Predictive Underwriting in the Caribbean: Pharmacy Data, Mortality Models, and the Future of Life Pricing

What pharmacy claims, electronic health records, and behavioural data can tell underwriters that medical questionnaires never could — and why the Caribbean is closer to building these models than most boards realise

 

Three years ago, the dominant question in U.S. life insurance underwriting was how to accelerate decisions without losing underwriting quality. The answer, increasingly, is predictive models drawing on pharmacy data, motor vehicle records, electronic health records, and credit-based mortality signals. Caribbean carriers are not yet using these tools at scale. The data infrastructure to do so, however, is closer than most boards realise — and the carriers that build first will own a structural pricing advantage that competitors will struggle to close.

 

Where the question is today

Caribbean life underwriting still looks much as it did fifteen years ago. A proposer completes a medical questionnaire, attends a paramedical examination, provides recent laboratory work, and waits — sometimes for weeks — while an underwriter consults reinsurance manuals and case files to reach a decision. For straightforward cases, the process produces a defensible answer. For the cases that matter most — large face amounts, accelerated underwriting candidates, atypical health profiles — the process is slow, inconsistent, and expensive. Worse, it leans heavily on what the proposer chooses to disclose.

Article 01 of this series argued that Caribbean group life pricing is heading toward a structural break. Individual life pricing is travelling on a slower but parallel curve: the underwriting decision is increasingly the binding constraint on both placement speed and pricing accuracy. Carriers that can underwrite more accurately, more cheaply, and more quickly will win on persistency, distribution economics, and ultimately loss-ratio control. The question is no longer whether predictive underwriting is coming to the Caribbean — it is which carriers will build it first.

 

The question is no longer whether predictive underwriting is coming to the Caribbean — it is which carriers will build it first.

What predictive underwriting actually is

Stripped of marketing language, a predictive underwriting model is a statistical engine that takes a set of attributes about a proposed life — some traditional, many non-traditional — and produces an estimate of that life’s mortality risk relative to a reference population. The output is not a binary accept-or-decline. It is a continuous score that allows the carrier to assign the proposer to a risk class, route the case to the appropriate underwriting path, and price the cover with materially greater accuracy than a medical questionnaire alone can deliver.

Three sources of data have proven most valuable in the U.S. market, and each is meaningful for the Caribbean — albeit at different stages of development.

Pharmacy claims data

Of all non-traditional underwriting inputs, pharmacy data has produced the strongest mortality lift in U.S. work over the last decade. What someone is prescribed reveals what they are being treated for, how compliant they are with treatment, how their conditions are progressing, and how their drug regimens compare with cohorts of similar mortality experience. A history of beta-blockers prescribed alongside ACE inhibitors and a recent introduction of a glucose-lowering agent tells an underwriter more about cardiovascular risk than three pages of a medical questionnaire ever will. In the Caribbean, pharmacy data is more accessible than most insurers assume — the major pharmacy chains in Jamaica, Trinidad, and Barbados have substantially digitised dispensing records over the last decade, and a structured industry conversation about consented access is overdue rather than premature.

Electronic health records and laboratory data

Where pharmacy data reveals what is being treated, lab and EHR data reveal underlying physiological state. HbA1c trends, lipid panels, kidney function markers, and prescribing-physician notes are mortality-relevant in ways no questionnaire response can replicate. The Caribbean lags the U.S. on EHR penetration — but several jurisdictions, notably the Bahamas, Barbados, and parts of the OECS, have made meaningful progress on national health information infrastructure. The CARPHA-led regional NCD surveillance work has further laid groundwork that insurance carriers could, with appropriate consent and data-sharing protocols, build on.

Behavioural, motor vehicle, and credit-based attributes

U.S. work has consistently found that what proposers do correlates with how long they live, not merely what conditions they carry. Motor vehicle records reveal risk-taking behaviour. Credit-based attributes — used in a regulated, mortality-validated way, not as a general financial-fitness measure — show statistically significant lift in mortality models. Behavioural data from app usage, wearables, and consented digital footprints adds another layer. Caribbean carriers face tighter constraints than U.S. peers on the use of credit-based attributes (data protection regimes in Jamaica, Trinidad, Barbados, and the Cayman Islands set the rules), but the principle stands: behavioural data is mortality-predictive, and even a fraction of what U.S. models use would represent a step change for Caribbean underwriting.

WHAT THIS IS NOT Predictive underwriting is not an algorithm replacing an underwriter. The strongest implementations in the U.S. market route cases — straight-through automated decisions for low-risk cases, accelerated paths for clear-cut cases, full traditional underwriting for genuinely complex cases. The model handles the volume; the human underwriter handles the judgement calls. Caribbean carriers that adopt predictive underwriting as a triage tool, rather than a replacement, get the speed-and-cost benefits without sacrificing decision quality.

Why this matters for Caribbean carrier economics

The case for predictive underwriting in the Caribbean is not theoretical. Four economic effects are well-documented from U.S. and U.K. deployments, and each translates directly into the Caribbean operating environment.

1. Acquisition cost compression. Predictive triage routes 40 to 60 percent of typical case volume to accelerated or automated underwriting paths, removing paramedical examinations, blood work, and underwriter time from a meaningful proportion of cases. Caribbean carriers pay similar per-case costs for these services as U.S. peers — a 50 percent reduction in those costs across half the book is a material expense saving.
2. Placement-rate improvement and persistency. The single largest leak in the new-business funnel for Caribbean life insurers is the time between application and policy issue. Proposers reconsider, distribution channels lose momentum, competitors close the gap. Predictive underwriting compresses average cycle time from weeks to days, and in clean cases to hours. Placement rates rise. The cases that do place persist better, because the underwriting was done before enthusiasm cooled.
3. Pricing accuracy in the middle of the book. Traditional underwriting tends to be highly accurate at the extremes — clear-standard lives at one end, declined risks at the other — but blunt in the middle. The 60 percent of cases that fall into the broad standard-to-mild-impairment range are precisely where predictive scoring most improves discrimination. Carriers that price this middle band more accurately capture profitable risk that less sophisticated competitors miscategorise.
4. A renewable data asset. Once a carrier builds the underwriting data infrastructure, it becomes a permanent capability. Each year’s experience updates the model; each model refresh tightens the pricing; each tightened price compounds the competitive advantage. This is the closest a Caribbean life insurer comes to a structural moat — and unlike capital strength, it is buildable by mid-sized carriers prepared to invest.

The Caribbean data infrastructure is closer than most boards realise

The standard objection to predictive underwriting in the Caribbean is that the data infrastructure does not exist. The honest answer is that it exists more than insurers think, and the gaps are narrower than the gaps that the U.S. market faced fifteen years ago when these models were first being built.

Pharmacy chains in Jamaica, Trinidad, Barbados, and Bahamas have moved substantially toward digital dispensing. Major chains hold structured, person-identified prescription histories spanning multiple years. With appropriate consent frameworks and Data Protection Act compliance, this data is accessible to underwriters in a form usable for both individual risk assessment and model calibration. The conversation is regulatory and commercial, not technical.

Health ministries across the region have invested in NCD registries, hypertension and diabetes surveillance programmes, and increasingly digitised public-sector EHR. Private healthcare networks — particularly hospital groups operating across multiple territories — hold meaningful longitudinal records. Insurance industry associations could convene credible regional discussions about data-sharing frameworks that respect privacy while enabling mortality model development. CARILEC, IAJ, ICATT-affiliated industry bodies, and the regional supervisory community are all natural participants.

Data protection regulation is increasingly aligned across the major Caribbean territories — Jamaica’s Data Protection Act 2020, Barbados’s 2019 framework, Trinidad’s evolving regime — and provides a workable legal scaffolding for consent-based predictive underwriting models. A carrier that designs from privacy-by-design principles, secures appropriate consents at application, and uses data only for the purpose disclosed will be on solid legal ground in every major Caribbean jurisdiction.

The regulatory considerations every board should anticipate

Caribbean insurance regulators have not yet issued specific guidance on predictive underwriting models. That should not be read as regulatory indifference — international supervisory attention to actuarial model governance, accelerated underwriting, and the use of non-traditional data in life insurance is increasing rapidly, and Caribbean supervisors generally follow developments in larger markets with a one-to-three year lag. The carriers that build with model governance frameworks already in place will be in materially stronger positions when guidance does arrive.

Four supervisory questions can be anticipated with high confidence. First, what data is used and how was the proposer’s consent obtained? Second, has the model been validated against actual mortality experience, and how is performance monitored over time? Third, can the model be audited — are decisions explainable, are inputs documented, is there a clear governance trail? Fourth, what happens when the model gets it wrong — what is the override process, who has authority, how are errors fed back into model improvement? A carrier that can answer all four questions with documented processes is in a position no Caribbean competitor currently occupies.

THE MODEL GOVERNANCE PRIORITY Boards considering predictive underwriting should ensure that model governance is treated as a first-class question from the outset, not as a downstream compliance exercise. The most common failure mode in U.S. deployments has been carriers building strong models without commensurate governance infrastructure — and finding the governance gap exposed only after the regulator or the auditor began asking questions. Building the model and building the governance must happen in parallel.

A realistic 24-month path for a Caribbean carrier

What follows is a sequenced path that a mid-sized Caribbean life insurer can realistically follow over a 24-month horizon, without enterprise-scale capital commitment in year one. The protocol is deliberately staged so that each phase builds optionality for the next.

Months 1-6 — Foundation and feasibility

Establish data-source inventory across pharmacy, health information, and behavioural domains within the carrier’s home territory. Engage Data Protection officers and external counsel on a consent framework that supports both individual risk assessment and aggregate model development. Run a small-volume retrospective analysis on the carrier’s own claims experience to confirm that pharmacy or laboratory data, where it exists, demonstrates statistical lift on the in-force book. The output of this phase is not a model — it is a decision-quality answer to the question, is this worth the carrier’s capital?

Months 7-12 — Pilot model build

If the feasibility analysis supports proceeding, the next phase builds a pilot predictive model — initially in a single product line, typically simplified-issue term life or accelerated-underwriting universal life. The pilot model uses a constrained data set (pharmacy plus basic biometrics) and is deployed in shadow mode, generating scores alongside traditional underwriting decisions without yet driving them. Six months of shadow-mode operation produces the validation evidence needed to support a regulator conversation, an auditor walkthrough, and a board approval for live deployment.

Months 13-18 — Live deployment and triage routing

The pilot model moves into live deployment as a triage tool — routing cases between accelerated, traditional, and decline-pending-review paths. Traditional underwriting capacity is preserved for the complex middle of the book. Expense savings and placement-rate improvements begin to compound. Model performance is monitored monthly, with feedback loops from emerging claims experience built into quarterly refresh cycles.

Months 19-24 — Expansion and platform

With one product line proven, the model architecture expands. Additional data sources are integrated (EHR access, behavioural data, motor vehicle records as available). The model is generalised across additional product lines (whole life, larger face amounts, group life conversion options). The carrier emerges from the 24-month cycle with a permanent underwriting data and analytics capability — and a structural competitive position no Caribbean competitor without similar investment can credibly match.

WHAT THIS PATH TYPICALLY PRODUCES Carriers that have followed this kind of staged path in markets at comparable points in their data evolution have, in published case studies, achieved acquisition-cost reductions of 30 to 50 percent on the routed portion of their book, placement-rate improvements of 8 to 15 percentage points, and material persistency improvements driven by faster decisioning. The investment to reach the 24-month stage is meaningful but not enormous — typically in the range that a mid-sized Caribbean carrier funds through ordinary capital expenditure, not strategic capital raises.

The carriers that move first will define the market

Caribbean life insurance is a market that has historically rewarded patience. Distribution relationships matter more than product innovation. Brand familiarity drives placement more than pricing precision. These are real characteristics of the market, and they are not disappearing. But they coexist with another truth: the carriers that built the foundational data and analytics capabilities in U.S. and U.K. markets fifteen years ago now sit in pricing positions their competitors cannot match through commercial effort alone. The window between when the technology becomes available and when it becomes a competitive necessity is the window in which structural advantages are built.

The Caribbean is in that window now. The first two or three carriers in each major territory to build credible predictive underwriting capability will define what the market expects of an underwriting function over the next decade. The carriers that wait until competitors have proven the model in the local market will be paying — in lost business, in adverse selection, in pricing disadvantage — for the comfort of having waited. The strategic question is not whether to invest, but how quickly to start.

The window between when the technology becomes available and when it becomes a competitive necessity is the window in which structural advantages are built.

 

ABOUT THE SERIES The Caribbean Actuarial Imperative is a 16-article series from Dawgen Global’s Actuarial & Insurance Regulatory Advisory Division. The series examines the structural shifts reshaping Caribbean insurance — pricing, reserving, reinsurance, enterprise risk, regulation, experience data, modelling technology, and transactions — and what insurance boards, executives, and regulators should be doing about them. The Actuarial & Insurance Regulatory Advisory Division is Fellowship-led, independent of any global broker or reinsurance group, and integrated with Dawgen Global’s broader Risk Advisory, Audit & Assurance, Tax Advisory, M&A, IT, and Cybersecurity practices. Enquiries: [email protected] Please reference ‘Actuarial Division’ in your subject line.

 

PREVIOUSLY IN THE SERIES Article 01 Why Caribbean Group Life Pricing Is About to Break

NEXT IN THE SERIES Article 03 IFRS 17, Three Years On: What Caribbean Insurers Got Right, What They Got Wrong, and What Comes Next

◆ ◆ ◆

Big Firm Capabilities. Caribbean Understanding.

About Dawgen Global