RASPRO Clinical Sensor Data Summarization

With connected medical devices fast becoming ubiquitous in healthcare monitoring there is a deluge of data coming from multiple body-attached sensors. Transforming this flood of data into effective and efficient diagnosis is a major challenge. To address this challenge we designed, developed, and tested a predictive healthcare data analytics and communication framework called RASPRO (Rapid Active Summarization for effective PROgnosis) in a collaborative work with doctors.

In RASPRO we built a novel three-step technique to derive high performance alerts from voluminous sensor data (as illustrated in Fig 1).

A data reduction algorithm that takes into account the medical condition at personalized patient level and converts raw multi-sensor data to patient and disease specific severity representation, termed as Personalized Health Motifs (PHM).
A Criticality Measure Index (CMI) score that is representative of the urgency of physician’s consultative attention for the patient is aggregated based on the criticality factors derived from PHM.
In the final step we generate alerts whenever the CMI crosses patient-disease-specific thresholds. These alerts are communicative of the interventional time available with the doctors (as shown in Fig.2).

The RASPRO framework was validated in clinical deployments across multiple specialities in our super speciality hospital.

One such clinical application used for validation of RASPRO is the onset of Acute Hypotensive Episode (AHE), which is a medical condition common in ICUs. For AHE, the medically acceptable time of intervention is at least half an hour prior to onset. However, the CMI based alerts for AHE generates valid alerts up to 180 minutes prior to its onset.
This pilot study also brought out a number of specific clinical applications in the field of cardiology, pulmonology and neurology to which RASPRO could be extended to.

The suitability of RASPRO framework in global health deployment demonstrates its advantages, namely:

Accessibility: Bandwidth and power are the two major limiting factors related to data infrastructure in remote rural areas. Even with the use of the most efficient data compression technique, the energy consumed for processing and transmission of the data is estimated to be orders of magnitude higher compared to the RASPRO algorithm. Such energy savings are becoming indispensable for ubiquitous deployment of IoT based sensors and devices.
Affordability: Based on the alert levels obtained from RASPRO technique, the low and normal patients can forgo unnecessary hospital visits and expensive medical procedures. This will result in overall reduction of healthcare service delivery cost in the global health setting.
Availability: The risk-stratified alerts mechanism that is achieved segregates the patients into four levels among which the physician can focus his consultative time and in-person attention only towards critical/high alert patients. Whereas, for those who are assessed by low & normal alert levels do not require the physician’s attention and forgo unnecessary hospital re-visits.

Related Funded Project

Funding agency: SERB, Govt. of India
Project Title: Development, optimization and pilot evaluation of edge-AI enabled photoplethysmograph IoT devices and clinical data summarization methods for remote patient monitoring and disease detection
Grant No.: SRG/2020/001119

International Collaboration

Dr. Prakash Ishwar, Professor (ECE, SE), Boston University

Related Patents

“Systems, methods, and devices for remote health monitoring and management using internet of things sensors”, MV Ramesh, RK Pathinarupothi, ES Rangan, P Durga, PV Rangan US Patent 10,433,726 (granted)
“Systems, methods, and devices for remote health monitoring and management”, MV Ramesh, RK Pathinarupothi, ES Rangan US Patent Grant No. 10542889 (granted)

Selected Publications

In Journals

K. Pathinarupothi, P. Durga and E. S. Rangan, “IoT-Based Smart Edge for Global Health: Remote Monitoring With Severity Detection and Alerts Transmission,” in IEEE Internet of Things Journal, vol. 6, no. 2, pp. 2449-2462, April 2019.
doi: 10.1109/JIOT.2018.2870068. IF 9.515. Link
Pathinarupothi, Rahul Krishnan, P. Durga, and Ekanath Srihari Rangan. “Data to diagnosis in global health: a 3P approach.” BMC medical informatics and decision making 18, no. 1 (2018): 78. IF 2.674. Link

In Conferences

K. Pathinarupothi, “Clinically Aware Data Summarization at the Edge for Internet of Medical Things,” 2019 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), Kyoto, Japan, 2019, pp. 437-438. DOI: 10.1109/PERCOMW.2019.8730765
K. Pathinarupothi, E. S. Rangan and P. Durga, “Deriving High Performance Alerts from Reduced Sensor Data for Timely Intervention in Acute Hypotensive Episodes,” 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Honolulu, HI, 2018, pp. 3260-3263. DOI: 10.1109/EMBC.2018.8512945
P Durga, Rahul Krishnan Pathinarupothi, Ekanath Srihari Rangan, and Prakash Ishwar. 2018. When Less is Better: A Summarization Technique that Enhances Clinical Effectiveness of Data. In DH’18: 2018 International Digital Health Conference, April 23–26, 2018, Lyon, France. ACM, New York, NY, USA, 5 pages. Link
Krishnan P. P, Alangot, B., and Rangan, E., “Severity Summarization and Just in Time Alert Computation in mHealth Monitoring”, Studies in Health Technology and Informatics, vol. 235, pp. 48-52, 2017. Link
K. Pathinarupothi and E. S. Rangan, “Consensus motifs as adaptive and efficient predictors for acute hypotensive episodes,” 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Seogwipo, 2017, pp. 1688-1691. DOI: 10.1109/EMBC.2017.8037166. Link
Rangan and R. K. Pathinarupothi, “Multi-sensor architecture and algorithms for digital health at every doorstep,” 2017 Second International Conference on Electrical, Computer and Communication Technologies (ICECCT), Coimbatore, 2017, pp. 1-5. DOI: 10.1109/ICECCT.2017.8117947
Durga, E. Rangan and R. K. Pathinarupothi, “Real-time identification & alert of ischemic events in high risk cardiac patients,” 2016 IEEE International Conference on Computational Intelligence and Computing Research (ICCIC), Chennai, India, 2016, pp. 1-5. DOI: 10.1109/ICCIC.2016.7919638
Pathinarupothi, R.K., Rangan, E., Alangot, B., Ramesh, M.V. “RASPRO: Rapid summarization for effective prognosis in wireless remote health monitoring”. In 2016 IEEE Wireless Health, WH 2016, art. no. 7764566, pp. 122-127. IEEE, 2016.

Book Chapters

Rangan and P, R. Krishnan P., “Adaptive motif-based alerts for mobile health monitoring”, Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST, vol. 192, pp. 168-176, 2017. Link