COVID-19 Data Intelligence System in Bangladesh

Anir ChowdhuryPolicy Advisor, a2i, ICT Division/Cabinet Division/UNDP

High risk case identifications

Hot zone identification

Resource Allocation:Health Workers, testing facilities, hospitals, isolation units, ICU, etc.

Timely policy response

Symptom monitoring of health care workers and field administration

Supporting Vital Decisions

of the Government for COVID-19

Response

3 Important Steps

1. Encourage citizens to do self-reporting and expedite finding high-risk cases for testing

2. Identify hot zones as early as possible for disease progression

3. Overnight create the largest pool of doctors for telemedicine

Use Big Data Analytics to Support Steps 1 and 2

Data Source List

• IVR Triage (333 & 16263)

• *3332# USSD Triage

• Doctors Verified (333 & 16263)

• IEDCR Heat Map

• Home Quarantine data

• Web user data (corona.gov.bd, coronatest.today, etc.)

• App users data (DGHS App, 20 most popular App integration, Chat-bot)

• Field reported data• BRAC Health Worker (app)

• Community Clinic data (app)

• Front Line Worker Tracking Data (SMS)

• Report from targeted SMS (teachers, imams, postal workers, social safety nets workers)

• DHIS2 Data

• DGHS Hospital Management Data

#First Questionআপনার শ্বাসকষ্ট

আছে? 1.হ্যা 2.না

#Second Questionআপনার জ্বর বা কাশি আছে? 1.হ্যা 2.না

#Third Question

গত ১৪শিছন জ্বরকাশি/শ্বাসকষ্টসহ্

কাছরা কাছে এছসছেন?

1.হ্যা 2.না

#Fourth Question

গত ১৪শিছন কছরানাছত আক্রান্ত কাছরা কাছে এছসছেন?

1.হ্যা 2.না

#Fifth Question# আপনার বয়স? 1.৬০ এর নীছে 2.৬০ এর উপছর

Self-reporting through 333and *3332#

Self-reporting through Web and App

Citizen Self-report Doctors’ Pool(4,000 doctors)

Medical advice e-Prescription

Data Analytics

Dashboard for Healthcare Providers(DGHS, IEDCR, public

and private)

Medical decisions

Dashboard for Govt. Administration

(Cabinet, field admin, police)

Admin decisions

Frontline Workers Self-report

Multi-stakeholder Communication

corona.gov.bdTV, radio,

community radio, phone, social media

COVID-19 Collective Intelligence System

Report by Community Health

Workers

Report from COVID-19 Tests

Verification by Doctors

High-risk casesHotzones

Recommendations

Data collection Data analysis Decision making

Data Process Flow

Citizen self-report Big data analyticsHigh Risk Case

identification through doctors verification

Tested by DGHS and IEDCR

3.7M 17,058 2232,216

Current: 48 hours, Target: < 24 hours

2,216 high-risk cases were identified by doctors and 223 tests have been conductedCommunity report

Just started

Volunteer Doctor Pool App of4,000 Doctors Created Overnight

Sample of Dhaka City COVID-19+ Cases

March 26 April 5 April 15

Possible hot zones can be identified before testing is done

High-risk suspected patients progression

Geo-fence in the lockdown areas Geo-fence for home quarantine

Geofencing to track violators

Tracing the violator of the lockdown/quarantine

Hospital capacity and suspected patients load management

Isolation Bed CapacityIsolation bed and

suspected patients comparison

Suspected PatientsRequired Isolation

bed

Resource mapping may be helpful for early preparation

Researchers

Psychologists

26.14%, 57

21%, 45

14%, 31

23%, 51

35%, 7

5%, 11

0 10 20 30 40 50 60

No response

Denial

Improve

Low risk

Moderate risk

High risk

No response Denial Improve Low risk Moderate risk High risk

Series1 57 45 31 51 7 11

Called by Doctors 14/04/2020

39%, 26

24%, 16

11%, 7

11%, 7

3%, 2

20%, 13

0 5 10 15 20 25 30

No response

Denial

Improve

Low risk

High risk

Irrelevant

No response Denial Improve Low risk High risk Irrelevant

Series1 26 16 7 7 2 13

Called by Doctors 15/04/2020

Observation: People are scared when a doctor calls for verification of symptoms. They have more trust when a psychologist/surveyor with good communication skills call

Partners

• Telecom: Posts and Telecoms Division, Robi, GP, Banglalink, Teletalk, National Telecom Monitoring Centre, Bangladesh Telecom Regulatory Commission

• Data: a2i, ICT Division, Directorate General of Health Services, Institute of Epidemiology, CMED, Cramstack, Nascenia, Brainstation, MCC, SoftBD, DataSoft, iDare, Audere, BRAC, Fordham University, Rochester University

• Epidemiology: DGHS, IEDCR, Harvard, Columbia, U California Berkley, FAO, UNFPA

• Behaviour science: GP, Beehub, Yale, IPA, BIGD, Dhaka University

• Telemedicine: 16263, Maya Apa, PulseHealth, DaktarBhai, Tonic, Olwel, Platform, Bangladesh Doctors’ Foundation

Thank You