Diabetes Management Skill

Comprehensive blood glucose monitoring and diabetes management to help control blood sugar and prevent complications.

Core Flow

code

User Input -> Identify Operation Type -> Extract Parameter Info -> Check Completeness -> [Need Supplement] Ask User
                                                      |
                                                   [Information Complete]
                                                      |
                                              Generate JSON -> Save Data -> Output Confirmation

Step 1: Parse User Input

Operation Type Recognition

Input Keywords	Operation Type	Description
record, glucose, bg	glucose_record	Log blood glucose
hba1c, a1c	hba1c_record	Log HbA1c
trend	trend_analysis	View glucose trend
tir	tir_analysis	View Time in Range
hypo, hypoglycemia	hypo_event	Log hypoglycemia event
screening, complication	complication_screening	Complication screening
target	target_view	View glucose targets
achievement	achievement_view	View achievement status
medication	medication_management	Medication management

Glucose Type Keywords

Input Keywords	Type Value	Target Range
fasting	fasting	4.4-7.0
postprandial, 2h	postprandial_2h	<10.0
bedtime	bedtime	6.0-9.0
random	random	-

Hypoglycemia Symptom Keywords

Input Keywords	Symptom
sweating	sweating
palpitations	palpitations
tremor	tremor
hunger	hunger
confusion	confusion
dizziness	dizziness

Complication Screening Types

Input Keywords	Screening Type
retina, 眼底, 视网膜	retinopathy
kidney, 肾脏, 肾	nephropathy
nerve, 神经, 神经病变	neuropathy
foot, 足, 足部	foot_complication

Retinopathy Severity

Input Keywords	Severity
none	none
mild	mild
moderate	moderate
severe	severe
proliferative	proliferative

Nephropathy Status

Input Keywords	Status
normal	normal
microalbuminuria	microalbuminuria
macroalbuminuria	macroalbuminuria

Step 2: Check Information Completeness

Glucose Record Required:

•Glucose value
•Glucose type (fasting/postprandial_2h/bedtime/random)

HbA1c Record Required:

•HbA1c value (%)

Hypoglycemia Event Required:

•Glucose value
•Symptoms (optional)

Complication Screening Required:

•Screening type
•Result status

Step 3: Interactive Prompts (If Needed)

Scenario A: Missing Glucose Type

code

Please select glucose measurement type:
- fasting (fasting blood glucose)
- postprandial (2-hour post-meal blood glucose)
- bedtime (bedtime blood glucose)
- random (random blood glucose)

Scenario B: Hypoglycemia Symptom Inquiry

code

Do you have any of the following symptoms?
- Sweating
- Palpitations
- Tremor
- Hunger
- Confusion
- Dizziness

Scenario C: Missing Complication Screening Parameters

code

Please provide detailed screening results:
- Retina: none/mild/moderate/severe/proliferative
- Kidney: normal/microalbuminuria/macroalbuminuria + UACR + eGFR
- Nerve: normal/abnormal
- Foot: normal/low_risk/high_risk + Wagner grade

Step 4: Generate JSON

Glucose Record

json

{
  "id": "glu_20250620070000001",
  "date": "2025-06-20",
  "time": "07:00",
  "type": "fasting",
  "value": 6.5,
  "unit": "mmol/L",
  "in_range": true,
  "target_range": "4.4-7.0"
}

HbA1c Record

json

{
  "date": "2025-06-15",
  "value": 6.8,
  "unit": "%",
  "target": "<7.0",
  "in_target": true,
  "change_from_previous": -0.3
}

Hypoglycemia Event

json

{
  "id": "hypo_20250618153000001",
  "date": "2025-06-18",
  "time": "15:30",
  "value": 3.4,
  "severity": "level_1",
  "symptoms": ["sweating", "palpitations"],
  "treatment": "glucose_tablets",
  "resolved": true
}

Complication Screening

json

{
  "retinopathy": {
    "status": "none",
    "last_exam": "2025-03-20",
    "next_exam": "2026-03-20"
  },
  "nephropathy": {
    "status": "microalbuminuria",
    "uacr": 45,
    "egfr": 78,
    "ckd_stage": "G2A2"
  },
  "neuropathy": {
    "status": "none",
    "monofilament_test": "normal"
  },
  "foot": {
    "status": "low_risk",
    "pulses_present": true,
    "ulcer": false,
    "wagner_grade": 0
  }
}

Complete schema definition: see schema.json.

Step 5: Save Data

•Read data/diabetes-tracker.json
•Update corresponding sections based on operation type
•Update statistics
•Write back to file

Execution Instructions

code

1. Identify operation type (record/hba1c/trend/tir/hypo/screening/target/achievement/medication)
2. Extract parameter information
3. Check required field completeness
4. If missing, ask user to supplement
5. Generate JSON compliant with schema.json
6. Save to data/diabetes-tracker.json
7. Output confirmation and medical recommendations

Example Interactions

Example 1: Record Fasting Glucose

code

User: record fasting 6.5

AI: Blood glucose measurement recorded
Time: 2025-06-20 07:00
Type: Fasting blood glucose
Value: 6.5 mmol/L
Status: Within target range (4.4-7.0)

Recommendation: Continue maintaining good blood glucose control

Example 2: Record Hypoglycemia Event

code

User: hypo 3.4 sweating

AI: Hypoglycemia event recorded
Time: 2025-06-18 15:30
Blood glucose: 3.4 mmol/L
Severity: Level 1 (mild)
Symptoms: Sweating, Palpitations

Hypoglycemia Management Recommendations:
━━━━━━━━━━━━━━━━━━━━━━━━━━
1. Immediately take 15g fast-acting glucose
   - 3-5 glucose tablets
   - 150ml juice or sugary drink

2. Re-check after 15 minutes

3. If still below 3.9 mmol/L, repeat step 1

4. After glucose normalizes, if >1 hour until next meal, eat small amount of long-acting carbohydrate

Example 3: Record HbA1c

code

User: hba1c 6.8

AI: HbA1c test result recorded
Value: 6.8%
Target: <7.0%
Status: On target

Change from last: -0.3% (improved)

Recommendation: Test HbA1c every 3 months

Example 4: View Achievement Status

code

User: achievement

AI: Blood Glucose Achievement Report
━━━━━━━━━━━━━━━━━━━━━━━━━━
HbA1c: 6.8% (target <7.0%) ✓ On target

Past 7 days glucose:
- Fasting: Average 6.2 mmol/L (achievement rate 85.7%)
- Post-meal 2h: Average 8.5 mmol/L (achievement rate 71.4%)
- Bedtime: Average 7.1 mmol/L (achievement rate 100%)

TIR: 72% (target >70%) ✓ On target
Hypoglycemia events: 1 (level 1)

Control Evaluation: Good blood glucose control, continue

For more examples, see examples.md.

Medical Safety Boundaries

Cannot Do:

•Provide specific medication dosage adjustment recommendations
•Directly prescribe or recommend specific medications
•Replace doctor's diagnosis and treatment decisions
•Predict disease prognosis or complication occurrence

Can Do:

•Provide blood glucose monitoring records and trend analysis
•Provide HbA1c tracking and achievement status
•Provide complication screening records and reminders
•Provide hypoglycemia event records and analysis
•Provide lifestyle recommendations and medical visit reminders

Emergency Medical Indications:

•Severe hypoglycemia (unconsciousness, coma)
•Diabetic ketoacidosis (nausea, vomiting, abdominal pain, deep breathing)
•Hyperosmolar hyperglycemic state (severe dehydration, confusion)
•Infection with fever and blood glucose >16.7 mmol/L