Private claims data expose uneven type 2 diabetes care across Saudi Arabia
Saudi Arabia’s privately insured population with type 2 diabetes mellitus is large, clinically complex and increasingly visible inside national claims systems. A retrospective cohort study using the National Platform for Health Information Exchange Services followed 187,797 patients and offers one of the clearest real-world snapshots yet of how antidiabetic therapy is being used outside the clinic walls. The cohort covered claims from 1 September 2022 through 31 December 2024, with eligibility restricted to patients aged 15 years or older who had continuous insurance enrolment and at least one coded diagnosis of T2DM.
The signal is unmistakable: treatment is common, but persistence is uneven.
Patients had a mean age of 53.7 years, and the prevalence of T2DM in the insured population reached 47.4 per 1000 insured men and 43.8 per 1000 insured women. Those figures sit within a broader regional reality in which diabetes has become a dominant chronic disease burden, yet the claims data also reveal how management varies once treatment begins. Nearly one-third of initial regimens, 29.4%, used combination therapy from the outset, suggesting that many patients are already beyond the earliest stages of pharmacologic management when they enter the captured claims stream. In practical terms, that pattern usually reflects either greater baseline glycaemic severity or earlier escalation driven by prior therapeutic failure, comorbidity burden or clinician preference for multi-drug control.
Switching and escalation were common after therapy began
The clearest pressure points emerged after first-line therapy was already underway. Among biguanide monotherapy users, switching occurred in 9% and escalation in 7%, while 35% of sulfonylurea users required escalation. Those differences point to diverging real-world trajectories across drug classes. Biguanides remain the backbone of initial pharmacotherapy for many patients because they are familiar, inexpensive and generally well tolerated, but a nontrivial share still needs intensification when glycaemic control slips. Sulfonylurea users escalating at a markedly higher rate suggests either waning efficacy, inadequate durability, or the limitations of agents that stimulate insulin release without altering the underlying insulin resistance that drives much of T2DM.
A single line from the data captures the broader picture: therapy often moved, rather than held steady.
Discontinuation further complicated the treatment course. It was observed in 22% of users of combination therapies, 22% of SGLT2 inhibitor users and 19% of sulfonylurea users. In claims research, discontinuation can reflect many bedside realities at once: intolerance, cost-sharing, refill gaps, therapeutic substitution, nonadherence, or administrative fragmentation. Even without direct clinical notes, the pattern matters because diabetes outcomes depend heavily on persistence. Patients who stop and restart therapy repeatedly rarely achieve the level of metabolic stability needed to reduce microvascular and macrovascular risk over time. The similar discontinuation burden across combination therapy and SGLT2 inhibitor users suggests that regimen complexity and access constraints may be as important as pharmacology alone.
Monitoring improved after treatment, but gaps remained before initiation
Laboratory monitoring showed partial adherence to guideline-based testing rather than complete uptake. Sixty-one percent of patients had at least one glycaemic test during the 6 months before therapy initiation, and testing reached 73% after treatment began. Those numbers suggest that surveillance improves once a patient is already inside the treatment pathway, yet they also show that a substantial minority still begin or continue therapy without clear laboratory confirmation in the expected window. For a disease in which therapeutic decisions depend on trends in HbA1c, plasma glucose and related metrics, incomplete monitoring can blunt timely intensification and delay recognition of inadequate control.
The inclusion of HbA1c, plasma glucose and oral glucose tolerance testing in the analysis is clinically important because each test serves a different purpose. HbA1c gives a longer view of glycaemic exposure and is central to treatment adjustment, while plasma glucose testing can capture immediate dysregulation and oral glucose tolerance testing remains relevant in selected diagnostic settings. Claims data cannot tell whether every test result changed management, but they do reveal whether the system is positioned to detect poor control at all. In a chronic disease as common as diabetes, the absence of testing is itself a clinical event.
Insurance claims can now support population-level diabetes management
The study’s greatest methodological strength lies in the infrastructure behind it. Claims were linked to the Council of Health Insurance internal product dictionary through the NPHIES product and service code, allowing standardised product names and activity types to be retrieved from a national source. Structured Query Language queries and Python-based transformation were used for extraction and processing. In a setting where fragmented records can obscure treatment trajectories, this kind of harmonised data environment creates a workable foundation for surveillance, benchmarking and policy planning.
For Saudi Arabia, that matters because diabetes care is no longer only a question of prescribing the right drug. It also involves knowing whether patients are staying on therapy, whether escalation is happening early enough, and whether the laboratory system is capturing the biochemical feedback needed to guide care. Private insurance adds another layer, because access and utilization patterns may differ from public-sector cohorts. This analysis, built from privately insured beneficiaries, therefore fills a specific gap rather than making claims for the entire national population.
The study also carries implications beyond the Kingdom. Many countries in the MENA region are expanding digital claims and e-prescribing platforms, yet few have used them at scale to quantify real-world diabetes management with this level of detail. Saudi Arabia’s experience suggests that when codes are standardized and data streams are linked, treatment movement becomes visible: who starts, who escalates, who stops, and who gets monitored.
Persistent therapy gaps remain the central operational problem
The clinical lesson from the cohort is not that antidiabetic therapy is absent. It is that treatment is often unstable. Escalation rates were substantial, discontinuation was common, and monitoring left room for improvement both before and after therapy initiation. Even the medication possession ratio of 0.87 among insulin users, while relatively strong by claims standards, still leaves room for missed fills and interrupted access in a therapy where consistency is critical to preventing hyperglycaemia swings.
Saudi Arabia now has a national claims structure capable of supporting stronger chronic disease measurement, but the data also expose how much of diabetes care still depends on continuity that is easy to lose and difficult to reconstruct after the fact. The cohort ends in late 2024, and the next questions are operational rather than rhetorical: whether higher-fidelity monitoring can be paired with better persistence, whether escalation occurs early enough to prevent avoidable deterioration, and whether private insurance systems can be used to close the gaps now visible in routine practice.