From Whatsapp to Prediction: How Automation Transformed Post-Discharge Surveillance of Surgical Infections

Brazilian legislation requires post-discharge surveillance of surgical patients to control Surgical Site Infections (SSI), but the process faces challenges such as difficulty in follow-up, communication failures, and high costs of traditional methods. The use of WhatsApp has proven to be a viable, sensitive, and low-cost alternative, being considered a successful contact when the patient views the message. If symptoms are reported, the team must call to investigate. Despite its effectiveness, the manual process still demands significant time. To evaluate the benefits and impact of a system that automates follow-up of post-discharge surgical patients via WhatsApp and, at the same time, to present a model to predict SSI based on patient responses, later validated by telephone. Post-discharge surgical patient data are imported from hospital systems into the SACIH Egressos system, which automates all contact at intervals of 30, 60, or 90 days, depending on the case. Additionally, an SSI prediction model was developed using multivariate logistic regression, built from data from nine hospitals and based on patient responses to a symptom questionnaire. Implementation of SACIH Egressos in a public hospital in Belo Horizonte in August 2023 resulted in an increase in adherence and identification of post-discharge SSI cases from 20% to 49%. This system outperformed the previous method, which did not generate real-time statistical reports and required manual message sending, leading to rework. For the SSI prediction model, a sample of 4,719 patient responses from nine hospitals collected between January and December 2024 was used, along with a validation sample of 2,005 patients. The final multivariate logistic regression model identified purulent secretion presence and total number of symptoms as the only significant predictors of SSI. Both the modeling and test datasets achieved an area under the ROC curve of 0.77. Implementation of the SACIH Egressos system improved patient surveillance and SSI detection by automating WhatsApp contact, eliminating manual workload and enabling real-time report generation. Additionally, a predictive model with moderate capacity was developed to help identify patients at risk.