关于「TDS 不是健康分:电导率、离子构成与镁的证据强度」,《碱性水与 TDS 迷思》给出的定量结论是什么?
结论:第五章说明 TDS 是总量而非健康分,饮水镁与冠心病死亡的合并相对风险 0.89 的 E 值只有 1.496。
数据口径:书中参数为示意;全部计算由书中程序生成。 本书不是医疗建议,不宣称任何饮用水或饮品对任何疾病有效。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.89, 1.496
| pooled RR 0.89 (95% CI 0.79 to 0.99), I2 70.6% | |
| log-RR -0.1165, SE 0.0576, z -2.02, two-sided p 0.043 | |
| E-value for the estimate | 1.496 |
| E-value for the CI bound nearest 1 | 1.111 |
| Cochran Q implied by I2 with k | 10: 30.6 |
| water Mg mg/L mg in 2 L share of 350 mg | |
| RO / distilled | 0.1 0.2 0.001 |
| soft tap | 2.0 4.0 0.011 |
| hard tap | 20.0 40.0 0.114 |
| Mg-rich spring | 60.0 120.0 0.343 |
| brackish NaCl | 2.0 4.0 0.011 |
查看完整验算程序
# Chapter 5. p52: How strong is the evidence that minerals in water
# matter? Jiang et al. (2016) pooled 10 observational studies: highest
# versus lowest drinking-water magnesium, coronary death RR 0.89, 95% CI
# 0.79 to 0.99, I2 70.6 percent. We back out the standard error and the
# z statistic, and the E-value (VanderWeele and Ding 2017): the minimum
# strength of association an unmeasured confounder would need with both
# exposure and outcome to explain the estimate away. Then: what 2 L a
# day of each p51 water adds to a 350 mg/day magnesium reference
# (illustrative round reference value).
# 镁与冠心病死亡:从置信区间反推标准误、z 值与 E 值。
import math
RR, LO, HI = 0.89, 0.79, 0.99
def norm_sf(z):
return 0.5 * math.erfc(z / math.sqrt(2.0))
def evalue(rr):
r = 1.0 / rr if rr < 1.0 else rr
return r + math.sqrt(r * (r - 1.0))
se = (math.log(HI) - math.log(LO)) / (2 * 1.959964)
z = math.log(RR) / se
print("pooled RR {:.2f} (95% CI {:.2f} to {:.2f}), I2 70.6%".format(
RR, LO, HI))
print("log-RR {:+.4f}, SE {:.4f}, z {:+.2f}, two-sided p {:.3f}".format(
math.log(RR), se, z, 2 * norm_sf(abs(z))))
print("E-value for the estimate: {:.3f}".format(evalue(RR)))
print("E-value for the CI bound nearest 1: {:.3f}".format(evalue(HI)))
print("Cochran Q implied by I2 with k = 10: {:.1f}".format(
9.0 / (1.0 - 0.706)))
MG = [("RO / distilled", 0.1), ("soft tap", 2.0), ("hard tap", 20.0),
("Mg-rich spring", 60.0), ("brackish NaCl", 2.0)]
print("water Mg mg/L mg in 2 L share of 350 mg")
for name, mg in MG:
print("{:15s} {:7.1f} {:9.1f} {:13.3f}".format(
name, mg, 2 * mg, 2 * mg / 350.0))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《碱性水胃酸缓冲化学证伪、总溶解固体迷思与饮水电解质剂量母典》第 5 章「TDS 不是健康分:电导率、离子构成与镁的证据强度」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-192 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026192,
author = {Fan, Yafu},
title = {Alkaline Water against Gastric Acid, the TDS Myth and Electrolyte Dosing in Drinking Water},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}