---
title: "Five Ways AI Can Stretch Access in a Small Health System"
date: 2026-11-17
author: "Nevena Nemes"
featured_image: "https://caribbeancurrents.org/wp-content/uploads/2026/11/cover-21.jpg"
categories:
  - name: "Regional Innovations"
    url: "/category/regional-innovations.md"
---

# Five Ways AI Can Stretch Access in a Small Health System

For resource-constrained healthcare systems, the most useful artificial intelligence may not be the most sophisticated. Its value may lie in helping patients reach care more efficiently by making better use of the data, knowledge and capacity already in the system.

Artificial intelligence is often discussed in healthcare at a scale that can make the conversation feel distant from the Caribbean. The examples tend to come from large academic hospitals, multinational technology companies and health systems with deep technical capacity.

> For resource- constrained healthcare systems, the mostuseful artificial
> 
>  intelligence may not be the most sophisticated.

For many organizations across the region, the reality is different: fragmented datasets, limited analytical capacity, competing priorities and small teams expected to do more with fewer resources. In that environment, the most relevant question is not whether a small health system can reproduce the AI capabilities of a global hospital network. It is whether AI can help patients reach care more reliably by reducing the friction inside the system.

Much of healthcare analysis still begins in Excel. An analyst may identify an unusual billing pattern, filter the data, build pivot tables, restructure the dataset and gradually develop a hypothesis. AI-assisted coding can help turn those repetitive steps into reusable analytical workflows, particularly for claims investigations involving laboratory combinations, recurring procedures, prescribing behavior or other patterns that only emerge after several stages of transformation. The access benefit is indirect but important: less time spent rebuilding analysis means more time identifying where resources are being wasted, where utilization is uneven and where patients may be encountering avoidable barriers.

![](https://caribbeancurrents.org/wp-content/uploads/2026/11/vertical1-8-768x1024-1.jpg)The same principle applies to small software tools. Healthcare organizations have traditionally faced a gap between spreadsheets and enterprise systems. A problem may be too complicated to manage manually, yet too narrow to justify a major procurement. AI-assisted development creates more room in between. A health system may need a simple application to track referrals, monitor authorizations, analyze waiting lists or show where cases repeatedly stall. These are not abstract technology problems. They are access problems, because a referral that cannot be seen, an authorization that cannot be tracked or a waiting list that cannot be understood can all become delays in care.

AI can also help organizations use information they already possess. Clinical notes, referral letters, discharge summaries, complaints and audit reports may contain valuable insights, but much of that information remains difficult to analyze because it sits inside narrative text. AI tools can assist in extracting medications, categorizing complaints, suggesting coding classifications for review or identifying variables for reporting. Used carefully, that can help systems identify recurring access barriers that would otherwise remain hidden inside thousands of individual records.

There is a similar opportunity around institutional knowledge. Every health system accumulates local rules, legislation, benefit packages, procedures and policies. Generic AI tools do not automatically understand that environment, but organizations can increasingly create secure assistants grounded in their own approved documentation. Staff can ask questions in ordinary language and receive answers tied back to relevant source material. The expert remains responsible for interpretation, but the pathway through the system becomes easier to navigate when staff can find the right answer quickly rather than searching through multiple documents or waiting for the one person who knows.

> AI can also help organizations use information they already possess.

Perhaps the most important application is less about automation than about asking better access questions. Why do some patients wait longer than others? Why do referrals fail between particular facilities? Why do utilization patterns differ across communities with similar needs? Why does one pathway move patients smoothly while another repeatedly stalls?

AI can help analysts move more quickly from observation to hypothesis and from hypothesis to investigation. For a small system, that may be where the greatest value lies: not in replacing people, but in helping the system see more clearly where access is being lost.

Artificial intelligence does not replace good data governance, reliable reporting or strong analytical foundations. The real test is whether it changes the patient journey: whether it exposes a bottleneck earlier, shortens an authorization process, improves referral visibility, makes scarce capacity easier to allocate or frees expertise for higher-value decisions.

That is a less dramatic vision of AI than autonomous medicine. For small health systems, it may also be the more important one – because access is often lost not at the point where care does not exist, but in the friction between the patient and the care that already does.

This version fits Issue 3 much better because it ties Dennis directly into the issue’s core thesis: access is not just whether capability exists, but whether the system can connect the patient to it.

> Artificial intelligence does not replace good data governance, reliable reporting or strong analytical foundations.

#### **ABOUT THE AUTHOR**

**Dennis Arrindell** is a health economist based in Curaçao whose work focuses on healthcare claims analytics, artificial intelligence, health economics and operational improvement. He is the creator of the Udemy course *Data Science for Healthcare Claims Data* and regularly explores practical approaches to helping resource-constrained healthcare organizations use data more effectively for decision-making.