Research on Intelligent Early Warning and Closed‑Loop Nursing Intervention Model of Pet Subclinical Health Risk

Abstract

This study addresses the challenge of detecting pet subclinical health risks, where physiological abnormalities occur without obvious clinical symptoms, leading to delayed intervention and high subsequent medical costs. Existing wearable pet monitoring devices mostly rely on single‑threshold alarms and lack multi‑indicator comprehensive risk evaluation, while few studies incorporate implementation compliance and response delay into health‑economic assessment. This paper constructs an intelligent early‑warning and closed‑loop nursing intervention model for pet subclinical health risks. An entropy‑weight method is adopted to fuse 12 physiological and behavioral monitoring indicators to calculate individual subclinical risk scores. A discrete‑time Markov chain model simulates monthly state transitions among health, subclinical, and clinical conditions. Three intervention scenarios (traditional manual observation, open‑loop intelligent monitoring, closed‑loop intelligent monitoring) are established, and cost‑effectiveness analysis is performed using quality‑adjusted health‑life‑month outcomes. Simulation results show that the risk‑scoring model achieves an AUC of 0.8395. Compared with traditional observation, the closed‑loop intervention reduces the three‑year cumulative clinical incidence by 32.3%. The incremental cost‑effectiveness ratio of the closed‑loop mode is 1424 yuan/QAHLY, which is economically acceptable under a conservative willingness‑to‑pay threshold. Findings indicate that improving implementation compliance and shortening response delay deliver greater health gains than merely boosting detection performance. This model provides technical and economic references for pet‑wearable product design, insurance cooperation, and popularization of early pet‑health intervention.