell whether its activities are the ones that move it.
Backward induction inverts the order. It begins with the terminal state — a death from cardiovascular disease, say — and asks: what was the immediately preceding state? An acute cardiac event. And before that? Sustained uncontrolled hypertension. And before that? Undiagnosed hypertension. And before that? A diet, an inactivity, a risk profile that was never screened. Each backward step is a link, and each link is a candidate intervention point.
Jamaica's NCD spending concentrates at the late links of the chain — the hospital, the acute event, the dialysis chair, the amputation. These are the most expensive links and the least effective, because by the time a patient reaches them the chain has almost run its course. The early links — screening the undiagnosed, controlling the newly-diagnosed, shaping the risk profile before disease begins — are where the chain is cheapest to break and where the least money currently goes.
The central quantitative insight of backward induction for NCDs is that the cost of breaking the chain is not constant along it. It rises steeply toward the terminal end. Preventing a case of hypertension through diet is inexpensive. Detecting and controlling hypertension once present is moderately expensive. Managing the kidney failure, stroke, or heart attack that uncontrolled hypertension eventually causes is enormously expensive — and often fails anyway.
This means that money spent at the early links buys far more prevented death per dollar than money spent at the late links. A rational NCD strategy therefore allocates against the chain, not against the disease category: it asks, for each link, how much death it can prevent per dollar, and moves spending toward the links where that ratio is highest — which are almost always the early ones.
Allocate NCD spending by cost-per-prevented-death along the causal chain, not by disease category or by hospital demand. The earliest breakable link is almost always the highest-yield place to spend, and it is almost always the place that is currently underfunded because it does not present as an emergency.
The largest reservoir of preventable NCD death is the undiagnosed — people with hypertension or elevated blood sugar who do not know it and will not until an acute event. Every contact with the health system, and community screening beyond it, is an opportunity to close that gap. Screening the undiagnosed is an early, cheap link; it converts an invisible risk into a manageable one.
A diagnosis is worthless if control does not follow. The link between "diagnosed" and "controlled" — reliable medication supply, adherence support, and routine follow-up — is where many chains continue to run despite the disease being known. Reliable access to first-line NCD medication is one of the highest-yield interventions available and one of the most operationally tractable.
Behind the risk factors sits the environment that produces them — the availability and price of healthy versus unhealthy food, the walkability of communities, the marketing of sugary products. This is the earliest link of all, upstream of any individual diagnosis. It is slow to move but moves the whole population's risk profile at once, and it connects directly to existing food-labelling and regulatory work.
Jamaica's NCD risk is not evenly distributed — obesity and certain risk factors fall disproportionately on women, while men are less likely to engage the health system before an acute event. A single undifferentiated strategy under-serves both. Targeting the intervention to where the risk actually concentrates raises yield without raising cost.
The Ministry should set an explicit, dated NCD mortality-reduction target and require every funded NCD activity to state its estimated contribution to that target. An activity that cannot connect itself to prevented death is a candidate for redirection toward one that can.
The Ministry should conduct a cost-per-prevented-death review across the NCD chain and progressively rebalance spending from late-link acute care toward early-link screening, medication access, and control — measured, not assumed. The rebalancing pays for itself as acute-event volume falls.
The single most tractable early-link intervention is uninterrupted supply of first-line hypertension and diabetes medication at the primary-care level. The Ministry should treat stock-outs of these medications as a reportable failure of the NCD strategy, because each one lets a known, controllable chain continue to run.
The Ministry should publish NCD incidence, control, and mortality disaggregated by sex, and design interventions against the disaggregated picture. What is measured only in aggregate is targeted only in aggregate — and Jamaica's NCD burden is not aggregate.
Seventy-eight percent is not a statistic to be recited; it is the terminal state from which the entire NCD strategy should be reasoned backward. Each of those deaths is the end of a chain that passed through screening, control, and environment — links where, at rising cost as the death approached, it could have been broken. A strategy that spends at the expensive end and measures itself by activity will always feel busy and rarely move the number.
The discipline proposed here is simple to state and demanding to apply: anchor to the mortality figure, map the chain behind it, and spend where death-prevented-per-dollar is highest — which is early, cheap, and currently underfunded. Human Intelligence LLC is prepared to support the Ministry of Health and Wellness in building the backward-induction model and the cost-per-prevented-death analysis it requires.