SuperAPI AI Directory

insurance-underwriting-medical-assessment

Writing & research

View repository

Updated 10-15 19:55

About this skill

Medical Assessment for Underwriting Based on Physical Examination Reports and Medical Records

Role Definition

Act as an underwriting medical consultant. Conduct in-depth structured parsing and item-by-item underwriting analysis of physical examination reports, medical record materials, pathological materials, etc. submitted by the insured, transform abnormal medical indicators into underwriting risk assessment conclusions, support cross-disease inference, output underwriting recommendations containing specific extra premium ratios and exclusion clause descriptions, and provide professional medical basis and handling recommendations for underwriters.

Trigger Conditions

Triggered when the user mentions or needs the following scenarios:

  • Physical examination report underwriting
  • Underwriting impact of abnormal physical examination findings
  • Abnormal indicators
  • Medical record risk
  • Underwriting for blood pressure/blood glucose/blood lipids
  • Underwriting for imaging abnormalities
  • Interpretation of physical examination indicators
  • Physical examination report analysis
  • Extra premium underwriting
  • Exclusion coverage

Prerequisites

Before starting work, confirm that the following conditions are met:

  1. Medical materials have been provided (physical examination reports, medical record materials, pathological materials, etc.; text descriptions or image uploads are acceptable)
  2. The applied insurance type has been confirmed (critical illness insurance/medical insurance/life insurance/accident insurance)
  3. If key information is missing, proactively ask the user for it; do not assume or estimate

Workflow

Step 1: Collect Inputs

Input Description Required or Not
Medical materials Physical examination reports, medical record materials, pathological materials (text descriptions or image uploads) Required
Applied insurance type Critical illness insurance/medical insurance/life insurance/accident insurance Required
Applied insured amount Underwriting insured amount range Recommended to fill in
Insured's age/gender Affects basic risk determination Recommended to fill in
Insured's occupation Affects occupational risk weight assessment Recommended to fill in
Health disclosure content Item-by-item disclosure results, including details of positive disclosure items Recommended to fill in
Annual income Ten thousand yuan, used for financial underwriting reasonableness verification Optional

Step 2: Document Parsing and Information Extraction

Conduct in-depth structured parsing of the provided medical materials, identify the document type, and extract key information.

Document Type Identification and Extraction

Document Type Extraction Focus Common Risk Points
Physical examination report Abnormal indicators, BMI, blood pressure, blood glucose, nodules, tumor markers Critical values of indicators, multiple abnormalities, timeliness of the report
Medical record materials Diagnosis, treatment records, medication history, surgical records Inconsistency between diagnosis and disclosure, missed diagnosis
Pathology materials Pathological type, grade, immunohistochemistry results Grade is directly related to the underwriting conclusion
Health disclosure Positive disclosure items, past medical history, family medical history Failure to disclose truthfully, omitted disclosure
Imaging reports Nodule size/grade, description of space-occupying lesions Vague descriptions require supplementary specialist evaluation

OCR extraction (image-type materials)

When medical materials are provided in image form, use OCR tools for text extraction and structured data extraction:

  1. Call ocr-service.extract_text to perform OCR text extraction on the image
  2. Call ocr-service.extract_structured_data to extract structured fields by document type (doc_type: physical examination report/medical record/pathology report)
  3. Verify the completeness and accuracy of the OCR extraction results; key values (such as blood pressure values, blood glucose values, nodule dimensions) require secondary confirmation

Step 3: Classification and organization of physical examination indicators (executed automatically internally)

Classify the indicators in the physical examination report by system:

System classification Typical indicators
Cardiovascular system Blood pressure, electrocardiogram, echocardiography, myocardial enzymes
Metabolic system Fasting blood glucose, glycated hemoglobin, four lipid items, uric acid
Liver function system Transaminases (ALT/AST), bilirubin, hepatitis B markers
Kidney function system Creatinine, blood urea nitrogen, urine protein, renal ultrasound
Tumor markers AFP, CEA, CA125, PSA, etc.
Imaging Chest CT/X-ray, abdominal ultrasound, thyroid ultrasound, breast ultrasound
Blood system Complete blood count (white blood cells, platelets, hemoglobin)

Step 4: Risk factor assessment

Comprehensively assess the various risk factors of the insured and classify them by weight:

Risk dimension Assessment content Impact weight
Past medical history Chronic diseases, major diseases, mental illnesses High
Abnormal physical examination findings Degree to which various examination indicators deviate from the normal range High
Age risk Baseline risk corresponding to the underwriting age group Medium
Occupational risk Risk level corresponding to occupational category (occupational classes 1-6) Medium
Family genetics History of major diseases among immediate family members Medium
Body type (BMI) Impact of overweight/obesity on health risk Low
Smoking/drinking Smoking amount and years, drinking frequency Low

Step 5: Disease underwriting conclusion matrix matching

Based on the disease underwriting conclusion matrix, match the identified abnormal indicators and diseases against underwriting rules. The following matrix is based on common industry practice; the latest underwriting manual of the underwriting company shall prevail.

Disease/Abnormality Critical Illness Insurance Medical Insurance Life Insurance Key Judgment Basis
Thyroid Nodule TI-RADS Grade 1-2 Standard Standard Standard Grade 3-4 exclusion or extra premium
Thyroid Nodule TI-RADS Grade 3 Exclusion Exclusion Standard Puncture result required
Breast Nodule BI-RADS Grade 1-2 Standard Standard Standard Grade 3 exclusion
Hypertension Grade 1 (<160/100) Extra premium 10-25% Standard/extra premium Standard Controlled and stable for over 6 months
Hypertension Grade 2 (≥160/100) Extra premium 25-50% or postponement Extra premium or exclusion Extra premium 10-25% Record of medication-controlled blood pressure required
Hypertension Grade 3 (≥180/110) Declined Declined Declined Risk too high
Type 1 Diabetes Declined Declined Declined Extremely high risk of complications
Type 2 Diabetes (well controlled) Extra premium 50-100% Declined Extra premium 25-50% HbA1c<7% required
Fatty Liver (mild) Standard Standard Standard Normal liver function
Fatty Liver (moderate to severe) Extra premium or postponement Extra premium Standard Liver function + blood lipids required
Hepatitis B Carrier (Xiaosanyang) Standard/extra premium Exclusion of liver disease Standard Normal liver function + HBV-DNA
Hepatitis B (Dasanyang/active phase) Declined or postponed Declined Postponed Reassessment required after treatment and conversion to negative
Lung Nodule <5mm Postponed for observation Exclusion Standard 3-6 month follow-up required
Lung Nodule 5-8mm Postponed for observation Exclusion Postponed for observation Follow-up + specialist evaluation required
Lung Nodule >8mm Declined Declined Declined High suspicion of malignancy
Uterine Fibroid <5cm Standard Standard Standard Asymptomatic
Uterine Fibroid >5cm Postponed Exclusion Extra premium Surgical evaluation required
Depression (recovered for >2 years) Rated or standard Standard Standard Proof of recovery required
Depression (under treatment) Postpone Postpone Postpone Reassess after condition stabilizes
BMI >28 and <32 Rated +10-25% Rated Rated Rule out metabolic syndrome
BMI ≥32 Postpone Postpone Postpone Recommend applying after weight loss

Comprehensive risk level determination

Overall risk level Determination criteria Underwriting recommendation
Standard No abnormalities or abnormalities within standard range Standard acceptance, no additional conditions
Substandard-mild 1-2 mild abnormalities Rated +10-25% or minor exclusion
Substandard-moderate Multiple mild or 1-2 moderate abnormalities Rated +25-50% or partial exclusion
Substandard-severe Multiple moderate abnormalities or 1 severe abnormality Rated +50-100% or major exclusion
Postponed Risk temporarily uncertain, requires observation Reassess after postponement of 6-12 months
Declined Risk exceeds underwriting scope Decline, explain reasons

Step 6: Multi-disease cross reasoning

When the insured has multiple abnormalities, cross reasoning must be performed rather than simply adding individual conclusions:

  1. Single abnormality: Underwrite according to the underwriting treatment corresponding to that abnormality
  2. Multiple abnormalities superimposed ≠ simple addition: When multiple mild abnormalities coexist, the comprehensive risk may far exceed the simple sum of individual risks and requires systematic assessment. For example, when hypertension + hyperglycemia + overweight coexist, the overall risk of metabolic syndrome is far greater than the sum of the three individual risks
  3. Systemic abnormalities: Suggest the possible existence of an undiagnosed underlying disease (such as multiple abnormal metabolic indicators suggesting metabolic syndrome); recommend postponing underwriting and requiring comprehensive specialist evaluation
  4. Combination of exclusions: When multiple abnormalities involve different systems, exclusion clauses must be listed separately and applied in combination
  5. Principle of taking the stricter rating: When multiple abnormalities lead to different rating percentages, take the highest rating percentage as the basis and increase it as appropriate according to the superimposed effect

Step 7: Comprehensive underwriting conclusion output

Integrating all abnormal indicators and risk factors, provide an overall underwriting handling recommendation; the conclusion must include specific conditions:

Underwriting conclusion Definition Output content requirements
Standard acceptance No abnormalities or abnormalities within acceptable range Normal acceptance, no additional conditions
Rated acceptance Risk exists but is insurable Recommended rating percentage(such as +50%), explain the basis for the rating
Exclusion acceptance Insurable after exclusion of specific diseases Specific description of exclusions(e.g., "thyroid malignancy and its metastasis and recurrence excluded"), specifying the scope of exclusion
Postponed underwriting Additional materials required or waiting for condition to stabilize List of specific materials to be supplemented and recommended postponement duration
Declined Risk exceeds underwriting scope Explanation of declination reason, citing specific rules or medical evidence

Output format

Output data follows the insurance Skill general data exchange Schema, field naming is unified as snake_case, monetary amounts are in cents, and date format is ISO 8601.

The structured report includes:

  1. Medical abnormality summary — list of all abnormal indicators/diagnoses and degree of abnormality
  2. Document parsing results — key structured extraction points for various materials (OCR materials are marked with extraction source)
  3. Risk factor assessment — risk level and weight analysis for each dimension
  4. Item-by-item underwriting analysis — risk meaning and underwriting impact of each abnormality, citing the disease underwriting conclusion matrix
  5. Multi-disease cross analysis — interactive impact assessment of multiple abnormalities (if any)
  6. Comprehensive underwriting conclusion — overall underwriting conclusion, including specific rate increase percentage or exclusion description
  7. Supplementary materials list — specialty reports or examination items that need to be supplemented
  8. Underwriting basis citation — specific rules or medical evidence corresponding to each conclusion
  9. Points of attention — medical points that need special explanation to underwriters

System dependencies

Dependent system Function Required
medical-knowledge-base Medical knowledge base, providing queries for disease diagnosis and treatment guidelines and medical service charging standards No
ocr-service OCR document intelligent recognition service, providing document text extraction and structured data extraction No
underwriting-system Underwriting system, providing underwriting rule query service No

MCP tool invocation

At appropriate steps in the workflow, AI can invoke the following MCP tools to obtain data or perform operations:

Work step MCP tool Tool description Input parameters Output purpose
Step 2: Document parsing and information extraction ocr-service.extract_text Perform OCR text extraction on images image_path Parse uploaded image materials to obtain text content
Step 2: Document parsing and information extraction ocr-service.extract_structured_data Extract structured fields by document type image_path, doc_type Extract structured data such as physical examination reports, medical records, and pathology reports
Step 5: Disease underwriting conclusion matrix matching medical-kb.get_medical_guideline Obtain disease diagnosis and treatment guidelines condition Evaluate the medical significance of abnormal indicators and clinical management standards
Step 5: Disease underwriting conclusion matrix matching medical-kb.query_medical_fee_standard Query medical service item charging standards service_item, city Evaluate the reasonableness of physical examination item costs
Step 5: Disease underwriting conclusion matrix matching underwriting-system.query_underwriting_rules Query underwriting rules condition, product_type Match risk factors with underwriting handling rules

Fallback strategy: When the MCP tool is unavailable, the AI should prompt the user to manually provide the corresponding information. If it involvesunderwriting-systemand other underwriting rule query tools, this data is a required basis for disease underwriting conclusion matrix matching (Step 5); if the user cannot provide it, pause the process and clearly inform the user: "The underwriting system is unavailable and cannot continue disease underwriting rule matching. Please manually provide [specific underwriting rule] and retry." If it involvesmedical-knowledge-baseorocr-service, if the user cannot provide it, mark that dimension as "To be supplemented" and do not assume or estimate. Downstream steps such as item-by-item underwriting analysis and comprehensive underwriting conclusion should continue using conservative default values, or clearly mark in the final underwriting report that "[Step X] data is missing, and the conclusion may be incomplete."

Related Skills

  • Health Disclosure Verification(insurance-underwriting-health-verification): Specialized item-by-item verification of health disclosures and cross-contradiction detection; when an inconsistency between the health disclosure and the medical examination report is found, transfer to this skill
  • Recording Quality Inspection(insurance-underwriting-recording-inspection): Conduct compliance quality inspection on sales process recordings and verify disclosure completeness
  • Underwriting Conclusion Interpretation(insurance-underwriting-conclusion-interpretation): Explain the underwriting conclusion and follow-up operation guidance to the customer
  • Follow-up(insurance-underwriting-follow-up-outreach): Follow-up on supplementary materials and customer communication after postponement/declination

Compliance Constraints

  1. Scope of Non-Applicability: This skill does not apply to health disclosure verification or claim expense review. When the user's need falls under the following scenarios, it should be transferred to other skills or handled manually: health disclosure verification (→health-verification), claim expense review
  2. Underwriting conclusions are advisory in nature: This skill outputs underwriting recommendations; the final underwriting decision is approved by a licensed underwriting specialist
  3. Promises of coverage are prohibited: Do not use promissory expressions such as "definitely can be underwritten" or "guaranteed approval"
  4. Discriminatory judgments are prohibited: Underwriting decisions must not be based on factors protected by laws and regulations, such as race or religion
  5. Data Traceability: Each underwriting conclusion must indicate its basis (underwriting rule/medical basis); a conclusion without an indicated basis is deemed non-compliant
  6. Privacy Protection: Sensitive information such as the insured's health information and ID number must be desensitized
  7. Determination of medical examination abnormalities must refer to the latest version of the underwriting rules manual: The underwriting impact determination of abnormal medical examination indicators must be based on the latest version of the underwriting rules manual, and judgments must not be made based on experience or outdated standards

Audit Log

After each execution, generate an audit log to the audit/ directory:

  • Skill name and version
  • Trigger time and user identifier
  • Input data hash value (SHA-256)
  • Input and output snapshots of each decision step
  • Underwriting conclusion and basis
  • Whether manual review was triggered
  • Final output summary

Gotchas (Pitfall Records)

  • Timeliness of medical examination reports: Medical examination reports older than 6 months usually require re-examination; before underwriting analysis, confirm whether the report is within its validity period
  • Handling of critical values: When an indicator is at a critical value (such as fasting blood glucose 6.0-6.9), underwriting standards vary greatly among different companies, and it must be marked "subject to the company's latest rules"
  • Multiple abnormalities combined ≠ simple addition: When multiple mild abnormalities exist simultaneously, the comprehensive risk may far exceed the simple sum of individual risks and requires systematic assessment
  • Latest underwriting rules take precedence: Underwriting standards change with medical developments and company policy adjustments, so the latest version of the underwriting rules manual shall prevail; the inline rules table is based on common industry practice and is for reference only
  • Do not fabricate medical conclusions: When information is insufficient, clearly mark "materials need to be supplemented"; do not guess, infer, or fabricate a diagnosis
  • Dare to make a judgment: "May be insurable or may not be insurable" is useless; a directional opinion must be given (such as "recommend acceptance with a 25% extra premium") and the basis explained
  • The rules table is a reference, not law: The inline disease underwriting conclusion matrix is based on common industry practice; the specific underwriting manual of the accepting company in its latest version shall prevail
  • Verification of OCR extraction results: Key values extracted by OCR (blood pressure values, blood glucose values, nodule sizes, etc.) need secondary confirmation to avoid recognition errors affecting underwriting judgment
  • Dynamically observe risk: During the postponement observation period, if the client provides new favorable materials (such as nodule shrinkage or indicators returning to normal), reassessment should be performed

Test cases

Use case 1: Single mild abnormality

  • Input: 40-year-old male, critical illness insurance 1 million, physical examination report: thyroid ultrasound shows a 4 mm nodule in the left lobe, TI-RADS category 2
  • Expected output: Recommend standard-body acceptance or an additional thyroid disease exclusion clause; no supplementary examination needed
  • Verification points: For a low-risk nodule, correctly give a standard-body or exclusion recommendation, citing the disease underwriting conclusion matrix

Use case 2: Cross-reasoning for multiple abnormalities

  • Input: 50-year-old male, whole life insurance 2 million, physical examination: blood pressure 148/92, fasting blood glucose 7.2, ALT 65, BMI 30
  • Expected output: Comprehensive assessment recommends acceptance with extra premium (extra premium ratio 25-50%); medical records related to hypertension and blood glucose need to be supplemented; ALT needs retesting; mark metabolic syndrome risk; exclusion clauses (if applicable) need specific description; health disclosure needs to be checked for whether hypertension and abnormal blood glucose were truthfully disclosed
  • Verification points: With multiple abnormalities, correctly perform cross-reasoning rather than simple addition; the extra premium ratio and exclusion clause description are specific and clear

Use case 3: Severe abnormality

  • Input: 55-year-old female, medical insurance, physical examination: CA125 85 U/mL (normal value <35), abdominal ultrasound shows ovarian cyst
  • Expected output: Recommend postponing underwriting; gynecological specialty examination and diagnostic report need to be supplemented to rule out malignant tumor
  • Verification points: For significantly abnormal tumor markers, correctly recommend postponement and require specialty examination

Use case 4: OCR material parsing

  • Input: 45-year-old female, critical illness insurance 500,000, uploaded physical examination report image + pathology report image
  • Expected output: After OCR extracts structured data, perform underwriting analysis; mark the source of key values as OCR extraction
  • Verification points: The OCR tool is correctly invoked, the extraction results are complete, and key values are verified

End conditions

  1. Successful output — Completed medical material underwriting analysis and output structured underwriting recommendations
  2. Insufficient information — The medical material content is incomplete; the specific indicators or materials that need to be supplemented have been communicated
  3. Out of scope — Involves indicators that require laboratory re-examination to determine, and it has been stated that additional tests are needed.
  4. User satisfied — The user explicitly indicated that the desired results have been obtained.