The study, conducted by Sungsu Oh and Hyunsu Lee of Pusan National University’s Department of Physiology together with neurosurgeons In Ho Han, Jae Il Lee, and Byung Kwan Choi of Pusan National University Hospital, was published in Medical & Biological Engineering & Computing. The work tackles a problem that has plagued health informatics for decades: electronic medical record systems are heterogeneous by design, shaped by local workflows, billing requirements, and clinical habits. When hospitals want to collaborate — whether for multi-site clinical trials, epidemiological surveillance, or artificial intelligence development — their data must first be normalized onto a common clinical ontology. SNOMED-CT, the Systematized Nomenclature of Medicine—Clinical Terms, is the most comprehensive such ontology in existence, but mapping free-text diagnoses to its precise concepts has traditionally required laborious manual coding or brittle rule-based systems.

The Korean team’s strategy begins with ClinicalBERT, a variant of the BERT transformer architecture that has been pretrained on clinical text, giving it an innate familiarity with the abbreviations, fragments, and idiosyncrasies of physician prose. Rather than treating diagnosis classification as a pure text-labeling task, the researchers framed it as a problem of geometry. They trained ClinicalBERT on electronic medical record data to produce latent representations — dense numerical vectors — for spans of text corresponding to diagnoses. In parallel, they encoded the fully specified names, or FSNs, of SNOMED-CT concepts: the ontology’s unambiguous, self-describing labels such as “myocardial infarction (disorder).” The goal of the fine-tuning stage was to pull these two families of vectors toward one another in the embedding space, so that the vector for a hospital’s idiosyncratic diagnosis text would land close to the vector for its correct SNOMED-CT concept.

The alignment was accomplished through mean squared error-based fine-tuning, a loss function that penalizes the squared distance between the EMR-derived diagnosis embeddings and their target SNOMED-CT FSN embeddings. This is a conceptually elegant move: instead of asking the model to memorize a lookup table of diagnoses, the researchers asked it to reshape its internal geometry so that semantic equivalence becomes spatial proximity. Once the alignment was complete, the resulting latent representations were fed into a downstream classification model that assigns each diagnosis span to one of 273 SNOMED-CT classes.

The performance numbers are striking. The proposed model achieved an accuracy of 0.934, a weighted F1-score of 0.923, and a macro-averaged F1-score of 0.823 across the 273 classes. The gap between the weighted and macro scores reflects a familiar reality of clinical data: some diagnostic classes appear far more frequently than others, and models naturally excel at common conditions while struggling with rare ones. A macro-averaged F1 above 0.82 nevertheless indicates that the model maintains strong performance across the board, not merely on the frequent diagnoses that dominate hospital records.

Perhaps the most thought-provoking finding, however, is what the fine-tuning did not do. Despite measurably improving semantic alignment in the embedding space, the fine-tuned model performed roughly on par with the base, unmodified ClinicalBERT when it came to the downstream classification task. Both remained competitive with dedicated biomedical entity-linking systems — SapBERT, which achieved an accuracy of 0.944 in the comparison, and BioSyn, at 0.931. This apparent paradox — better alignment, similar classification — is one of the study’s central insights. The researchers analyzed the latent representations directly, examining how diagnosis vectors were distributed before and after fine-tuning, and found that the fine-tuned model exhibited reduced similarity distances and more distinct class separation. Fine-tuning reduced the average Manhattan distance between paired representations by 36.2 percent and the average cosine distance by 53.1 percent, evidence that the embeddings had indeed converged toward their ontological targets.

In other words, the geometry of the latent space became cleaner and more semantically coherent, even if the final classification accuracy did not climb. The authors suggest that base ClinicalBERT, having been pretrained on large volumes of clinical text, already encodes much of the structure needed for discrimination, and that MSE-based alignment improves the interpretability and organization of the space without necessarily adding discriminative power. This trade-off between semantic alignment and discriminative performance is a subtle but important consideration for anyone designing medical NLP systems: a model whose embeddings cluster cleanly by concept may be more useful for retrieval, normalization, and cross-dataset integration even when its headline accuracy is unchanged.

The implications reach beyond a single hospital’s coding workflow. Medical concept normalization is a foundational enabling technology for the secondary use of health data — the practice of reusing clinical records for research, quality improvement, and public health. Studies have documented both the enormous opportunity and the serious technical and privacy barriers involved. Manual coding is expensive and inconsistent; rule-based systems fail when clinicians invent new abbreviations; and naive deep learning classifiers trained at one institution often collapse when deployed at another, where documentation habits differ. An embedding-based approach that explicitly anchors diagnoses to SNOMED-CT fully specified names offers a path toward models that transfer more gracefully, because the anchor points are institution-independent.

Privacy considerations loom large in this domain, and the Korean team’s approach has advantages here as well. Because the model operates on diagnosis spans rather than whole patient records, and because the data used to train it were de-identified retrospective records — a design approved by the Institutional Review Board of Pusan National University Hospital, with informed consent waived for the de-identified data — the pipeline is compatible with privacy-preserving deployment. The authors describe the method as having strong potential for scalable, privacy-preserving medical concept normalization in real-world clinical environments. Notably, the trained artifacts are embeddings and classifiers rather than raw patient data, meaning institutions could in principle share model components without sharing records — a property that aligns with the growing movement toward federated and privacy-conscious health AI.

The study also contributes to a rapidly maturing field. Automated medical coding has seen an explosion of deep learning approaches in recent years, with transformer models applied to ICD-10 and SNOMED-CT coding tasks across multiple languages and healthcare systems. Comparisons in this literature are notoriously difficult because datasets, class counts, and annotation standards vary widely; by benchmarking directly against SapBERT and BioSyn, two widely cited biomedical representation models, the Pusan National University team situates their result within a recognizable landscape. Their 0.934 accuracy sits within roughly one percentage point of SapBERT’s 0.944, achieved with a straightforward MSE alignment strategy rather than specialized contrastive pretraining.

There are caveats, as with any machine learning study in medicine. The model was trained and evaluated on data from a single hospital network, and while 273 SNOMED-CT classes is a substantial label space, it is a small fraction of the ontology’s several hundred thousand concepts. Rare diagnoses, ambiguous abbreviations, and multilingual clinical text all remain open challenges. The authors also emphasize that the observed tension between semantic alignment and classification performance deserves further study — future work may find alignment objectives that improve both properties simultaneously, or demonstrate that better-aligned embeddings pay off in downstream tasks beyond classification, such as cross-lingual normalization or retrieval-augmented clinical decision support.

Still, the work represents a meaningful step toward a long-promised vision: hospital data that can travel. When a diagnosis written in a Busan emergency department can be automatically and reliably expressed in the same standardized language as one written in a Baltimore clinic, the foundations for multi-institutional research, fairer AI training datasets, and population-scale health insights become dramatically more accessible. The Korean team’s results suggest that the key may lie not just in better classifiers, but in sculpting the hidden geometric spaces where machines understand medicine — one embedding at a time.

Atasoy’s recent publication in the September issue of MIS Quarterly brings to light the intricate relationship between electronic health record (EHR) technologies and insurance claim denials. The study argues that software interoperability, or the lack thereof, plays a critical role in the accuracy and success of healthcare claim submissions. This connection, if addressed properly, holds the potential to alleviate some of the financial burdens currently plaguing the healthcare industry.

The foundation of Atasoy’s research rests on the Health Information Technology for Economic and Clinical Health Act (HITECH) introduced in 2009. The legislation aimed to accelerate the adoption of electronic health record systems across U.S. healthcare providers, promoting a digital infrastructure as a pillar for reformative change. Although this policy succeeded in increasing EHR uptake, it inadvertently triggered interoperability issues owing to a diverse array of software vendors employed by different institutions.

This diverse software environment creates significant data inconsistencies and gaps when patient information is transferred or accessed across different systems. Such lapses often result in erroneous or incomplete submissions to insurance companies, thereby increasing the likelihood of claim denials. Atasoy’s work meticulously examines how these technological disparities translate into real-world financial consequences for medical providers.

Between 2000 and 2020, an eye-opening $745 billion worth of healthcare services were rendered without compensation, with claim denials representing a substantial portion of this uncompensated care. To dissect the role of healthcare information technology in these denials, Atasoy and her colleagues utilized a large dataset comprising over 19 million anonymized patient visits collected from 48 hospitals in Maryland. This dataset included detailed records of whether initial claims were approved or denied.

By mapping these claims against patterns of EHR adoption and software sourcing strategies, the researchers unveiled compelling evidence: hospitals relying on multiple software vendors experienced higher rates of claim denial compared to those standardizing on a single vendor’s system. A seemingly modest reduction from 1.18% to 1.03% denial rate was observed when a single vendor approach was employed, but at scale, this marginal improvement translates to millions of dollars saved and substantial administrative efficiencies gained.

The study further delved into the influence of physicians’ familiarity with EHR systems across their workplaces. It found that when doctors practiced within hospitals utilizing similar EHR platforms, claim denials declined, underscoring the critical role of user proficiency in navigating software systems. Familiarity promotes accurate clinical data entry and fosters compliance with insurance payer protocols, ultimately enhancing the integrity and acceptability of submitted claims.

Atasoy and her team highlight this alignment between user experience and technological uniformity as a powerful lever in addressing claim denials. By fostering standardized and user-friendly interfaces within EHR systems, healthcare organizations can minimize data entry errors and streamline the flow of clinical information critical for claims processing. This insight marks an important paradigm shift, positioning software management not just as a tool for digitization, but as an integral component of financial stewardship in healthcare.

As healthcare costs spiral ever upwards, a multifaceted approach is necessary to stem mounting expenditure. While technology is not a panacea, Atasoy’s findings suggest targeted improvements in EHR interoperability and user interface standardization could deeply impact the rate of denied claims and reduce downstream costs. This efficiency gain could alleviate the financial strain faced by both providers and patients, whose care and wallets are often caught in the crossfire of administrative inefficiencies.

Moreover, the research carries significant implications for federal healthcare policy. It advocates for a strategic emphasis on EHR certification processes that ensure interoperability and minimize vendor variability. Strengthening regulatory oversight in this domain could enforce higher standards for software usability and data accuracy—critical factors that determine whether an insurance claim sails through or gets inexplicably rejected.

The standardization agenda extends beyond software to the very design of user interfaces and workflows, which when optimized, can propel data accuracy to new heights. “Placing greater emphasis on usability and standardization of user interfaces and layouts,” the researchers argue, “can substantially reduce the extent of errors in claim processing.” Such improvements offer tangible benefits, including fewer surprise bills for patients and more predictable revenue streams for healthcare providers.

Atasoy’s investigation also illuminates the underappreciated connection between clinician experience with EHR applications and operational costs linked to denials. Healthcare systems that invest in training physicians within standardized EHR environments are better positioned to curb these costs by reducing inaccuracies and incomplete submissions. This dimension underscores an often-overlooked facet of healthcare IT: the human-technology interface.

The broader takeaway from this comprehensive study is clear: technology adoption in healthcare must transcend mere presence to embrace integration and alignment. When systems are fragmented and users struggle with disjointed applications, the resulting data errors can cascade into significant financial losses. Conversely, harmonized EHR ecosystems empower providers to offer seamless, accurately documented care, bolstering both clinical and financial outcomes.

At a time when the U.S. healthcare system is under intense pressure to contain costs while improving care quality, leveraging technology strategically could be a game-changer. Atasoy’s work not only quantifies the economic impacts of EHR interoperability but also charts a path toward systemic improvements that benefit providers, payers, and patients alike. By addressing the root causes of denied insurance claims through smarter software management, the healthcare industry can move closer to a more efficient, equitable future.

Subject of Research: People
Article Title: The Interplay Between Healthcare Information Technologies and Denied Claims
News Publication Date: Not provided (Article Publication Date: 1-Sep-2025)
Web References: http://dx.doi.org/10.25300/MISQ/2024/17469
References: Hilal Atasoy et al., MIS Quarterly, September 2025
Keywords: Medical facilities, Health care, Health and medicine