Job Description
Prospective Clinical Reviewer - LLM Training
Eclat Health Solutions is seeking experienced Prospective Clinical Reviewers to support the training and refinement of our Large Language Model (LLM) for prospective risk adjustment workflows.
Reviewers will analyze longitudinal medical records to identify clinically supported suspected conditions that the LLM should learn to surface for prospective provider review. Using clinical judgment, Clinical Documentation Improvement (CDI) principles, TAMPER evidence, and knowledge of chronic disease processes, reviewers will identify potential conditions supported by patterns within the medical record and distinguish meaningful clinical suspects from historical, resolved, unsupported, or otherwise non-actionable findings.
Reviewer output will serve as clinical training and validation data for the LLM, helping improve the model's ability to identify appropriate suspected conditions while minimizing false-positive and clinically unsupported recommendations.
This role does not independently establish a patient diagnosis. Suspected conditions identified through the review process represent potential clinical opportunities for subsequent provider evaluation and confirmation.
-
Perform longitudinal medical record review to identify clinically supported suspected conditions for LLM training and validation.
-
Review diagnoses, medications, laboratory results, imaging, procedures, specialist documentation, historical clinical information, and other relevant evidence to identify potential suspects.
-
Determine whether available clinical evidence is sufficient for a condition to be surfaced as a prospective suspect.
-
Apply Clinical Documentation Improvement (CDI) principles when evaluating clinical documentation and potential suspect conditions.
-
Apply TAMPER (Treatment, Assessment, Monitor, Plan, Evaluate, Referral) principles when evaluating clinical support.
-
Identify and label appropriate suspected conditions according to established LLM training and clinical review guidelines.
-
Document the clinical evidence and rationale supporting each identified suspect.
-
Distinguish active or potentially active conditions from historical, resolved, ruled-out, unsupported, or otherwise non-actionable findings.
-
Identify potential HCC-related suspects while maintaining appropriate clinical and coding boundaries.
-
Review and validate LLM-generated suspects as model training and refinement progresses.
-
Identify false-positive suspects, missed suspects, and clinically unsupported model outputs.
-
Provide structured clinical reviewer feedback that can be used to improve model performance and suspect-surfacing accuracy.
-
Participate in clinical calibration to maintain consistency in how suspected conditions and supporting evidence are identified and labeled.
-
Collaborate with Clinical, Quality, Coding, Product, and LLM training teams as needed to clarify clinical criteria and review expectations.
-
Escalate ambiguous clinical scenarios, complex cases, and edge cases to Clinical/Quality Leadership.
-
Meet established accuracy, productivity, and turnaround time expectations for LLM training activities.
Candidates must meet all of the following:
-
Registered Nurse (RN/USRN) or MBBS graduate.
-
Minimum of 3 years of experience in Clinical Documentation Improvement (CDI), clinical documentation review, medical record review, utilization review, clinical quality, or a related clinical review function.
-
Experience reviewing medical records to identify documentation gaps, clinical indicators, and clinically meaningful findings.
-
Strong understanding of chronic disease processes and the clinical evidence used to support diagnoses.
-
Working knowledge of clinical documentation requirements and ICD-10-CM diagnosis terminology and principles.
-
Experience reviewing outpatient physician documentation and longitudinal medical records.
-
Ability to evaluate clinical information including diagnoses, medications, laboratory results, diagnostic findings, procedures, and specialist documentation.
-
Strong clinical reasoning, analytical, and written communication skills.
-
Experience using electronic medical records and clinical documentation systems.
-
Ability and willingness to complete project-specific training in prospective risk adjustment, HCC concepts, LLM training methodology, and suspect identification.
-
Prior exposure to risk adjustment or HCC coding concepts is preferred but not required. Candidates without direct risk adjustment experience must successfully complete project-specific HCC training and competency assessment.
-
Experience with outpatient or prospective CDI.
-
Experience supporting prospective clinical review, HCC validation, suspecting, or risk adjustment initiatives.
-
Knowledge of CMS-HCC risk adjustment methodologies.
-
Certified Risk Adjustment Coder (CRC) preferred but not required.
-
CDI certification such as CCDS or CDIP.
-
Experience collaborating with coding or risk adjustment teams.
-
Clinical quality audit or medical record abstraction experience.
-
Experience with US healthcare clients.
-
Experience using clinical review, CDI, coding, or risk adjustment platforms.
-
Experience with AI-assisted clinical documentation, abstraction, annotation, model validation, or coding workflows is a plus.
-
Strong clinical judgment and critical thinking.
-
Ability to interpret clinical indicators and patterns across complex longitudinal medical records.
-
Strong understanding of clinical documentation integrity and CDI principles.
-
Ability to identify clinically supported suspected conditions without independently establishing a diagnosis.
-
Ability to distinguish meaningful prospective suspects from historical, resolved, ruled-out, or unsupported conditions.
-
Ability to consistently apply defined clinical labeling and LLM training criteria.
-
Excellent attention to detail and documentation accuracy.
-
Strong written communication and ability to clearly document clinical rationale and supporting evidence.
-
Sound judgment regarding when clinical or coding escalation is appropriate.
-
Ability to work independently while consistently meeting quality, productivity, and turnaround expectations.
Selected reviewers will receive project-specific training and calibration related to:
-
Prospective suspect identification and clinical review methodology.
-
Risk adjustment and HCC fundamentals relevant to the review workflow.
-
LLM training, labeling, and validation methodology.
-
Identification of clinically supported suspected conditions.
-
Longitudinal medical record review and interpretation of supporting clinical evidence.
-
Differentiation of active or potentially active conditions from historical, resolved, ruled-out, and unsupported conditions.
-
TAMPER and supporting clinical evidence.
-
Identification of false positives, missed suspects, and clinically unsupported model outputs.
-
Documentation of suspect rationale and supporting clinical evidence.
-
Appropriate escalation of coding-specific, clinically ambiguous, and edge-case scenarios.
Reviewers will be expected to successfully complete project-specific HCC and LLM suspect-identification training and demonstrate competency through assessment and calibration prior to independent production. Reviewers must maintain established quality and productivity standards throughout the project.
The ideal candidate is an experienced RN/USRN or MBBS-trained clinician with at least three years of clinical documentation or medical record review experience and strong clinical reasoning skills.
They are comfortable navigating complex longitudinal medical records, interpreting clinical patterns and supporting evidence, and identifying suspected conditions that may warrant provider evaluation. They understand the distinction between identifying a clinically supported suspect for prospective review and independently establishing a diagnosis.
Prior HCC or risk adjustment experience is beneficial but is not required. The successful candidate has a strong clinical and CDI foundation and can apply that expertise to structured LLM training and validation following project-specific training and calibration.
