Long-Term Efficacy of an AI-Based Health Coaching Mobile App in Slowing the Progression of Non-Dialysis Dependent CKD: A Retrospective Cohort Study
Jianwei Ma,Jiangyuan Wang,Jiapei Ying,Shasha Xie,Qin Su,Tianmeng Zhou,Fuman Han,Jiayan Xu,Siyi Zhu,Chenyi Yuan,Ziyuan Huang,Jingfang Xu,Xuyong Chen,Xueyan Bian
DOI: https://doi.org/10.2196/54206
2024-10-14
Abstract:Background: Chronic kidney disease (CKD) is a significant public health concern, with an escalating global prevalence ranging from 11% to 13%. Therefore, practical strategies for slowing CKD progression and improving patient outcomes are imperative. There is limited evidence to substantiate the efficacy of mobile app-based nursing systems for decelerating CKD progression. Objective: This study aimed to evaluate the long-term efficacy of the KidneyOnline intelligent care system in slowing the progression of non-dialysis-dependent CKD. Methods: In this retrospective study, the KidneyOnline app was utilized for CKD patients in China who were registered between January 2017 and April 2023. Patients were divided into two groups: an intervention group using the app's nurse-led, patient-oriented management system and a conventional care group that didn't use the app. Patients' uploaded health data were processed via deep learning optical character recognition, and the AI system provided personalized healthcare plans and interventions. Conversely, the conventional care group received suggestions from nephrologists during regular visits without AI assistance. Monitoring extended for an average duration of 2.1 years post-recruitment, with the study's objective being to assess the app's effectiveness in preserving kidney function. The primary outcome was the eGFR slope over the follow-up period, and secondary outcomes included changes in albumin-to-creatinine ratio (ACR) and mean arterial pressure (MAP). Results: A total of 12297 eligible patients who registered on the KidneyOnline app from January 2017 to April 2023 were enrolled for the analysis. Among them, 808 patients were successfully matched using 1:1 propensity score matching, resulting in 404(50%) patients in the KidneyOnline care system group and another 404(50%) patients in the conventional care group. The eGFR slope in the KidneyOnline care group was significantly lower than that in the conventional care group (-1.3, 95% CI: -2.4, -0.1 mL/min/1.73 m2 per year vs. -2.8, 95% CI: -3.8, -1.9 mL/min/1.73 m2 per year，P=.009). Subgroup analysis revealed that the effect of the KidneyOnline care group was more significant in males, patients over 45 years old, and patients with worse baseline kidney function, higher blood pressure, and heavier proteinuria. After 3 and 6 months, the MAP in the KidneyOnline care group decreased to 85.6 (SD 9.2) and 83.6 (SD 10.5) mmHg, respectively, compared to 94.9 (SD 10.6) and 95.2 (SD 11.6) mmHg in the conventional care group (P<.001). The ACR in the KidneyOnline care group showed a more significant reduction after 3 and 6 months (736 mg/g vs. 980 mg/g and 572 mg/g vs. 840 mg/g, P=.074, .027)); however, there was no significant difference in ACR between the two groups at the end of the follow-up period (618 mg/g vs. 639 mg/g, P=.904). Conclusions: The utilization of KidneyOnline, an AI-based, nurse-led, patient-centered care system, may be beneficial in slowing the progression of non-dialysis-dependent CKD.