A study in Frontiers in Medical Technology by KGMU researchers found that Dozee's contactless AI system for early detection of patient deterioration could save 2.1 million lives and cut healthcare costs by ₹6,400 crore annually
Published Oct 29, 2024 | 7:00 AM ⚊ Updated Oct 29, 2024 | 7:00 AM
Imagine you are managing a big garden with many different plants, each needing special care. Some plants are delicate and need constant attention, while others seem fine but can suddenly start wilting without warning. If you could only check on each plant every few hours, you might miss early signs of trouble — a drooping leaf or drying soil. By the time you notice, it might be too late to save the plant.
Now, picture a smart gardener watching each plant continuously. This gardener does not touch the plants but senses subtle changes in their health, spotting early signs that a plant might need water, nutrients, or shade. When something is off, the gardener alerts you, allowing you to step in and save the plant before it is too late.
In hospitals, doctors and nurses face a similar challenge with patients in regular wards. These are patients who do not need intensive care monitoring, but their conditions can worsen suddenly, often without clear signs.
Checking every few hours might mean doctors miss early warning signs. Researchers at King George’s Medical University (KGMU), Lucknow, have studied a technology that works like the “smart gardener” for patients — a system called Dozee, designed to detect signs of deterioration early.
In a large-scale observational study published in Frontiers in Medical Technology, researchers from KGMU revealed that early detection of patient deterioration using Dozee’s contactless AI-based system could save up to 21 lakh lives each year and reduce healthcare costs by ₹6400 crore annually.
At the heart of the study is Bengaluru-based Dozee’s Continuous Contactless Monitoring system — an automated, contactless, continuous monitoring system that detects heart rate, respiratory rate, and blood pressure through ballistocardiography (a technology that detects subtle vibrations created by the heart and respiratory movements) — a remote health monitoring device that anticipates critical patient deterioration up to 16 hours in advance.
Unlike traditional monitors, the Dozee system operates without physical contact, minimising patient discomfort and enabling continuous monitoring in non-intensive settings.
This crucial lead time provides healthcare teams with a rare window for timely intervention, potentially transforming patient outcomes.
The study brings to light the limitations faced by nearly 1.9 million patients across India who currently rely on traditional spot checks in general hospital wards. Dozee’s innovation leverages ICU-grade, continuous monitoring, which could effectively expand critical care across 95 percent of a hospital’s capacity at a fraction of the usual cost.
The study employed a prospective observational cohort design, observing patients admitted to King George’s Medical University from October 2022 to January 2023.
The institution approved the study, and all patients (or their representatives) provided informed consent.
Data was collected from Dozee sensors placed under the patient’s mattress. These sensors captured vital signs continuously, converting subtle bodily vibrations into readings that were analysed and stored securely.
The system, equipped with a tiered alert framework, categorised patients’ alerts based on severity levels, ensuring that only meaningful changes prompted alerts. This reduced false alarms and minimised “alert fatigue” among healthcare providers.
Alerts were triggered based on three vital parameters: heart rate (HR), respiratory rate (RR), and blood pressure (BP).
When a patient’s readings deviated significantly from the established baseline, an alert was issued, escalating depending on the degree of deviation.
The study, conducted at King George’s Medical University in Lucknow, involved 706 patients over approximately 84,448 hours of monitoring.
Patients were split into two groups: those who experienced clinical deterioration, totaling 33 patients (5%), and those who were discharged without complications (95%).
The research was focused on the following:
Assessing the effectiveness of the Dozee Early Warning System (Dozee-EWS) in identifying early signs of clinical deterioration.
Evaluating the potential time savings for healthcare providers, who spend a significant portion of their day on periodic vital checks and documentation.
This study’s findings offer insight into how contactless monitoring can detect deterioration early and improve patient outcomes, marking a notable shift in general ward care.
The study found that Dozee-EWS is good at catching warning signs when a patient’s health might be worsening, identifying these changes in most cases. The system was able to correctly spot signs of deterioration between 67% and 94% of the time, depending on the situation. While the system did sometimes give false alarms, it was still effective in noticing changes early enough for doctors and nurses to step in, especially in the 24 hours before a serious event.
For patients whose health was getting worse, Dozee gave more alerts — around 11 times in the 24 hours before things got serious — compared to just four alerts on average for patients who were stable.
This steady increase in alerts helped doctors prioritise which patients needed urgent attention. On average, the system alerted healthcare staff 16 hours before a patient’s condition became critical, giving them a valuable window to act.
The study also showed that Dozee-EWS can save healthcare providers time. Normally, about 10% of their day is spent checking patient vitals and noting them down.
With Dozee constantly monitoring vitals and sending alerts, healthcare providers can save roughly 140 minutes every day. This saved time means they can focus more on direct patient care, especially helpful in busy hospitals with limited staff.
By tracking over 700 patients for 85,000 hours, the study confirmed Dozee’s potential to save lives by providing early alerts, allowing medical teams to act quickly. It also reduced the daily workload of healthcare staff by about 2.4 hours, making patient care more efficient and responsive.
Dr Himanshu Dandu of KGMU, one of the authors of the paper , underscoring the importance of such technology in settings with limited resources said in a statement, “Dozee’s Early Warning System (EWS) provides scalable, affordable solutions that align with the needs of healthcare systems managing high patient volumes.”
Dr Jean-Louis Teboul, a critical care expert at Paris-Saclay University, another author, also commented on the global applicability of this breakthrough, calling it a “milestone that could reshape healthcare worldwide, addressing the universal need for timely, accessible, and affordable care.”
In places where there aren’t enough healthcare staff, systems like Dozee-EWS can be especially helpful. Dozee-EWS monitors patients automatically and gives alerts when needed, so staff does not have to do as many routine checks.
This setup helps ensure that signs of worsening health aren’t missed and that care is given on time, even in busy hospitals. It also lets healthcare workers focus more on patients who urgently need help, which could lower the number of patients needing ICU care and help hospitals make better use of resources.
The study did find that the system sometimes creates “false alarms,” which can lead to “alarm fatigue” — when frequent alerts cause staff to become less responsive to them. To address this, Dozee-EWS categorises alerts by urgency, so only the most serious changes demand immediate attention.
Future versions of the system might also include extra health measures, like oxygen levels and body temperature, to make alerts even more accurate.
The study’s positive results suggest that Dozee-EWS could be useful in many different healthcare settings, from small rural clinics to large urban hospitals.
Adding more health measures and improving alert accuracy could make this system even more valuable, providing continuous monitoring for a wide range of patients and health facilities.
This study did not account for the relationship between the care giver and the patient, and the importance of non-verbal communication between, for example, the patient and the nurse. There are studies that show that compassion, touch and positive eye contact make a difference to patient outcomes. These elements are crucial to the provision of healthcare, and interpersonal relations are important in determining outcomes.
(Edited by Rosamma Thomas)
