儿童生长曲线分析技术深度解析：原理、实现与预警机制

# 儿童生长数据模型 (使用Pydantic)
from datetime import date
from typing import Optional
from pydantic import BaseModel

class ChildGrowthRecord(BaseModel):
    record_id: str  # 记录唯一ID
    child_id: str  # 关联儿童ID
    record_date: date  # 记录日期
    age_in_months: int  # 月龄
    gender: str  # 'M' or 'F'
    height_cm: float  # 身高(cm)
    weight_kg: float  # 体重(kg)
    head_circumference_cm: Optional[float] = None
    bmi: Optional[float] = None
    
    def calculate_bmi(self):
        if self.height_cm and self.weight_kg:
            self.bmi = self.weight_kg / ((self.height_cm / 100) ** 2)

# 标准生长曲线参考数据模型
class GrowthReferenceData(BaseModel):
    gender: str
    age_in_months: int
    p3_height_cm: float  # 3rd percentile
    p50_height_cm: float  # 中位数
    p97_height_cm: float  # 97th percentile
    sd_neg2_height_cm: float  # -2 SD
    sd0_height_cm: float  # 中位数(0 SD)
    sd2_height_cm: float  # +2 SD

# Flask/FastAPI 数据提交接口示例
from flask import Flask, request, jsonify

app = Flask(__name__)

@app.route('/api/v1/growth_records', methods=['POST'])
def add_growth_record():
    data = request.get_json()
    try:
        record = ChildGrowthRecord(**data)
        record.calculate_bmi()
        
        # 持久化到数据库
        # save_growth_record(record)
        
        return jsonify({
            "message": "记录添加成功",
            "record_id": record.record_id
        }), 201
    except Exception as e:
        return jsonify({"error": str(e)}), 400

from scipy.stats import norm

def calculate_percentile_and_zscore(
    gender: str,
    age_in_months: int,
    measurement_value: float,
    reference_data: list,
    measure_type: str
) -> dict:
    """
    基于WHO/CDC的LMS方法计算百分位和Z-score
    LMS公式：Z = [(X/M)^L - 1] / (L*S) (L≠0)
    """
    
    # 查找参考数据点(实际需插值计算)
    ref_point = next((r for r in reference_data
                   if r.gender == gender and r.age_in_months == age_in_months), None)
    
    if not ref_point:
        return {"percentile": None, "z_score": None, "message": "未找到参考数据"}
    
    # 使用LMS参数计算(简化示例)
    median_val = getattr(ref_point, f'sd0_{measure_type}_cm')
    sd_neg2_val = getattr(ref_point, f'sd_neg2_{measure_type}_cm')
    
    # 计算近似Z-score(实际需完整LMS参数)
    if sd_neg2_val and median_val:
        # 简化计算，实际应使用完整LMS公式
        z_score = (measurement_value - median_val) / (median_val - sd_neg2_val) * 2
        percentile = norm.cdf(z_score) * 100
        return {"percentile": percentile, "z_score": z_score}
    
    return {"percentile": None, "z_score": None, "message": "计算失败"}