Personal tools

AI4D

Mobile BP

Mobile BPThis project’s aim is to develop a low cost, easy-to-use device to assist a minimally-trained person in taking the BP using a cuff, a cell-phone and small signal acquisition box that connects both. The main ideas are the replacement of the high-cost manometer of the sphygmomanometer, which can be done using a small, low-cost, signal acquisition box attached to the manometer tube in order to get pressure signal; eliminate the need to be trained (only picture instructions) to identify Korotkoff sounds by connecting the signal acquisition box to a cell phone that would process and display the blood pressure result. This system won the 2011 EWH Competition, and later the Best Innovation Leveraging Technology in the Dell Social Innovation Challenge in 2012.

The hardware needed to make this software work is: Android Device with USB-OTG + USB-OTG Adapter (device specific) + BP Box + Blood pressure Cuff with manual Pump and Valve.

The main features of the software are:

  1. Measure Systolic, Diastolic pressure and pulse, from cuff with manual pump.
  2. Save blood pressure and pulse measure into device database.
  3. View saved blood pressure and pulse measures.
  4. Detect arrhythmias and estimate respiration rates
  5. Upload data to remote medical record system (using the MIT Sana framework)

Software Installation

Ways to install software with an android device:

  1. Take a picture of the QR code bellow with a QR Reader App from your android phone. Download the software and click on it to install it.
    Mobile BP QR Code
  2. Follow the instructions in our User Manual

Device Pre-validation

The device is being pre-validated according to standard protocols in China and South Africa.

Contributors:

Carlos Arteta (DPhil)
Joao Domingos (DPhil)
Marco Pimental (DPhil)
Ali Maraci (DPhil)
Mauro Santos (DPhil)
Gari Clifford (Project Lead)

Telespiro

TelespiroLabelThe burden of chronic lung disease in the developing world is both staggering and growing as a result of increased air pollution, tobacco use and workplace exposures. Chronic obstructive pulmonary disease (COPD) affects 210 million people globally and was the fourth leading cause of death worldwide in 2004. Widely expected to become the third leading cause of death in a matter of a few years, COPD is a disabling combination of emphysema (destruction of lung’s elastic recoil) and chronic bronchitis that can be managed with inhaled medications if caught early; however, even rudimentary screening procedures and clinical management of symptoms has not been employed in the world’s poorest regions. Part of the problem arises with the cost of equipment and lack of personnel necessary to measure lung function reliably via spirometry.

Chronic respiratory disease in the developing world is both increasingly prevalent and difficult to manage as a result of shortages of medications, accurate equipment and qualified medical personnel. Development of a low-cost mobile phone based spirometer device named “TeleSpiro” specifically designed for resource limiting setting. Telespiro is a open tube differential pressure sensing spirometer that can detect ambient humidity and pressure at a low price point. The key contributions of this design, calibration and signal processing is repeat-use sterility, low cost of development, no requirement for independent power source, operation independent of computer hardware via use of new Android USB host mode capabilities, respiratory tube design and integration with an electronic health record system. The tube and mouthpiece were designed via turbulent CFD and international specifications to optimize collection of key pulmonary function tests (FEV1/FVC). PIC Microcontroller and Android software applications are being implemented on the collected signal from the differential pressure sensor and filtered to obtain accurate and precise pulmonary function tests. Android software will be built into the OpenMRS system as a way of transmitting data and also provided user feedback thus acting like a technician.

Contributors:

Will Carspecken (MEng, MD)
Gari Clifford (Supervisor)

Electronic Medical Records

This collection of projects is aimed at producing a streamlined and accurate data recording and communication system for health workers and patients. We aim to create both centralized and personalized medical record systems, which communicate with each other. Interactions between a patients, medical experts and caregivers (including friends and family) are, as far as is possible, mediated in a hardware-independent and communication channel-independent manner.

Our projects include:

  • SanaMobile: A mobile interface to medical records for resource-constrained systems, in collaboration with MIT.
  • OpenMRS: We support development of this open source EMR in collaboration with Partners in Health.
  • HIV iChart - Drug-Drug Interaction Application in collaboration with Liverpool School of Tropical Medicine

Contributors:

Joachim Behar (DPhil)
Maxim Osipov (DPhil)
Phil Houghton (Research Engineer)
Gari Clifford (Supervisor)

Mobile Heart Sounds

msStethoscopeThe aim of this project is to allow low-cost, remote screening for rheumatic heart disease (RHD) - a condition that is still prevalent in developing areas of the world. RHD is the leading cause of heart failure in children and young adults worldwide and results in heart murmurs that are almost always audible during auscultation. Screening for RHD is performed  using a low-cost stethoscope attachment for a mobile phone to record heart sounds and a phone app, which should be able to distinguish murmurs in cases of RHD from normal subjects. Data for this project is being collected in  partnership with Cardiac Clinic at Groote Schuur Hospital, Cape Town, South Africa.

By using a (smart) mobile phone to record (and analyse) the data, we can leverage the high quality audio capture and signal analysis facilities of the phone, as well as the intuitive user interface and capability of uploading to a remote upload to a medical record. In particular, the development of signal quality metrics allow even an untrained user to know if the recording is of good enough quality for analysis or if another recording is needed. By doing this at the point of data capture (and automatically uploading the data when the phone is in range of a suitable network), the minimum amount of data loss is experienced. Furthermore, the development of signal segmentation techniques can help address the shortage of trained professionals and provide automated analysis either in the cloud, or on the ground, even when there is no connectivity to the cloud.

Contributors:

David Springer (DPhil)
Gari Clifford (Supervisor)
Lionel Tarassenko (Supervisor)

CVD Screening

Cardiovascular diseases (CVD) are a major cause of premature death and disability in both rural and urban India. Wide gaps in the translation of best practice into primary health care exist. mhealth interventions for overcoming these gaps hold promise but as yet robust evaluations of such systems is missing.

In collaboration with the George Institute for Global Health we have develop a multifaceted non-physician healthcare worker (NPHWs) intervention utilizing an electronic clinical decision support system (CDSS) and preliminarily evaluate it for utility, effectiveness and acceptability by the community, NPHWs and primary health care (PHC) physicians in this setting. The CDSS algorithm for screening and management of CVD is based on WHO and country specific guidelines and built as an Android V4.0 application for a 7 inch tablet. The algorithm has been validated using a large de-identified existing data-set and the CDSS has been field tested in eleven rural villages of Andhra Pradesh in India.  Results form the pilot indicate that a tablet based CDSS, integrated with existing PHC system could contribute to improved CVD detection, prevention and management in Indian primary health care system. Following refinements to the system, the CDSS will now be evaluated in a cluster randomized clinical trial involving 54 villages and 16000 people funded by the Australian National Health & Medical Research Council (NHMRC) under a million dollar Global Alliance for Chronic Disease (GACD) grant.

Contributors:

Arvind Raghu (DPhil)
Gari Clifford (Supervisor)
Lionel Tarassenko (Supervisor)

Smart Water

This project attempts to connect the provision of clean water with downstream communicable diseases such as diarrhoea, one of the largest global killers. Delivering a sufficient, reliable and affordable supply of safe water is an enduring public policy challenge. We are part of an interdisciplinary, cross-departmental research cluster that sets out to design, build and evaluate mobile technology innovations for water security called mobile/water for development. The cluster have recently completed a cross-country analysis of mobile water payment innovations in urban Africa and have started the first global studies that are testing 'smart handpumps' in rural Kenya and Zambia. They are also examining how 'smart river systems' may be designed to negotiate fairer and more sustainable outcomes in Kenya under increasing hydro-climatic risk and competing water demands. We are using the same SMS-server configuration for monitoring incidences of diarrhoea and oral rehydration therapy in highland Guatemala.

Contributors:

Joachim Behar (DPhil)
Contributors:
Ali Maraci (DPhil)
Joao Jorge (DPhil)
Alessandro Guazzi (DPhil)
Gari Clifford (Project Lead)

Open Source ECG Analysis

We are currently developing an open source Java-based ECG analysis system that can run both retrospectively and in real time to perform beat detection, rhythm classification and signal quality feedback. The aim is to enable untrained users to acquire diagnostically useful ECGs for either immediate analysis by a smart phone or remote experts. An Android based phone app has been developed to allow upload to an open source medical record where algorithms are used to combine many different algorithms or humans to provide a maximally accurate diagnosis (given the skill levels of each human or algorithm).

Contributors:

Joachim Behar (DPhil)
Tingting Zhu (DPhil)
Maxim Osipov (DPhil)
Qichen Li (Intern)
Julien Oster (DPhil)
Gari Clifford (Supervisor)