According to sources, India is termed as the Tuberculosis capital of the world (WHO) for the second consecutive year. Multi Drug Resistant (MDR) TB is on the rise and can be fatal if undetected or untreated. However, it has been observed that TB screening in India does not have a proactive approach because of the social stigma attached to the disease and in general.
A personal experience related to IGRA test for TB coupled with his mother’s diagnosis of diabetes followed by TB contraction motivated Rahul Pathri to select this field of work. Adding on to the personal experiences, he faced a huge loss by the death of his friend due to a TB induced lung collapse which only made his resolve stronger to come up with a permanent solution to this problem. He wanted to explore it for a larger benefit of finding an early, easy and affordable method of TB detection while tracking various nuances of the disease not limited to different geographies such as coal mines, slums, urban rural and rural population, with HIV co-infection not being an exception. This incident was the eureka moment for him and that’s when the whole idea of the smartphone application, TimBre, was conceived.
What is the startup all about?
Docturnal’s TimBre involves spectral analysis (.WAV/MP3) of a recorded sound file of a person’s cough and its processing in the home grown machine/deep learning algorithms of classification of various cough patterns in conjunction with Clinical, Demographic and Socio-Economic data. This analysis has the ability to finally confirm whether the condition is TB or not (which includes LTBI, MDR-TB).
The team at Docturnal Private Limited comprises of professionals who came together to bring about a change in the field of non-invasive screening and diagnostics of Tuberculosis in India. Rahul Pathri if the founder and CEO at Docturnal, who gave up his cushy job in the BI industry and found his perfect band of like-minded collaborators who have joined him in this fight against TB. Arpita Singh, ace marketer, is the acting co-founder and works pro bono with him. Practicing pathologist Dr. Kalpana Singh has over 30 years of medical experience and acumen of working with Government hospitals. Ramoji Bashupolu (RNTCP) and Ajay Keskar are involved in the spectral and clinical data sample collection of infected TB patients’ front with various Government hospitals.
The founder’s words
“Docturnal’s vision statement is ‘Simply Health which means simplifying health while implying simplicity.’ Docturnal aims to utilize the phenomenal advancements in Mobility, Deep/Machine Learning to deliver healthcare to individuals who may not always have access to the solutions”.
Docturnal’s TimBre app does non-invasive diagnosis of Tuberculosis by spectral analysis of sound signatures of coughs. Given the deep tech nature of this startup and its relevance to India (the TB capital of the world), AVISHKAR has provided Docturnal an INR 10 lakh seed fund and a 6-month accelerator program.
Docturnal is AVISHKAR’s startup. AVISHKAR is a deep tech accelerator program of IIIT-H Foundation that supports start-ups working on deep learning technologies and machine intelligence, with future expansion plans into. Till date IIITH-Foundation has supported 104 start-ups and made seed investment in 15 start-ups. From being a sector agnostic ICT incubator, the foundation has changed its focus onto start-ups working on deep learning technologies and machine intelligence, and plans on expanding into the AR/VR and gaming domain as well.
“Currently there are no apps in India that use cough and machine/deep learning to detect TB and MDR TB. A provisional patent has been filed with a priority date of April-2016 and a Final Patent is under review. Docturnal is looking forward to the TimBre App being approved by NABL. This will enable us to become ‘diagnostic providers’ once the Specificity and Sensitivity numbers are close to 90%. Until then, we will cater to mass screening as ‘screening providers’.” Says the Proud founder, Rahul.
How does TimBre work?
To be able to use this useful app by Docturnal, you will need to download this free app on your phone. Once downloaded, it will ask you to enter your basic information including your name, mobile number, age, occupation, marital status, height, weight and Pin Code. Apart from this, it will also prompt you for some other clinical information and your daily habits like smoking and alcohol. If you have any history of TB or HIV, sleep and cough patterns, H1bAc or blood pressure, you will need to input those details. In fact, you can upload an image of reports and tests conducted before. Once, this basic information is punched in, you will need to record your cough in a noiseless room, preferably in the night. This data gets recorded in the home grown database of the app for algorithms to process it. Post processing, you will receive an SMS confirming whether the cough pattern is TB related or not. The app also suggests the next steps with an SMS for a nearest Sputum Microscopy tests or a visit to a Physician for the next course of treatment.
How does TimBre do the analysis?
The prediction by the algorithms app backend is based on a two-step process. A neural net (deep learning – CNN) followed by traditional machine learning algorithms operating upon clinical and demographic data and sometimes in conjunction with feature variables from WAV file.
“A combination of spectral and clinical/demographic data avoids higher error rates. Most importantly the Deep Learning results are not interpretable on the cough file while Machine Learning results of Clinical, Summary Variables (Intensities) & Demographic are interpretable.” Says Rahul.
Success Stories of Docturnal
Today, this startup is associated with Gandhi Hospital, Hyderabad. The hospital is helping Docturnal with both training data collection (model creation) as well as implementation (Beta). Docturnal has entered into a partnership with the World Vision (an NGO) facilitated the RNTCP. The startup has also been successful in raising a Rs. 10 lakh seed fund and a 6-month accelerator program by the Indian Institute of Information Technology, Hyderabad’s Incubator 2016-17.
Apart from the above partnerships and seed funding, for 2017, Doctrunal has already bagged the Digital Edge 50 Award Winner 2017 and NASSCOM Social Innovation Finalist in Primary Healthcare Category 2017.
Currently the team is working on signing up Beta Customers for its pilot launch. For now, it is only an android mobile-based point of care device. The team is working towards creating iOS and Windows app and Desktop in order to broaden their reach. In the coming few years, the team aims at expanding their operations to foreign countries where TB rates are high, such as Indonesia, China, Africa, and Russia.