REMOTE CARE MODEL FOR COVID-19
Cover image attribute - <a href='https://www.freepik.com/free-photos-vectors/technology'>Technology vector created by freepik - www.freepik.com</a>
Once I had finished my XLRI BM & HR interviews in late March, I was eager to get back to active practice to deal with COVID19. Thankfully my previous employer, Dr Marudhupandian M.D, CEO of Pon Malligai hospitals Chennai, was more than delighted to have me back. Being a social entrepreneur himself, very soon both our ethos got aligned towards making COVID care affordable, especially considering the exorbitant rates that were charged during that period. We both knew it was the right thing to do, and so we decided to get to work immediately. I am very grateful that he entrusted me and my team with the responsibility of aiding him in developing an affordable online care model. It was a great learning experience and one that was immensely fulfilling on a personal level.
Dual Objectives -
Design and implement a REMOTE / ONLINE COVID19 CARE MODEL to timely cater to the unaddressed mild & moderate cases in the city of Chennai by complementing online expert follow-up with round the clock offline paramedical support and to provide a streamlined path for hospitalization if necessary (onboarding or referral).
Develop a Machine Learning model for Diagnosis-Support (based on extensive in-house data set of early-stage pneumonia Chest PA-X-rays) in order to mitigate false-negative cases that evade triple diagnosis (clinical + imaging + testing) in early stages.
Team members and their responsibilities,
Deepak Ashan Ailani, Operations Consultant - Patient prioritization by dynamic queuing, prediction of bottlenecks & mitigation of operational constraints of supplies & personnel, establishing KPIs for evaluation.
Suhail Hafiz khan J, Artificial Intelligence and Machine Learning developer – Building end to end categorical deep learning model, Image processing, data modelling and preparing final interface.
Timeline elaboration - 10 weeks
Week 1 & 2 - Were dedicated for primary & secondary market research + insight analysis with the following objectives
Identify the optimal quantum of supplies and personnel needed (minimum + buffer)
Decide optimum pricing by deliberating with experts in the value chain of service delivery
Identify best practices followed by similar or relevant models elsewhere or historically
Understand public sentiment and apprehensions for online + offline support model
Week 3 & 4 - Designing the model under the guidance of management + designing marketing strategy. Details are as follows.
Online - 24*7 expert video & audio consultation with regular follow up (based on patient’s criticality). Timely checks via counselling session scheduled based on observed patterns to ensure the adequate upkeep of mental health. Dietary advice and query addressal for the entire family on all relevant matters of personal and social life.
Offline - Paramedical support team providing all necessary equipment (such as pulse-oximeter, CBG) and supplies such as medicine as per expert advise. Ground assistance and mobilization services (via ambulance) provided for necessary tests such as Ecg, X-rays, CT-Chest, RT-PCR, metabolic profile, etc.
Digital Marketing strategy - Testimony based. Leveraging in-house patient database and those available with providers like Practo. The objective of the campaign was to show that successful 360-degree remote care was possible with ONLINE expert Consult + OFFLINE paramedical support + INITIATIVE on side of the patient.
End of week 4 - By this time, the dedicated physician team led by Dr. Marudhupandian M.D. had successfully handled more than 15 COVID19 patients from diagnosis to recovery all within the comfort of their home. The testimonies generated immense goodwill which translated into more patients reach out to us and also more referrals from peers in the city.
Week 5 - 8 - The objective was to achieve phased weekly improvisation of the online model and offline support based on patient testimonies and operational constraints faced by ground personnel.
Optimized the duration and the time slot for each patient consultation and respective follow-up via dynamic queuing based on regularly updated status of patient criticality and predicted prognosis
Week 9 & 10 - Monitoring of care model, providing value-added support for expansion, and designing streamlined onboarding of patients from remote to inpatient care based on case to case analysis.
By the time the project was completed, more than 150 patients have benefited from the model saving them roughly 97% for institutional private care for the same patient & severeness grade.
The AI & ML Chest PA Xray model, trained with 2500+ high-quality x-rays, played a key role in the mitigation of false-negative diagnosis (to near zero), through achieving an accuracy of 88%.
In recognition of the wonderful work done by the MD and this team, Pon Malligai hospitals got Government authorization for an entire Block for in-patient COVID19 care (20 beds + 5 ICU units).
