追加：上了 50 岁培养一个爱好 · 核实记录（2026-11-04）
任务来源：读者 issue #143。原文：「能攒下钱的人（非力工）需要一个长期、稳定、健康的爱好以避免陷入精神焦虑、无所事事、行差踏错。例如乐器、钓鱼，或者其他输出能力如写文章、做视频。」没有附来源。
原有覆盖：全书没有讲爱好的条目（grep「爱好」「乐器」「艺术活动」只在第 11 节出现一次，说的是「业余爱好者」测漏洞）。最近的是第 22 节第 6 条（运动抗抑郁）、第 9 条（定期和人见面）和第 3 节第 18 条（情绪低落时先做）。Fancourt 2020 报告读书、社交、运动的变化只解释了爱好与抑郁关联的约 7%，所以这条不算运动或社交条目的换个说法。
四问：① 普通读者都有闲暇，会碰上。② 「有个爱好好」近乎常识，但读者不知道的是证据有多弱：差别只有约 0.1 个标准差，两项合唱分组试验在抑郁上结果不一，证据只覆盖 50 岁以上。这条的价值一半在把预期压下来。③ 动作是新的，全书没有。④ 有期刊原文。
落点：第 22 节末尾新增第 13 条，节首引言补一句指路。本节不到 20 条，没有节首导览。
关键数字
| 文献 | 复核 | 数字与原文 |
|---|---|---|
| Mak HW 等 (2023). Nature Medicine 29(9):2233-2240. https://doi.org/10.1038/s41591-023-02506-1（Europe PMC 摘要与 PMC10504079 全文 XML） | 是 | 「We harmonized measures of hobby engagement and multiple aspects of mental wellbeing across 16 nations represented in five longitudinal studies (N = 93,263). Prevalence of hobby engagement varied substantially across countries, from 51.0% of Spanish respondents to 96.0% of Danish respondents.」「having a hobby was associated with fewer depressive symptoms (pooled coefficient = -0.10; 95% confidence intervals (CI) = -0.13, -0.07), and higher levels of self-reported health (pooled coefficient = 0.06; 95% CI = 0.03, 0.08), happiness (pooled coefficient = 0.09; 95% CI = 0.06, 0.13) and life satisfaction (pooled coefficient = 0.10; 95% CI = 0.08, 0.12).」「little variance in findings was explained by country-level factors (<9%)」。方法段：「we harmonized and recoded all variables, and standardized the outcome variables. We created a binary indicator of hobby engagement (yes, no)」，所以系数是标准差单位。先后顺序：在 ELSA 上用 Arellano–Bond 法，「hobby engagement was still associated with subsequent changes in depressive symptoms (coefficient = −0.38; 95% CI = −0.63, −0.12)」；讨论段「these associations were not merely the result of good psychological health predicting hobby engagement. In actuality, the relationship between hobbies and mental wellbeing is likely bidirectional」。局限段：「Because of the use of observational data, causality cannot be established even with sophisticated longitudinal data analysis modeling.」 |
| Fancourt D, Opher S, de Oliveira C (2020). Psychother Psychosom 89(2):111-113. https://doi.org/10.1159/000503571（致编辑信，PMC7158225 全文） | 是 | 「Data came from 8,780 adults aged 50+ from the English Longitudinal Study of Ageing, with biennial measures from 2004/5 to 2016/17」。「taking up a hobby was associated with a decrease in depressive symptoms (coefficient −0.28, 95% CI −0.34 to −0.23) and a 30% lower odds of experiencing depression (OR 0.70, 95% CI 0.64 to 0.76)」。「a further 7% was explained by time-varying reading habits, social engagement, and physical activity」。无抑郁无爱好者：「a 32% lower odds of developing depression (… OR 0.68, 95% CI 0.56–0.83)」。有抑郁无爱好者：「a 272% higher odds of recovering from that depression (… OR 2.72, 95% CI 2.09–3.53)」。基线 15.6% 超过 CES-D 抑郁阈值，结局常见，按全书译法只写「比值比 OR」，说人话只写方向，没用原文的「30% lower odds」「272% higher odds」 |
| Johnson JK 等 (2020). J Gerontol B 75(3):549-559. https://doi.org/10.1093/geronb/gby132（Europe PMC 摘要；PMC7328053 全文用 curl 取 pmc.ncbi.nlm.nih.gov 页面） | 是 | 「Twelve Administration-on-Aging-supported senior centers were cluster randomized into two groups: the intervention group started the choir immediately and a wait-list control group began the choir 6 months later.」「The sample (N = 390) had a mean age of 71.3 years」，招的是 60 岁以上。「The 90-min choir sessions took place at the senior centers, and each choir met weekly」。「There were no significant group-by-time differences at 6 months on the three primary outcome measures: PHQ-8, TMT, or chair stands」。「significantly greater improvements in loneliness (p = .02; standardized effect size [ES = 0.34] and interest in life (p = .008, ES = 0.39)」 |
| Coulton S, Clift S, Skingley A, Rodriguez J (2015). Br J Psychiatry 207(3):250-255. https://doi.org/10.1192/bjp.bp.113.129908（Europe PMC 摘要；Cambridge 全文页没取到） | 摘要 | 「A pilot pragmatic individual randomised controlled trial comparing group singing with usual activities in those aged 60 years or more.」「A total of 258 participants were recruited across five centres in East Kent. At 6 months post-randomisation, significant differences were observed in terms of mental health-related quality of life measured using the SF12 (mean difference = 2.35; 95% CI = 0.06-4.76)」「At 3 months, significant differences were observed for the mental health components of quality of life (mean difference = 4.77; 2.53-7.01), anxiety (mean difference = -1.78; -2.5 to -1.06) and depression (mean difference = -1.52; -2.13 to -0.92).」抑郁焦虑用的是哪张量表、6 个月时抑郁还在不在，摘要没写，正文里也就没写。这篇是从 Johnson 2020 的引言里顺藤摸到的，初稿漏了它、写成「分组试验只有一项」，后来改成「两项，结果不一致」 |
定级与量级
- 证据定 A 并标争议：有 9 万多人的多国合并跟踪数据和一项分组试验，能给具体数字，和本节第 9、10 条同一做法。反方是美国那项合唱试验（Johnson 2020）：主要指标抑郁症状半年时没有差别；英国那项（Coulton 2015）3 个月时抑郁评分低 1.52，两项对不上。
- 口径记死亡率（含健康终点），和本节第 6、7 条（抑郁、焦虑）一致。
- 收益记「小」，凭判断：最严格的同人自比下只差约 0.1 个标准差；比值比结局常见，不能换成降幅套阈值；两项分组试验在抑郁上结果不一。
- 人群标签没加：证据只覆盖 50 岁以上，但年轻人培养爱好不算「毫无用处」，按「拿不准就不加」处理。标题和备注写明了年龄。
刻意没写的
- issue 里的「行差踏错」：没找到任何证据，没写。「能攒下钱、非力工」也没写进条目，证据不按职业和收入分。
- Fancourt D, Steptoe A (2019). BMJ, l6377（https://doi.org/10.1136/bmj.l6377）：去博物馆、看戏、听音乐会的频率与 14 年死亡率，经常去的风险比 HR 0.69（0.59–0.80）。没收，因为测的是出门看展看戏（receptive arts），不是自己做的爱好；而且身体好的人才出得了门，这层干扰比上面两项更重，作者也写「causality cannot be assumed」。
- Tomioka K 等 (2016). J Epidemiol 26(7):361-370（https://doi.org/10.2188/jea.je20150153）：日本 1853 名 65 岁以上老人，「既没爱好也没人生意义」对「两样都有」死亡 HR 2.08（1.47–2.94）。没收，暴露把爱好和人生意义（ikigai）捆在一起，单看爱好的数字不是主结论。
- Pressman SD 等 (2009). Psychosom Med 71(7):725-732：1399 人一次性调查，休闲活动多与血压、皮质醇、腰围低相关。只是同一时间点的相关，没收。
- Mak 2023 里「各类爱好有共同的作用成分、差别不大」是作者引的理论观点，不是本研究结果，没写成「哪种爱好都一样」，只写问卷不分种类。