Raberahona Mihaja, Zimmer Alexandra, Rakotoarivelo Rivonirina Andry, Randrianarisoa Patrick Andriniaina, Rambeloson Garcia, Rakotomijoro Etienne, Andry Christophe Elody, Razafindrakoto Haingonirina Anique, Andriantahiana Dera, Randria Mamy Jean de Dieu, Rakotosamimanana Niaina, Grandjean Lapierre Simon
Department of Infectious Diseases, University Hospital Joseph Raseta Befelatanana, Antananarivo, Madagascar.
Centre d'Infectiologie Charles Mérieux, University of Antananarivo, Antananarivo, Madagascar.
ERJ Open Res. 2025 Mar 24;11(2). doi: 10.1183/23120541.00655-2024. eCollection 2025 Mar.
Recent advances in digital and wearable technologies with artificial intelligence (AI) enable the use of continuous cough monitoring (CCM) to objectively monitor symptoms as surrogate markers of treatment efficacy in pulmonary tuberculosis (PTB). The objectives of this study were to describe the evolution of cough during PTB treatment in adults and to assess the feasibility of community-based CCM.
We prospectively enrolled PTB adult participants upon treatment initiation. Participants' coughs were continuously monitored during 6 months with a smartphone loaded with an app able to detect cough by using an AI algorithm.
22 participants were included. The median (interquartile range (IQR)) age was 28.5 (22-42) years and 62% were male. The median (IQR) coughs per hour (medCPH) was 11.0 (7.0-27.0) at week 1. By the end of the intensive phase of PTB treatment at week 8, the medCPH was 3.5 (1.5-7.0), which was significantly lower than the medCPH at week 1 (p=0.002). At week 26 (end of treatment), the medCPH was 1.0 (1.0-2.5). The adherence to CCM was high during the first 13 weeks of PTB treatment and then waned over time. The adherence was similar during daytime and night-time.
Cough counts rapidly drop during the intensive phase of PTB treatment and then slowly decrease to a low baseline level by the end of the treatment. Community-based CCM using digital technology is feasible in low-resource settings but requires evaluation of alternative approaches to overcome adherence issues and technical limitations (mobile internet and electricity availability).
数字技术和可穿戴技术与人工智能（AI）的最新进展使得连续咳嗽监测（CCM）能够用于客观监测症状，作为肺结核（PTB）治疗效果的替代指标。本研究的目的是描述成人PTB治疗期间咳嗽的演变情况，并评估基于社区的CCM的可行性。
我们在治疗开始时前瞻性纳入PTB成年参与者。使用装有能够通过AI算法检测咳嗽的应用程序的智能手机，在6个月内对参与者的咳嗽进行连续监测。
纳入了22名参与者。年龄中位数（四分位间距（IQR））为28.5（22 - 42）岁，62%为男性。第1周时每小时咳嗽中位数（medCPH）为11.0（7.0 - 27.0）。到PTB治疗强化期第8周结束时，medCPH为3.5（1.5 - 7.0），显著低于第1周的medCPH（p = 0.002）。在第26周（治疗结束时），medCPH为1.0（1.0 - 2.5）。PTB治疗的前13周内CCM的依从性较高，然后随时间下降。白天和夜间的依从性相似。
咳嗽次数在PTB治疗强化期迅速下降，然后在治疗结束时缓慢降至低基线水平。在资源匮乏地区，使用数字技术的基于社区的CCM是可行的，但需要评估替代方法以克服依从性问题和技术限制（移动互联网和电力供应）。
