IMO Health Quit Building the Medical Record - and Became the Dictionary Inside Nearly All of Them

The most important product in your doctor's computer may be the one nobody mentions. When a physician types “left ankle swelling after a fall,” the medical record must preserve what the physician meant, connect it to a stable clinical concept, and eventually make sense to billing codes, quality reports, analytics software, another hospital and perhaps an AI model. Those audiences do not naturally speak the same language. IMO Health makes the translation.

That is a wonderfully unshowy business. IMO Health, formerly Intelligent Medical Objects, maintains a clinically curated vocabulary and knowledge graph that sits beneath familiar electronic health record interfaces. It lets clinicians document in their own words while software maps those words to systems such as ICD-10-CM and SNOMED CT. Around that foundation, the company sells tools for cleaner problem lists, procedure documentation, surgical scheduling, data normalization, value sets, risk adjustment, denials prevention and clinical trials.

The reach is larger than the name recognition. IMO Health says every major U.S. EHR vendor uses its solutions and more than 740,000 healthcare professionals rely on them in clinical settings. The company maintains more than five million clinical terms. Its customers and partners span Epic, Oracle Health, MEDITECH and NextGen; hospitals such as Piedmont and MetroHealth; data businesses such as CuriMeta; and research teams, payers and health-tech developers.

01The first product was too much product

IMO Health's origin story begins somewhere between a medical school and outer space. In the early 1990s, a Chicago Medical School team worked on MEDAS, a differential-diagnosis tool requested by NASA for possible use aboard the space shuttle. Data entry was tedious, so the group built a history-and-physical writer. Glaxo later evaluated the early record system in Cook County's emergency room, and a team led by informatician Frank Naeymi-Rad formed IMO in 1994 to commercialize the work.

Then came the useful mess. Naeymi-Rad has said the company spent its first decade trying nearly everything, including rewriting the same electronic health record three times on three platforms. The broad ambition kept producing software, but not a crisp center. In 2004, his mentor, physician-informatician Charles Safran, pushed him toward the “hedgehog” idea: identify the one capability the company could do unusually well, cared deeply about, and could turn into an economic engine.

That piece was interface terminology, under development since 1995. It captured the “DNA” of the encounter in the clinician's language, then connected it to administrative and reference standards without flattening the meaning. Epic adopted the terminology in its model EHR in 2004. IMO later mapped it across SNOMED CT and ICD-9-CM, launched its formal Clinical Interface Terminology product, and moved deeper into the workflow.

02A translator with three audiences

The easiest way to understand IMO is as a three-way translator. Clinicians want expressive, familiar language. Machines want consistent concepts. Finance and regulation want recognized codes. A clinician should not need to pause an encounter to become a coding specialist, yet a hospital cannot run billing, population health or interoperability on poetic free text.

IMO Core handles the point-of-care end: diagnosis search, documentation and problem-list management. Core Procedure supplies terminology for procedures, labs and medications. Core Periop keeps surgical dictionaries and associated CPT codes aligned. Precision Normalize uses NLP to extract, standardize and enrich both structured records and free text. Precision Sets gives teams a way to build and maintain consistent cohorts and reporting definitions. IMO Studio is the SaaS doorway, while the broader IMO Health Platform packages the vocabulary, knowledge graph and clinical AI as infrastructure.

The company is different from a public code set or a simple crosswalk. Codes are designed for particular administrative or reference jobs; a clinician's idea can be more specific, contextual or messy. IMO's claimed advantage is the accumulated layer between them: decades of real-world phrasing, continuous clinical review, mappings across multiple standards and distribution already inside the workflow. Competing options include Wolters Kluwer's Health Language, Clinical Architecture, in-house terminology teams, direct use of SNOMED CT, and open-source NLP. Each can be sensible. IMO's pitch is that recreating its maintenance burden is the expensive part.

03How the plumbing became a platform

For years, the terminology wedge served providers and EHR vendors. In 2019, IMO reorganized its portfolio into Core for workflow terminology and Precision for data insight. In 2020 it launched Precision Normalize, its first cloud-native product. The COVID-19 emergency forced faster updates: the company created pandemic-specific concepts and gave related value sets to customers without charge. In 2023 it acquired Houston-based Melax Tech, adding biomedical NLP and AI expertise, then released Normalize with NLP, Precision Sets, Studio and Discovery for Problems.

The next expansion is outside the hospital's immediate workflow. Health-tech companies can call normalization and entity-extraction APIs instead of assembling their own clinical language stack. Payers can improve risk adjustment and coding review. Life-sciences and research organizations can use more precise cohorts, evidence extraction and patient identification. In 2025, IMO launched an EHR-integrated clinical-trial suite and named its first general manager for life sciences.

Those figures are company-reported outcomes from particular implementations, not universal promises. That qualification matters. The business model is enterprise software: terminology and data licenses, APIs, SaaS access, integration and support under negotiated contracts. Public list pricing is not available. The clearest public price tag belongs to the company itself. Thomas H. Lee Partners acquired a majority interest from Warburg Pincus in 2022 in a transaction that sources told Axios valued IMO above $1.5 billion, or more than 25 times an estimated $60 million of forward EBITDA. The deal price was not formally disclosed.

04The bit worth stealing

IMO's strategy is copyable even if its clinical corpus is not. It preserved the user's natural behavior and moved complexity backstage. It entered through a narrow, high-frequency job. It distributed inside systems customers already used. It kept updating an asset that becomes more expensive to reproduce with time. Only after that layer was trusted did it expand into adjacent workflows and new buyer groups.

This is also a story about what changed a founder's mind. It was not a sudden market forecast. It was a decade of evidence that the full EHR was too diffuse, followed by a mentor's demand for focus and a platform customer's validation. Epic's adoption showed that IMO could be more valuable as a component than as the whole application. The company stopped competing for the complete screen and became harder to remove from underneath it.

05The AI chapter still has to earn its keep

IMO now describes itself as AI-native. The sensible part of that claim is not that a large model suddenly replaces its old dictionary. It is the opposite: probabilistic software needs a deterministic clinical foundation. A model extracting diagnoses from notes is more useful when its output resolves to maintained concepts, knows whether a condition was negated or historical, and can show the code relationship. Melax added NLP machinery; the existing terminology supplies context and review.

That combination fits an awkward moment in healthcare. Hospitals want fewer denials, cleaner records and less clerical work. Developers want usable data without building a terminology department. Researchers want cohorts that mean the same thing across sites. All want AI, but none can afford casual hallucination in a clinical workflow. IMO sits in the middle of that demand, selling the connective tissue rather than the final diagnosis.

The danger is platform sprawl. A company celebrated for focus can dilute itself by chasing every use case attached to the phrase “clinical data.” Life sciences, revenue cycle, ambient AI, trials and payer workflows each have different buyers and proof requirements. The original lesson still applies: usefulness beats breadth. IMO's next decade will depend on whether its knowledge layer produces measurable results in those adjacent markets, not merely whether it can be described as a platform.

For now, the enduring idea is almost comic in its modesty. A doctor says what they mean. A hospital gets the code it needs. A machine receives a concept it can compute. IMO Health built a valuable company by standing between those sentences and refusing to treat them as identical.