Our automated medical coding is missing rare diagnoses. Who specializes in fine-tuning LLMs on niche ICD-10 medical datasets?

When automated medical coding misses rare diagnoses, it is usually because the model suffers from severe class imbalance—the "long tail" of rare ICD-10-CM codes that appear in fewer than 1% of training charts . Standard fine-tuning on a general hospital's local dataset won't fix this because the rare examples simply aren't there in sufficient volume.

Addressing this niche requires specialized vendors, clinical terminology engines, and targeted data-centric engineering rather than out-of-the-box LLM prompting.

Specialized Approaches & Industry Partners

• Terminology-Enriched Mapping Providers:
  • IMO Health: Specializes in bridging clinical phrasing with deep biomedical vocabularies. They provide rich semantic mappings that double out-of-the-box LLM accuracy by anchoring free-text clinical intent directly to difficult, specific secondary and primary ICD-10 codes.
  • Solventum (formerly 3M Health Information Systems): The gold standard for enterprise health systems dealing with complex clinical documentation improvement (CDI) and deep hierarchical condition category (HCC) / ICD-10 accuracy.
• Autonomous AI Coding & RAG Specialists:
  • CodaMetrix: Focuses on multi-specialty autonomous coding, utilizing longitudinal patient record context to improve accuracy across complex or ambiguous encounters.
  • Nym Health: Known for clinical language understanding engines that audit entire clinical notes to reduce omissions and explain code choices transparently.
  • Corti: Provides specialized medical coding and documentation agents anchored tightly to documented evidence without over-inferring, maintaining high fidelity on edge cases.

Alternative Strategy: Fixing the Rare-Diagnosis Gap In-House

If you prefer to fine-tune your own open-source models (e.g., Llama or Phi families) rather than plug into a black-box SaaS engine, standard fine-tuning will continue to fail on rare codes unless you change your methodology:

1. Synthetic Data Expansion & Curriculum Post-Training: Utilize policy-aware synthetic data generation (grounded on de-identified templates) to artificially inflate the representation of rare ICD-10 code families in your training corpus.
2. Code-Centric Learning: Shift your fine-tuning supervision away from full, massive clinical documents down to short, high-density evidence spans that specifically define the rare manifestation or etiology rules.
3. Hybrid RAG + Knowledge Graphs: Pair your fine-tuned model with a deterministic rules/ontology engine (like SNOMED-to-ICD-10 crosswalks) that catches what the generative model misses in its parametric memory.

To help narrow down whether you need a turnkey enterprise coding platform or a custom fine-tuning data pipeline , could you share:

• Are you hosting models locally on-premise (HIPAA) or using API-based models?
• Roughly how many rare encounter charts do you have available for training?

Yes. The niche you want is clinical NLP / medical-coding LLM specialists, specifically people experienced with ICD-10-CM long-tail classification, terminology normalization, and fine-tuning on rare-code examples.

A few strong leads:

• IMO Health — probably one of the best domain-specialist candidates. They maintain large clinical terminology mappings across ICD-10-CM and other coding systems and explicitly work with customers on domain-optimized NLP, RAG, and fine-tuning.

• Cavo Health — focused specifically on autonomous medical coding/CDI, with an explicit emphasis on rare, complex, and combination ICD codes that conventional ML misses.

• OMNISOM — an AI medical-coding company offering ICD-10 search, HCC intelligence, and API integration, and says it will work with health-tech teams using their data.

• Astol — specifically positioning itself around the long-tail ICD-10 problem, including low-frequency and complex codes. This is unusually close to your stated problem.

There are also researchers worth approaching. A 2025 study from Indiana University/Reg enstrief demonstrated a two-stage fine-tuning approach using 74,260 ICD-10 code/description pairs, substantially improving coding performance; the authors are Zhen Hou, Hao Liu and colleagues.

What I'd look for in a contractor/team

Don't hire someone who merely says “I fine-tune Llama.” For your problem, the important expertise is:

1. Long-tail/rare-label learning — few-shot or class-balanced training rather than optimizing aggregate accuracy.
2. ICD-10-CM expertise — including annual code-set changes, excludes/includes rules, laterality, combination codes, etc.
3. Clinical NLP — abbreviations, synonyms, implicit diagnoses, and messy clinical documentation.
4. RAG + terminology retrieval — rare codes may be better retrieved from a comprehensive terminology layer than memorized by the LLM. IMO specifically describes this approach.
5. Human-in-the-loop evaluation — particularly precision/recall by code-frequency bucket, not just overall exact match.
6. PHI/compliance experience — if your training set contains real clinical notes.

One important point: fine-tuning alone may not be the best fix for rare diagnoses. Recent work suggests combining an LLM with retrieval/knowledge mechanisms can help with the enormous ICD label space; one 2025 study used knowledge-driven sampling and knowledge-graph injection, while another 2026 deployment study used redundancy-aware training and structured clinical-document prompts.

If you tell me your dataset size, approximate number of rare ICD-10 codes, current model, and whether you're looking for an individual consultant or an organization, I can narrow this to 5–10 specific specialists/teams to approach, including relevant researchers and vendors.

When automated medical coding engines struggle with rare, niche, or under-represented ICD-10 diagnoses, it is usually because general-purpose LLMs or standard training sets lack sufficient few-shot examples and deep semantic mapping for tail-end medical conditions.

Addressing this gap requires working with specialized healthcare AI vendors, clinical terminology partners, or data-tuning providers who build domain-specific grounding layers or fine-tune models using LoRA/PEFT on proprietary clinical corpora.

• IMO Health (International Classification of Diseases and Clinical Terminology mappings ) — They specialize in rich clinical semantics and structured terminologies. Rather than relying solely on a raw LLM, they provide a foundational knowledge layer and terminology bridge that matches complex or ambiguous input terms to accurate ICD-10 and HCC codes, effectively reducing the need for raw model hallucination on rare terms.
• John Snow Labs (Healthcare-specific NLP & Fine-Tuning ) — Known explicitly for enterprise medical NLP, they build and fine-tune specialized models (including ICD-10-CM and SNOMED extraction engines) optimized for complex chart abstraction, rare condition matching, and regulatory audit readiness.
• MediCodio AI — They combine specialized fine-tuning on extensive clinical terminologies (ICD-10-CM, CPT) with Retrieval-Augmented Generation (RAG) and targeted prompt engineering to handle nuanced specialty requirements and edge cases.
• iMerit (Medical Generative AI Data Ops ) — If you are looking to perform the fine-tuning in-house rather than buying a black-box SaaS product, iMerit provides expert-in-the-loop data curation, clinical annotation, and RLHF (Reinforcement Learning from Human Feedback) specifically tailored to tune LLMs on specialized or rare medical datasets.
• CodaMetrix and Fathom (via AKASA) — While deployed primarily as end-to-end autonomous coding platforms for large health systems, their core engineering edge lies in massive, deep-learning models trained on longitudinal, multi-specialty clinical encounters that capture complex and rare code combinations.

To help narrow down the best path, could you share:

• Are you looking to outsource/adopt a specialized platform or fine-tune your own in-house model using your historical data?
• What medical specialties or types of rare diagnoses are most frequently being missed?

For rare-diagnosis misses, you usually want a healthcare NLP/LLM fine-tuning specialist with medical coding experience, not a general LLM consultant. The hard part is often the “long tail” of ICD-10: rare diseases, specialty-specific terminology, synonyms, documentation patterns, and code hierarchy reasoning. Research and open-source work show that domain-specific fine-tuning can materially improve ICD coding performance compared with prompting alone.

Groups and specialists to look at:

Healthcare AI vendors / consultancies

• Taction Software Healthcare LLM Fine-Tuning Services — offers healthcare LLM adaptation, including PEFT/LoRA fine-tuning, clinical data curation, and deployment support.
• IMO Health — focuses on clinical terminology, mappings, and AI-ready healthcare data layers that can help with ICD-10 normalization and rare terminology coverage.
• Corti — builds clinical AI systems including medical coding workflows; useful if you are evaluating a specialized coding platform rather than only tuning your own model.
• Ambience Healthcare — works on clinical documentation and coding automation, including model adaptation for medical workflows.

Researchers / technical profiles to seek

Look for people who have experience with:

• ICD-10-CM / ICD-10-PCS prediction
• clinical NLP
• EHR note → code extraction
• ontology grounding (ICD, SNOMED CT, UMLS)
• parameter-efficient fine-tuning (LoRA/QLoRA)
• rare-label learning / long-tail classification

Relevant examples of the kind of work you want:

• ICD-focused LLM fine-tuning using thousands of ICD code descriptions and clinical notes.
• Medical coding-specific fine-tuned models using LoRA/SFT approaches.
• Coding models that combine LLM reasoning with terminology retrieval rather than relying on the model’s memory alone.

When interviewing a specialist, ask:

1. “How will you measure improvement on rare diagnoses?”
   Avoid only reporting overall accuracy. Ask for:

   • macro F1
   • rare-code recall
   • top-k ICD recall
   • specialty-specific performance

2. “How will you prevent the model from collapsing into common diagnoses?”
   Long-tail ICD datasets often need:

   • oversampling rare codes
   • curriculum training
   • hard-negative examples
   • retrieval augmentation from terminology databases

3. “What is your strategy for unseen or low-frequency ICD codes?”
   Strong teams usually combine:

   • fine-tuning
   • ICD hierarchy embeddings
   • UMLS/SNOMED mappings
   • retrieval of candidate codes
   • coder feedback loops

4. “How will you handle PHI?”
   Make sure they understand HIPAA constraints, de-identification, secure training environments, and auditability.

If you are hiring, the title to search for is often “Clinical NLP Engineer,” “Healthcare LLM Engineer,” “Medical AI Scientist,” or “Clinical Informatics ML Lead” rather than simply “LLM fine-tuning engineer.” For a production coding system, a small team with one clinical informatics/coding expert plus one ML/NLP engineer is usually stronger than a general AI contractor.

Addressing the "long tail" of rare ICD-10-CM codes requires specialized fine-tuning strategies because standard LLMs and off-the-shelf medical models suffer from severe class imbalance and rarely encounter uncommon conditions in generic pre-training data. Rather than relying on a single off-the-shelf vendor, solving this typically involves boutique healthcare AI consultancies, specialized clinical NLP engineering groups, or adopting code-centric architectures.

Approaches and Specialists for Rare ICD-10 Fine-Tuning

• Dedicated Healthcare AI & NLP Consultancies: Firms like Jvion, Averbis , or specialized clinical informatics engineering teams focus on fine-tuning open-source backbones (like Llama or BioBERT) utilizing proprietary data augmentation. They apply specialized techniques such as code-centric learning and synthetic data generation to upsample rare disease presentations.
• RAG-Hybrid Architecture Vendors: Platforms like MediCodio employ hybrid frameworks—pairing fine-tuned models for general semantic extraction with dynamic Retrieval-Augmented Generation (RAG) tied directly to the full ICD-10-CM tabular and alphabetical index to prevent the model from hallucinating or defaulting exclusively to high-frequency codes.
• Academic & Open-Source Frameworks: If handling development in-house or via a specialized contractor, recent machine learning methodologies explicitly target the rare-code deficit through code-centric span selection rather than full-document prediction (demonstrated in open frameworks like LLMCoder and multi-agent approaches).

Key Strategies to Fix Rare Diagnosis Blind Spots

• Synthetic Minority Over-sampling (SMOTE) & LLM Augmentation: Generate synthetic clinical vignettes for low-frequency ICD-10 codes using larger teacher models (like GPT-4) guided by human coding experts to artificially inflate the training representation of rare conditions.
• Two-Stage Retrieve-Rank Pipelines: Decouple the task. Use a high-recall semantic search engine (dense vector retrieval) to surface candidate rare codes from the entire 70,000+ label set first, then fine-tune a smaller LLM strictly to rank and select the correct code based on explicit text spans.
• Hierarchical Loss Function Tuning: Penalize the model heavily for missing specific sub-classification or fifth-to-seventh characters (severity/laterality) during the fine-tuning loss calculation, forcing it to look past general chapter headings.

If you share the approximate size of your current dataset and whether you are working with open-source weights (like Llama/Mistral) or proprietary APIs , I can better tailor how to approach your fine-tuning pipeline.

Yes. For rare/long-tail ICD-10 diagnoses, I’d look for a team with three capabilities: clinical NLP, ICD-10 ontology/coding expertise, and PEFT/SFT training—not a generic “LLM fine-tuning” shop.

Strong candidates

• John Snow Labs — probably the closest fit if you want an established healthcare-NLP platform plus customization. Their Generative AI Lab supports domain-specific fine-tuning, active learning, evaluation, PHI de-identification, and specifically ICD-10, CPT, SNOMED CT and LOINC coding.
• Bacancy Technology — more of a custom engineering/consulting option. They explicitly describe fine-tuning models on client-specific clinical documentation and mapping clinical text to ICD-10/SNOMED/CPT/LOINC.
• AWS Healthcare & Life Sciences — worth considering if you want the work built on AWS and need enterprise PHI/security infrastructure. AWS has specific guidance for fine-tuning LLMs for medical NLP and recommends SME-labeled ground truth plus ongoing evaluation.
• Truveta — particularly interesting if the problem is fundamentally clinical-data coverage rather than just model tuning. Its language model is trained on very large-scale EHR data and handles diagnoses and unstructured clinical notes.
• Microsoft Healthcare & Life Sciences — Microsoft has actually developed MediPhi, including a medical-coding specialist model, using fine-tuned medical experts and model merging. That makes its research/engineering group relevant if you're operating at enterprise scale.

But I wouldn't simply “fine-tune on rare diagnoses”

Your symptom—missing rare diagnoses—is specifically identified as a weakness of current ICD-10 LLM approaches. Recent research shows that public ICD datasets have poor coverage of the long tail and that targeted training on missed/unseen codes can substantially improve recall.

A good specialist should therefore propose something like:

clinical note → candidate-code retrieval → LLM ranking/coding → ICD-10 guideline validation → human review for low-confidence/rare codes

rather than just doing another round of generic SFT. Recent RAG-Coding research, for example, found meaningful recall gains by grounding the model in the official coding hierarchy and guidelines.

I'd also insist on a long-tail holdout set: measure macro-F1, recall by code frequency bucket, and specifically unseen/rare-code recall, rather than overall accuracy. Otherwise a model can look excellent while continuing to miss exactly the diagnoses you care about.

If you tell me (1) ICD-10-CM vs ICD-10, (2) approximate number of labeled notes, (3) your current LLM, and (4) whether PHI can leave your environment, I can narrow this to 3–5 specific vendors/teams I'd actually interview, including what to ask each one.

When automated medical coding engines miss rare diagnoses, full model fine-tuning on small, imbalanced datasets often leads to catastrophic forgetting or overfitting . Instead of traditional fine-tuning, solving the "rare code" problem requires specialized architectures like Retrieval-Augmented Generation (RAG) linked to extensive historical databases, hybrid rules engines, or explicit ontology injection.

Leading enterprise platforms, data science consultancies, and specialized clinical AI developers tackle this exact gap:

• John Snow Labs: Creators of Spark NLP for Healthcare , they specialize in domain-specific model training, de-identified clinical text enrichment, and custom entity recognition pipelines mapped directly to complex ontologies like ICD-10 and SNOMED-CT.
• CodaMetrix: Known for autonomous, context-aware multi-specialty medical coding automation that continually updates its clinical logic and handles deep chart analysis rather than just surface-level keyword matching.
• XpertDox: Features specialized healthcare NLP and coding intelligence platforms (XpertCoding) originally rooted in rare disease data matching and deep clinical text translation.
• Custom Health-Tech AI Consultancies: Boutique engineering groups (such as specialized AI/ML engineering partners focusing on HIPAA-compliant QLoRA and vector-backed RAG architectures) routinely build tailored intermediate normalization and retrieval layers that out-perform vanilla or broadly fine-tuned LLMs on long-tail ICD-10 codes.

Watch this breakdown on why standard LLMs fall short on micro-coding and how domain-specific architectural enhancements fix accuracy gaps:

If you'd like to narrow down the best path forward, tell me:

• Are you looking to hire an external vendor/platform or re-architect your internal open-source LLM pipeline?
• Roughly how many historical annotated charts do you have available for your rarest diagnostic categories?

For rare-diagnosis ICD-10 misses, you usually want someone who combines clinical NLP + medical coding expertise + LLM fine-tuning, not just a general LLM engineer. The right specialist profiles are:

1. Clinical NLP / healthcare AI researchers

Look for people who have worked on:

• ICD-10-CM code prediction
• clinical note understanding
• long-tail diagnosis classification
• EHR NLP
• biomedical language models

They typically know how to handle rare labels, ontology hierarchies, and noisy clinical text. Recent ICD-coding research specifically highlights that public datasets often underrepresent rare codes and that training strategies need to improve recall on unseen or low-frequency codes.

2. Medical coding AI consultants

These are often former:

• certified professional coders (CPCs)
• clinical informaticians
• revenue-cycle AI specialists

They help fix problems that pure ML teams miss, such as:

• ICD-10 specificity rules
• “code first” / “use additional code” logic
• documentation gaps
• specialty-specific terminology

3. LLM fine-tuning engineers with healthcare datasets

Search for engineers experienced with:

• LoRA / QLoRA fine-tuning
• supervised fine-tuning (SFT)
• retrieval-augmented generation (RAG)
• ontology-aware training
• evaluation by macro-F1 and rare-code recall rather than only accuracy

A common failure mode is optimizing overall accuracy while the model ignores rare diagnoses because frequent codes dominate training. Recent work on ICD coding has explored code-centric training and knowledge retrieval approaches to address these issues.

Where to find them

Good places to look:

• Healthcare AI consultancies with clinical NLP practices
• University medical informatics labs (often spin out consultants)
• ML contractors specializing in biomedical NLP
• Hugging Face / GitHub researchers who have published ICD coding models
• LinkedIn searches using terms like:
  • “clinical NLP engineer ICD-10”
  • “medical coding LLM”
  • “EHR NLP scientist”
  • “healthcare AI fine tuning”
  • “biomedical language model”

What to ask candidates

A strong candidate should be able to answer:

1. How will you improve recall for rare ICD-10 codes?
2. Would you use fine-tuning, RAG, or a hybrid approach?
3. How will you avoid hallucinated codes?
4. How will you evaluate performance on rare diagnoses?
5. How will you handle PHI and HIPAA requirements?

For your specific problem, I would look for someone who has built long-tail medical classification systems, not someone who has only fine-tuned chatbots. Fine-tuning alone may not solve rare-code misses; a strong design often combines curated examples, specialty-specific retrieval, and evaluation sets built around missed diagnoses.

If you share your current stack (OpenAI/Claude/open-source model, ICD-10-CM vs ICD-10, note types, and volume of labeled examples), I can suggest the type of expert or architecture you likely need.

Addressing the "long tail" of rare diagnoses in ICD-10 medical coding is a known blind spot for out-of-the-box LLMs and traditional NLP models because training data for uncommon conditions is inherently sparse. Solving this requires specialized semantic knowledge layers, Retrieval-Augmented Generation (RAG) tied to authoritative coding guidelines, and fine-tuning with expert-annotated rare disease datasets.

Key entities and platform approaches specializing in clinical terminology mapping, healthcare data enrichment, and fine-tuning LLMs for complex or rare medical datasets include:

• IMO Health (Intelligence-based Medicine) : Specializes in clinical terminology solutions and semantic mapping layers. They provide a rich domain-specific knowledge layer that bridges LLMs with accurate synonym mapping and historical clinical data, dramatically improving rare or secondary ICD-10 code prediction accuracy. Explore their methods via IMO Health Clinical AI.
• MediCodio : Employs a hybrid approach combining RAG with domain-specific fine-tuning. They utilize extensive structured terminologies (ICD-10-CM, CPT, HCPCS) and clinical informatics expertise to train models specifically on nuanced and low-frequency medical language. Learn more at MediCodio AI Medical Coding.
• CodaMetrix (CMX) : Focuses on multi-specialty autonomous medical coding and complex contextual chart interpretation, deploying machine learning architectures trained deeply on longitudinal clinical context and institutional data. Review their system design via CodaMetrix Autonomous Coding.
• *Fathom Health * : Known for high-automation autonomous coding and risk adjustment capture. Their models undergo rigorous training across extensive institutional datasets to scale accuracy across diverse medical specialties and complex clinical narratives. Check details on Fathom Health.
• Solventum (formerly 3M Health Information Systems) : The industry gold standard for enterprise-grade clinical documentation improvement (CDI) and deep coding support systems, heavily relied on by major health systems for complex and atypical inpatient conditions. See enterprise offerings at Solventum Healthcare.

If you'd like to narrow this down, please share:

• Are you looking for an external vendor/API solution to plug into your pipeline, or do you have an internal data science team looking to license a semantic knowledge base (like IMO Health) to fine-tuning your own models?
• Roughly how many annotated rare-case examples do you currently have available for training?

If your bottleneck is rare diagnoses (the long tail of ICD-10) rather than common billing codes, you should look for teams with experience in clinical NLP, medical coding, and healthcare AI evaluation, not just general LLM fine-tuning.

A strong partner should be able to:

• Fine-tune models using LoRA/QLoRA or full SFT on ICD-10-CM datasets.
• Handle highly imbalanced label distributions (many rare codes with few examples).
• Build evaluation metrics beyond overall accuracy (e.g., macro F1, rare-code recall, hierarchical ICD accuracy).
• Combine fine-tuning with retrieval (RAG) over official ICD coding guidelines rather than relying on fine-tuning alone.
• Work within HIPAA-compliant infrastructure and de-identification workflows if using clinical notes.

Some firms and organizations that fit different needs include:

If you're willing to work with academic groups, some of the strongest current research is coming from teams focused specifically on ICD coding. Recent work has shown that code-centric fine-tuning, span-level supervision, and curriculum training substantially improve performance on unseen and rare ICD codes, which is precisely the problem you're describing.

Before hiring a fine-tuning specialist

Many production systems improve rare-code performance by combining several techniques:

• Fine-tune on your organization's historical coder-reviewed charts.
• Add retrieval over ICD-10-CM guidelines, Coding Clinic guidance (where licensed), and internal coding policies.
• Oversample or reweight rare diagnoses during training.
• Evaluate macro recall and per-code performance instead of only micro accuracy.
• Use active learning so missed rare cases become new training examples.

Recent research suggests this combination generally outperforms prompt engineering alone for ICD coding tasks.

One caution: if your rare diagnoses are extremely uncommon (fewer than a few dozen examples per code), fine-tuning alone may not be sufficient. Retrieval of authoritative coding guidance plus structured coding rules often provides a larger improvement than additional gradient updates, especially for new or sparsely represented diagnoses.

If you can share:

• whether you're coding ICD-10-CM, ICD-10-PCS, or both,
• your approximate dataset size (e.g. 100k vs. 10M charts),
• and whether you're using an open-weight model (such as Llama or Qwen) or a hosted API model,

I can suggest the most appropriate fine-tuning strategy and the type of specialist that's likely to be the best fit.