Gestational Blood Sugar Tracker (GBST)

Gestational diabetes mellitus (GDM) is a serious pregnancy complication, in which women without previously diagnosed diabetes develop chronic hyperglycemia during gestation. In most cases, this hyperglycemia is the result of impaired glucose tolerance due to pancreatic β-cell dysfunction on a background of chronic insulin resistance. Risk factors for GDM include overweight and obesity, advanced maternal age, and a family history or any form of diabetes. Consequences of GDM include increased risk of maternal cardiovascular disease and type 2 diabetes and macrosomia and birth complications in the infant. There is also a longer-term risk of obesity, type 2 diabetes, and cardiovascular disease in the child. GDM affects approximately 16.5% of pregnancies worldwide, and this number is set to increase with the escalating obesity epidemic. 
According to the most recent (2017) International Diabetes Federation (IDF) estimates, GDM affects approximately 14% of pregnancies worldwide, representing approximately 18 million births annually[1]. Risk factors include overweight/obesity, westernized diet and micronutrient deficiencies, advanced maternal age, and a family history of insulin resistance and/or diabetes. While GDM usually resolves following delivery, it can have long-lasting health consequences, including increased risk for type 2 diabetes (T2DM) and cardiovascular disease (CVD) in the mother, and future obesity, CVD, T2DM, and/or GDM in the child. This contributes to a vicious intergenerational cycle of obesity and diabetes that impacts the health of the population as a whole. Unfortunately, there is currently no widely-accepted treatment or prevention strategy for GDM, except lifestyle intervention (diet and exercise) and occasionally insulin therapy—which is only of limited effectiveness due to the insulin resistance that is often present. While emerging oral antidiabetics, such as glyburide and metformin, are promising, concerns remain about their long-term safety for the mother and the child [16, 8]. Therefore, safe, effective, and easy-to-administer new treatments are sought. 

1. What are the legal implications for using client generated medical data for autonomous diagnosis? 
2. Are women comfortable talking to a machine about gestational diabetes mellitus issues?
3. Which artificial intelligence algorithm yields the best recommendations based on blood sugar and patient data available? 
4. What is the impact of dietary modification and physical activities in the control of blood sugar in women at risk of, and with GDM?

This project aim to develop and pilot test an artificial intelligence mobile application that detects, tracks and suggests treatment for gestational diabetes mellitus using tests based on global standards and a localized behavioural lifestyle modification intervention.

Specific Objectives

1. To evaluate the role of artificial intelligence algorithm in the prevention and early detection of GDM
2. To describe the preference of pregnant women on the choice of illustration used or GDM in the artificial intelligence App
3. To evaluate the impact of lifestyle modification (diet alone vs diet and physical activities vs physical activities alone) on the blood glucose levels amongst pregnant women
4. To describe the perceptions and concerns of pregnant women and healthcare providers on the use of artificial intelligence in the diagnosis of GDM and storage of data
5. To determine the diet quality of the women at risk of GDM.
6. To evaluate influence of lifestyle behavioural modification intervention on GDM or risk in pregnant women.

The solution is targeted at women during pregnancy and six-weeks post partum. The migration of qualified medical personnel to other countries such as the United Kingdom has left many healthcare centres under staffed. This situation denies pregnant women of adequate medical attention. In some areas of Ibadan, the care of pregnant women is left in the hands of Nurses only thus complications or threats such as gestational diabetes mellitus (GDM) are detected very late. 

The pilot stage is within Ibadan metropolis where there are primary healthcare centres, secondary healthcare centres and tertiary healthcare centres. The project will use one healthcare centre from each level. 

The solution is a mobile app that uses AI algorithm to detect GDM using blood sugar results and other personal details. The solution is available on the Android platform because it is supported by low and high-end smartphones. Pregnant women use the mobile app to get an interpretation of their blood sugar tests. If the results are good, no treatment is suggested. If the results indicate the patient is on the borderline, diet and exercise suggestions will be made.  If GDM is identified, the patient is sent to see a specialist at the tertiary hospital.  

The GBST team is a multidisciplinary group made up of an artificial intelligence expert, obstetrics and gynaecologist consultant, human nutrition and dietetics professional and a legal practitioner.  

The team is supported by nursing staff, medical laboratory staff and research assistants.

The obstetrics and gynaecologist consultant along with the human nutrition and dietetics professional engage with the pregnant women on a weekly basis in the ante-natal clinic. They have access to the women and know how long it takes them to see a medical doctor. They also see the impact of treatments suggested by friends in the absence of access to a medical personnel.   

For this MIT Solve solution, the artificial intelligence (AI) model has been developed using 15 features of the pregnant woman. The mobile app is in its second version of development. 

None. The mobile application is in development. Two versions have been developed so far with changes being made to the user interfaces. 

The product is borne out of research. There is a need to develop a business model that will sustain the product and increase its use by pregnant women in locations where medical personnel are limited. Development of a finanancial plan for sustainability is also important. 

Our claim to innovativeness comes from 

1. The use of artificial intelligence to provide suggested interventions to manage gestational diabetes mellitus 

2. the ability to monitor patient activities using existing technology

3. the triage of GDM status using international standards with available medical data

Health information systems exist that can be integrated into the GBST app for data on patient routine and lifestyle modifications that will corroborate or deny information provided during treatment at the healthcare facility.

Access to patient generated data will revolutionize treatment has additional non-medical data that contributes to healing is provided. Health Information Systems will adapt to include the data.   

The impact goals for GBST are:

1. Encourage pregnant women with or without GDM to be conscious of their diet, exercise and well-being of the unborn child
2. educate pregnant women about GDM using slides/voice notes/etc within the mobile app
3. integration of the mobile app into the ante-clinic routine 
4. access to data (diet/exercise/sugar levels) by medical personnel using an Application Program Interface (API) made by our team 
5. integration of data from wearable devices into the GBST mobile app

The impact of this solution will be measured by

1. Mortality rate attributed to gestational diabetes mellitus (GDM)
2. Number of downloads from the Google Play Store
3. Review ratings on the website or Google Play Store
4. Proportion of women aged 15-49 years who make their own informed decisions regarding sexual relations, contraceptive use and reproductive health care
5. Proportion of pregnant women who own a mobile telephone 

The Theory of Change for Gestational Blood Sugar Tracker (GBST) is based on the migration of qualified medical personnel to other countries thus increasing the patient to doctor ratio. Pregnant women require constant monitoring hence the biweekly appointment at the ante-natal clinic. Availability of medical personnel to consult with pregnant women is limited, so non-communicable diseases are not detected early. Most pregnant women have a smartphone for communication and entertainment. There is the need to ensure care for pregnant women in a timely and personal manner. Using the smartphone with the pregnant women, it is expected that the gestational blood sugar tracker app would do the triage based on available data and offer interventions as required. The change required is for pregnant women to use the smartphone to record their activities using text and images. The data collected from the recorded activities help to determine how to treat the gestational diabetes mellitus (GDM). medical personnel have access to the data either in chart form or text format.  

This solution is powered by artificial intelligence (AI) models that are used to predict the gestational diabetes mellitus (GDM) status of a pregnant woman. The incorporation of the AI model within a mobile app built for the Android platform.

The AI model uses classification algorithms to predict the GDM status of pregnant women using 15 variables that are collected during registration, hospital visit and medical laboratory visit. 

The team is gender-balanced. There are two (2) women and two (2) men in the multidisciplinary team. The Project coordinator is a lady and works full time on the project. 

There are christians and muslims in the team. All activities are done to ensure full participation irrespective of faith. 

Opportunities to talk about the project, travel on the project are shared amongst team members. 

The team is now well-bonded and moving on to create the impact desired. 

The solution provides value to pregnant women by ensuring the blood sugar is monitored and the child is alive. The need to reduce number of deaths from non-communicable diseases such as diabetes is the social good. The sustainability of the service will come from in-app purchases, medical referral and product licensing.  

The team will need help about sustainability. 

The current thought is to license the mobile app to hospitals/maternity clinics for their own use. This would be the Free-for-Service business model. 

In-app sales or advertising would also help. 

The project is currently sustained by a grant of thirteen thousand euros (€13,000) from the French Embassy in Nigeria.