Department of Computer Science, COMSATS University Islamabad, Islamabad 45550, Pakistan.
Department of Preventive Medicine and Public Health, University of Granada, 18071 Granada, Spain.
Int J Environ Res Public Health. 2023 Mar 23;20(7):5246. doi: 10.3390/ijerph20075246.
Smartphone applications or apps are increasingly being produced to help with protection against the risk of domestic violence. There is a need to formally evaluate their features.
This study systematically reviewed app-based interventions for domestic violence prevention, which will be helpful for app developers.
We overviewed all apps concerning domestic violence awareness and prevention without language restrictions, collating information about features and limitations. We conducted searches in Google, the Google Play Store, and the App Store (iOS) covering a 10-year time period (2012-2022). We collected data related to the apps from the developers' descriptions, peer reviewed research articles, critical reviews in blogs, news articles, and other online sources.
The search identified 621 potentially relevant apps of which 136 were selected for review. There were five app categories: emergency assistance ( = 61, 44.9%), avoidance ( = 29, 21.3%), informative ( = 29, 21.3%), legal information ( = 10, 7.4%), and self-assessment ( = 7, 5.1%). Over half the apps ( = 97, 71%) were released in 2020-22. Around a half were from north-east America ( = 63, 46.3%). Where emergency alerts existed, they required triggering by the potential victim. There was no automation. Content analysis showed 20 apps with unique features, including geo-fences, accelerometer-based alert, shake-based alert, functionality under low resources, alert auto-cancellation, anonymous communication, and data encryption. None of the apps deployed artificial intelligence to assist the potential victims.
Apps currently have many limitations. Future apps should focus on automation, making better use of artificial intelligence deploying multimedia (voice, video, image capture, text and sentiment analysis), speech recognition, and pitch detection to aid in live analysis of the situation and for accurately generating emergency alerts.
智能手机应用程序(apps)越来越多地被开发出来，以帮助防范家庭暴力风险。有必要对其功能进行正式评估。
本研究系统地回顾了基于应用程序的家庭暴力预防干预措施，这将有助于应用程序开发者。
我们综述了所有关于家庭暴力意识和预防的应用程序，没有语言限制，整理了有关功能和限制的信息。我们在谷歌、谷歌应用商店和苹果应用商店(iOS)中进行了搜索，时间跨度为 10 年(2012-2022 年)。我们从开发者的描述、同行评议的研究文章、博客的评论、新闻文章和其他在线来源中收集了与应用程序相关的数据。
搜索确定了 621 个潜在相关的应用程序，其中 136 个被选来进行审查。有五个应用程序类别：紧急援助(=61，44.9%)、避免(=29，21.3%)、信息(=29，21.3%)、法律信息(=10，7.4%)和自我评估(=7，5.1%)。超过一半的应用程序(=97，71%)是在 2020-22 年发布的。大约一半来自北美东北部(=63，46.3%)。在存在紧急警报的地方，它们需要由潜在的受害者触发。没有自动化。内容分析显示，有 20 个应用程序具有独特的功能，包括地理围栏、基于加速度计的警报、基于摇晃的警报、在资源有限的情况下的功能、警报自动取消、匿名通信和数据加密。没有一个应用程序使用人工智能来帮助潜在的受害者。
目前应用程序有许多限制。未来的应用程序应该专注于自动化，更好地利用人工智能部署多媒体(语音、视频、图像捕捉、文本和情感分析)、语音识别和音高检测，以帮助实时分析情况，并准确生成紧急警报。
