Safe Heart

A low cost perinatal screening and referral technology

What is it?

Our proposal combines object and service into a low-cost set of open API devices designed to aid prenatal, perinatal, and neonatal care in underserved countries. These devices (Doppler ultrasound, blood pressure monitor, and pulse oximeter) will be linked to cellular smartphones, together with automated decision support and end-to-end electronic medical record (EMR) integration. Working with an NGO in rural Guatemala, we are piloting parts of the system to create a streamlined and culturally appropriate system. We will demonstrate the impact that mHealth technology can have on the quality of care provided by lay midwives in a rural under-resourced setting, particularly with regard to improved rates of referral for prenatal complications and timely neonatal medical care. The system will contribute new technical avenues for improving prenatal, perinatal, and neonatal care and for ensuring continuity of care in these communities. The system is designed to be highly flexible and, therefore, scalable or adaptable to other contexts and disease entities to accommodate specific care models.

How is it used?

All three devices are intended to be worn on the body by the user at the direction of a midwife or healthcare professional, and are easily sterilizable, which allow them to be reused and reapplied to other patients.

What technologies does it incorporate?

The three sensor devices that will be used in the project are 1) A low cost 1 dimensional Doppler ultrasound transducer, 2) a low cost blood pressure monitor, and 3) a low cost pulse oximeter. All devices are connected to and powered by a low cost smart phone (typically $35-100).
Data is recorded by the smartphone, fused together to check for errors, passed through a series of scientifically validated algorithms to identify need for action/onward referral.
Our system will allow the assessment of three key clinical antenatal or perinatal problems which will trigger automated alarms on the mHealth platform automatically prompting the midwife or healthcare professional using the device to refer.
1. Preeclampsia screening
In Latin America, 25% of maternal deaths are attributed to complications from high blood pressure, and rural detection is low because of the difficulty that traditional midwives have with manual blood pressure measurements. Our system will include an automatic blood pressure cuff that attaches to the cell phone.
2. Fetal distress
We intend on using algorithms based on the Dawes-Redman criteria (1) to screen for fetal distress. This requires automatic algorithmic analysis of the signal from the cell-phone enabled Doppler ultrasound device, identifying abnormal patterns in the heart rate (such as marked sinusoidal variations).
3. Intra-uterine growth restriction (IUGR)
In developing countries 11% of neonates are estimated to be affected by IUGR (2) and prevalent methods to detect IUGR require 2-D ultrasound, which is technically complex and expensive. However, markers derived from fetal heart rate traces (such as heart rate variability), collected with a simple Doppler ultrasound (3) have the potential to detect IUGR cases (4) (5) (6) (7) (8). An IUGR- detection algorithm (based on a logistic regression of multiple features) will support the midwife with a more detailed fetal assessment and enable health personnel to prepare for delivery and anticipated neonatal complications.
The algorithms for screening each of these three clinical problems have been widely published and, except in the case of IUGR, extensively evaluated. We therefore intend to implement these algorithms as described in the literature (1) (9) (10), and pilot the use of the IUGR screening algorithm, and subsequently follow up in a much larger follow-up cohort study. (Please see figure 1 Hardware Diagram in included attachments) Schematic of our low-cost devices linked to a smartphone.

How does it work?

Using our specialty of low-power devices with our extensive research team and partners, the device works under the same principle of pulse oximetry, Doppler ultrasound and blood pressure monitors, but focused on use with a smartphone. This ensures that no additional energy source (such as batteries) other than the electricity drawn from the smartphone is needed to power the instrument, and will attain results based on algorithms that are calculated from the smartphone’s CPU through our application designed to operate in tandem with the device. 
The blood pressure cuff is powered through the USB port on the phone and is used at the start of the patient encounter. Then, the midwife plugs the oximeter into the head phone socket and the ultrasound unit (which requires more power) into the USB port of the phone. The unit is strapped to the patient and they are asked to wear it for 20 minutes during antepartum monitoring. The phone processes both signals simultaneously and compares heart rates. If the heart rates are too similar (indicating that the Doppler is focused on a maternal artery), or the data is too low quality, the phone issues an alert to the midwife and provides appropriate decision support (adjust the device, call the central medical center, etc.).
Geocoded epidemiological data is also collected and made available in the EMR with analysis results to track patient progressed over time. This is achieved via wireless transmission in a compressed, encrypted SMS-format in areas with 1 and 2G access (for alerts and referral), and (for bulky data such as the raw ultrasound recording) through data where 3G (or above) and Wi-Fi are available.
Readings are recorded and cached locally when there is no signal, and transmitted when a connection is established, and referral information is cross referenced to ensure text messages were not lost (as happens in 5% of cases). Personally identifiable information will be stripped and stored on cloud-based servers, although the healthcare provider will retain a link to the patient in their EMR (currently OpenMRS).

Who uses it?

These devices are worn by the pregnant women at the direction of a health care professional, and in the case of our study in Guatemala lay midwives (traditional birth attendants) after a relatively small amount of training. The design allows for ease of use on both the wearer and medical professional end. Navigation of the application is driven by pictograms and audio help recordings, so low literacy rates are not a barrier.
Doctors working at a tertiary or secondary location can access all the information recorded during the medical encounter with the midwife and take appropriate action. Automatic alerts and cross-checks between multiple databases (SMS-enabled messaging systems, the EMR and a secure cloud backup) allow for minimal response times and minimal data loss.

Why does it help?

Our project is innovative because it addresses, for the first time, the systemic barriers to maternal-child care in rural areas posed by the lack of access to basic diagnostic technology, decision support, and improved linkages between lay midwives and higher levels of care. By assisting practicing lay midwives to elevate the level of services they provide, and by promoting better referrals through mobile technology, we are activating and empowering an existing community health resource with a cost-efficient approach that also strengthens community health systems.


Team

Team's Location

USA

Team's Occupation

Engineers

Team Members

Dr. Gari Clifford, MD, Yale Zhang, Dr. Peter Rohloff, MD

Focus Area(s)

Diagnosis/Treatment/Referral, Data Collection/Data Insight

UNICEF Pillar(s)

Health, Child Protection



These pages have been pulled directly from applications submitted to the Wearables for Good Challenge in 2015. They represent the work of the individual teams and have subsequently not been edited.

Comments

Write a Comment

ALL PROJECTS