Safety alarm RCT in India

Background Sexual violence is a major threat to women in Indian slums, particularly among lower caste communities. Traditional interventions are often insufficient due to structural and environmental limitations. This study evaluated the effectiveness of a low-cost, community-based alarm intervention designed to prevent sexual victimisation in urban informal settlements. Methods A single-blind cluster randomised controlled trial was conducted six residential blocks in three slums of Delhi. A total of 500 adult women were randomised by residential block into intervention (9 blocks, n=239) and control (11 blocks, n=261) groups. The intervention consisted of distributing personal safety alarms and conducting community-based training sessions. Data on sexual victimisation and psychological sense of trust were collected before and after the six-months-intervention. Statistical analyses included logistic regression and repeated-measures analysis of variance. Findings A total of 485 participants (intervention: 226, control: 259) completed both the pre-test and post-test. The vast majority of participants have experienced wide range of sexual victimisation in their lifetime, including stalking (n=413, %), verbal sexual harassment (n=338, %), and non-consensual physical contact (n=304, %). Statistically significant group-by-time interaction effects were observed for underwear theft (OR=0·207, 95%CI [0·062; 0·694]) and non-consensual physical contact (OR=0·199, 95%CI [0·059; 0·670]), favouring the intervention group. Other victimisation types declined in both groups over time, including stalking (OR = 0·30, 95% CI [0·075, 1·176], p = 0·082), verbal harassment (OR = 0·86, 95% CI [0·187, 3·987], p = 0·852), voyeuristic behaviour (OR = 0·49, 95% CI [0·040, 6·126], p = 0·583), and indecent exhibition (OR = 0·31, 95% CI [0·054, 1·746], p = 0·183). Trust scores improved significantly in the intervention group, suggesting psychosocial benefit (F(1, 966) = 41·12, p < 0·001). Interpretation The intervention demonstrated potential in reducing opportunistic forms of sexual violence and promoting trust in vulnerable communities. This context-sensitive strategy may serve as a scalable model for similar settings lacking formal protection systems.

研究背景：性暴力是印度贫民窟女性面临的主要威胁，在低种姓群体中尤为突出。受结构性与环境层面的局限，传统干预手段往往难以发挥应有作用。本研究针对城市非正式定居点，评估了一款低成本、社区化警报干预措施在预防性侵害方面的有效性。
研究方法：本研究在德里三处贫民窟的六个居住区块中开展了单盲整群随机对照试验（single-blind cluster randomised controlled trial）。共计500名成年女性按居住区块被随机分配至干预组（9个区块，n=239）与对照组（11个区块，n=261）。干预措施包括发放个人安全警报器，以及开展社区培训课程。研究在为期6个月的干预前后，分别收集了性侵害经历与心理信任感相关数据。统计分析方法涵盖逻辑回归（logistic regression）与重复测量方差分析（repeated-measures analysis of variance）。
研究结果：共计485名受试者（干预组226名，对照组259名）完成了前测与后测。绝大多数受试者在一生中曾遭遇过多种类型的性侵害，包括跟踪骚扰（n=413，%）、言语性骚扰（n=338，%）以及非自愿身体接触（n=304，%）。针对内衣盗窃（比值比（odds ratio, OR）=0.207，95%置信区间（confidence interval, CI）[0.062, 0.694]）与非自愿身体接触（OR=0.199，95%CI[0.059, 0.670]），观察到具有统计学意义的组间-时间交互效应，且干预组获益更显著。其余侵害类型在两组中均随时间呈下降趋势，其中跟踪骚扰（OR=0.30，95%CI[0.075, 1.176]，p=0.082）、言语骚扰（OR=0.86，95%CI[0.187, 3.987]，p=0.852）、窥淫行为（OR=0.49，95%CI[0.040, 6.126]，p=0.583）以及露阴行为（OR=0.31，95%CI[0.054, 1.746]，p=0.183）。干预组的信任感评分显著提升，提示其具备社会心理获益（F(1, 966)=41.12，p<0.001）。
研究结论：本干预措施展现出在弱势社区中减少机会主义性暴力、提升群体信任感的潜力。这种贴合当地情境的策略，可为缺乏正式保护体系的类似场景提供可推广的干预范式。