1 of 18

OFFICIAL PARTICIPANT BRIEFING · 2026

DNA Hack for Health

– Participants

Welcome, builders. This weekend you'll design digital health tools on top of Bangladesh's real national health system, the Shared Health Record. Some of you have never touched health tech before. That's fine. This briefing starts from zero: the problem, the system, your ten themes, and how judging works.

THE PROBLEM

THE SYSTEM

THE THEMES

HOW YOU WIN

Official rules and rubric: dnahealth.co/hackathon

2 of 18

01 · WHY WE'RE ALL HERE

The Problem You're Here to Solve

Before you write any code, you need to understand what's actually broken. It comes down to four things.

Records That Don't Travel

Say you get treated at a hospital in Dhaka, then visit a clinic in Sylhet next year. The Sylhet doctor can't see your Dhaka file. It stays on that hospital's computer, or in a paper folder. Your treatment starts over from a blank page.

A New Identity at Every Counter

There's no single ID number that follows you between hospitals, so each one registers you as a brand-new patient. After a few visits, three different files describe the same person. They don't always agree with each other.

Policy Making in the Dark

The health ministry decides where vaccines and budgets go. To do that well it needs current numbers from the whole country. When data sits locked inside individual clinics, those numbers arrive late or never.

Systems That Can't Talk

A hospital, a lab, and a pharmacy each buy different software. None of it was built to share data with the others. Even fully computerized records stay stuck wherever they were typed.

DNA Hack for Health · dnahealth.co/hackathon

02 / 18

3 of 18

02 · THE HEAD START

The Foundation Already Exists

Here's the part most people don't know: Bangladesh already runs serious digital health infrastructure. These numbers are real and live today.

1.42 Cr

Citizens with a Health ID

Over 14 million people are registered in the national patient index, each with one verified, duplicate-free identity.

99.6 L

Health Records Indexed

Nearly 10 million medical encounters, stored in FHIR, the standard format we'll explain on the next slide.

110+

Hospitals Connected

Public hospitals plus 10 private platforms push live patient data into the national system every day.

India, Estonia, and Kenya built their national systems on these same standards. Bangladesh did too. You get to build on top of all of it.

DNA Hack for Health · dnahealth.co/hackathon

03 / 18

4 of 18

03 · LEARN THE WORDS FIRST

Six Terms You'll Hear All Weekend

Health tech is drowning in acronyms. Here are the six that matter, in plain words. Come back to this slide whenever you get lost.

EMR

Electronic Medical Record. The software a hospital uses instead of paper files. The doctor types your diagnosis and prescriptions into it.

FHIR (say "fire")

A standard format for health data. If two systems both use FHIR, each can read the other's records. Like PDF, but for medical information.

API

A doorway one program opens for another. Your app can use an API to ask the national system a question, such as "show me this patient's history."

Interoperability

A long word for a simple idea: different systems being able to exchange data and actually understand what it means.

Health ID

One number per citizen for everything health-related. It's linked to your NID behind the scenes, but the NID itself stays hidden from hospitals.

HIE

Health Information Exchange. The network that moves records between hospitals, labs, and apps. Bangladesh's HIE is called the SHR.

DNA Hack for Health · dnahealth.co/hackathon

04 / 18

Before we go deeper, let's deal with the vocabulary, because health tech people talk in acronyms and it can feel like a wall. EMR is just the hospital's record-keeping software. FHIR, pronounced fire, is a file format; the comparison I like is PDF. Any computer can open a PDF, and any health system can read a FHIR record. An API is a doorway between programs. When your app needs patient data, it knocks on the API and asks. Interoperability is the umbrella word for all of this working together. Health ID is the citizen's single health number, and notice the privacy trick: it's connected to the NID underneath but hospitals never see the NID itself. And HIE means Health Information Exchange, which is the whole network. Bangladesh's version is called the SHR, the Shared Health Record. From this point on, when I say SHR, I mean that national network.

5 of 18

04 · THE BIG PICTURE

Meet the SHR: Five Parts, One System

The national system has five parts. Each one does a single job, and they pass work to each other like runners in a relay.

MCI

Keeps the master list of patients. One verified identity each.

→

HIE

Moves records safely between hospitals, labs, and apps.

→

FHIR Store

Saves every medical encounter in the standard format.

→

Registries

Lists approved facilities, providers, and clinical terms.

→

EPI / NID

Links health records to national identity and vaccines.

ALREADY RUNNING ON GLOBAL STANDARDS

FHIR R4

ICD-11

LOINC

SNOMED CT

OpenHIE

SMART on FHIR

Each part talks to the others through APIs. Your project can join the conversation the same way.

DNA Hack for Health · dnahealth.co/hackathon

05 / 18

Here's the map of the whole system, left to right. MCI, the Master Client Index, answers one question: who is this patient? It keeps the single national list of identities. The HIE is the transport layer. When a record needs to move from a hospital to a lab or to a phone app, the HIE carries it. The FHIR Store is where encounters actually live, saved in that standard fire format we just covered. The Registries are reference lists: which hospitals are real, which doctors are licensed, what the official code is for each disease. And on the far right, the system plugs into EPI, the national immunization program, and the NID database, which anchors everything to a real person. The chips at the bottom name the international standards each piece follows. You don't need to memorize them yet; we'll go through the important ones slide by slide.

6 of 18

05 · SYSTEM DESIGN

How a National Health System Is Designed

Every national health system, from the UK to India, is built in layers. Bangladesh uses four. Knowing them tells you where your project fits.

CITIZEN & PROVIDER LAYER

What people see on a screen

Everything a person actually opens: patient apps, hospital EMRs, lab software, pharmacy systems, telemedicine.

INTEROPERABILITY LAYER

How the layers communicate

The middle layer that connects everything. It holds the FHIR APIs and checks patient consent before any data moves.

CORE SERVICES LAYER

The shared building blocks

Services the whole country shares: Health ID, the patient index, the encounter store, the terminology dictionary, the registries.

DATA & INTELLIGENCE LAYER

Where data becomes insight

Where data lands. Databases for storage, search engines for fast lookups, dashboards for the health ministry.

DNA Hack for Health · dnahealth.co/hackathon

06 / 18

Think of the system as a four-storey building. The top floor is what people touch: the apps, the hospital software, the lab screens. Most of your hackathon projects will live here. Below that sits the interoperability layer. Nothing passes between floors without going through it, and it has two jobs: translate everything into FHIR, and check consent. If the patient hasn't said yes, the data doesn't move. Full stop. The third floor holds the core services, the shared machinery every app uses: identity, the patient index, record storage, the dictionary of medical codes. And the ground floor is data and intelligence: the actual databases, plus the dashboards where the ministry watches disease trends. When you pitch on Sunday, being able to say which layer your project sits in, and which APIs it calls in the layers below, will make you sound like you've been doing this for years.

7 of 18

06 · CORE SERVICES · PART 1 OF 2

Identity and Trust: Who Are You, Who Can Look?

Healthcare starts with two questions: who is this person, and who's allowed to see their data? Three services answer them. Use these instead of building your own.

Health ID

Every citizen gets one health number, created from their NID or birth certificate. The clever part: hospitals see the Health ID, never the NID itself, which protects privacy. Children get one too, even before they have an NID.

Master Client Index

The master list of patients, with duplicate detection built in. If the same woman registers at two hospitals, MCI works out that she's one person, not two. It already holds 1.42 crore identities.

Consent Manager

Nothing moves without the patient saying yes. Every permission is recorded, has an expiry date, and can be withdrawn at any time. Your app has to work this way too. It's the law.

DNA Hack for Health · dnahealth.co/hackathon

07 / 18

These three services handle identity and permission, and I want you to use them rather than rebuild them. Health ID first. When a citizen registers, the system creates a new health number from their NID or birth certificate. Hospitals only ever see the health number. The NID stays hidden, which matters because your national ID unlocks your bank and your SIM card; it shouldn't be floating around hospital databases. Babies and children get Health IDs through birth registration, so coverage starts at day one. Second, the Master Client Index. Its superpower is de-duplication. Names get misspelled, ages get guessed, and without MCI one patient quietly becomes three records. MCI compares the details and merges them into one. Third, the Consent Manager. Think of it as a permission ledger. Every yes is logged, every yes expires, and every yes can be revoked. When you build this weekend, put the consent step in before the features. Judges will ask about it, and the Cybersecurity Act 2023 requires it.

8 of 18

06 · CORE SERVICES · PART 2 OF 2

Clinical Data: What Happened, In One Language

Three more services handle the medical side: storing what happened, naming it correctly, and recording where it took place.

Encounter Repository

Every visit, diagnosis, and lab result in the country gets saved here as a FHIR record. It works like a national medical memory. Authorized systems can read from it and write to it through APIs.

Terminology Registry

Doctors describe the same illness in different words. This registry holds the official codes (ICD for diseases, LOINC for lab tests, SNOMED for clinical terms) so every system means the same thing by "type 2 diabetes."

Facility & Provider Registry

The verified list of hospitals and licensed clinicians. Every record in the system points back to a real facility and a real provider, so the data always has an accountable source.

DNA Hack for Health · dnahealth.co/hackathon

08 / 18

Now the clinical side. The Encounter Repository is the simplest to explain: it's where the medicine lives. A visit to a doctor, a diagnosis, a blood test result, each becomes one FHIR record in this national store. Your app can read a patient's history from it, with consent, and write new encounters into it. The Terminology Registry solves a sneaky problem. One doctor writes 'sugar problem', another writes 'T2DM', a third writes 'diabetes mellitus type 2'. A computer can't tell those are the same disease unless everyone uses an agreed code, and that's what ICD gives us. LOINC does the same for lab tests and SNOMED for clinical concepts. The registry is where those official codes live. Finally the Facility and Provider Registry. Every record carries the stamp of a verified hospital and a licensed clinician. That's what makes the data trustworthy: nothing in the system is anonymous at the source.

9 of 18

07 · THE WHOLE SYSTEM IN ONE STORY

One Patient, One Visit: Watch the Data Flow

Enough theory. Meet Rahima. She visits one clinic, once. Here's what happens to her data, step by step.

Walk In

Rahima gives her Health ID at the desk. The system checks it's really her and asks for her consent.

→

See the Doctor

The doctor types the visit into the EMR. The software checks every code against the national dictionary, then sends it off as a FHIR bundle.

→

Stored Safely

Her record lands in the national encounter store. A copy reaches the fast search layer within seconds.

→

In Her Pocket

Rahima opens her health app and sees the new entry with her full history. She can share it with any doctor she picks.

→

Seen Nationally

Her name is stripped off. The anonymous data joins millions of records on the ministry's live dashboards.

That's the whole system: one visit becomes better care for her and better insight for the country.

DNA Hack for Health · dnahealth.co/hackathon

09 / 18

Let's follow one real visit from start to finish, because this is the story that makes everything click. Step one: Rahima arrives at a clinic and gives her Health ID. The system verifies her identity against the national index and records her consent for this visit. Step two: the doctor examines her and types the consultation into the clinic's EMR. Before anything leaves the building, the software validates every diagnosis code and test code against the national terminology dictionary, then packages the visit as a FHIR bundle and sends it. Step three: the bundle is written into the national encounter store, and a copy is pushed to a separate search layer so future lookups are instant. Step four: that evening Rahima opens her health app and the new entry is there, alongside her whole history. If she sees a specialist next month, she can grant them access in seconds. Step five: meanwhile her record, with her name and ID removed, becomes one data point on the ministry's dashboards. If a hundred clinics suddenly report the same fever, the ministry sees the spike the same week, not next year. One visit, five steps, and everyone in the chain is better off.

10 of 18

08 · GLOBAL STANDARDS · PART 1 OF 2

Speak the World's Health Language

Why do standards matter? Because an app that speaks the same data language as everyone else works everywhere. Learn these three first.

DATA EXCHANGE

FHIR

Pronounced "fire." The worldwide standard format for exchanging health data. The UK's NHS uses it, US Medicare uses it, India's ABDM uses it. An app that speaks FHIR can connect to any of them, including Bangladesh's SHR.

TERMINOLOGY

ICD-11 · LOINC · SNOMED

Three official code books. ICD-11 gives every disease a code, LOINC does the same for lab tests, and SNOMED covers clinical concepts. Shared codes are what make a blood test in Khulna comparable with one in Dhaka.

IDENTITY

Health ID / MCI

The principle of one person, one record. A single de-duplicated identity sits underneath every other service. Bangladesh already operates this for 14.2 million people.

DNA Hack for Health · dnahealth.co/hackathon

10 / 18

Standards sound boring until you realize what they buy you. FHIR is the big one. It defines how a diagnosis, a prescription, or a lab result is structured as data. Because the NHS, Medicare, and India's ABDM all use it, an app you build this weekend on FHIR could, in principle, run in London or Delhi with minimal changes. That's the pitch you can make to judges. The terminology trio works underneath FHIR. ICD-11 is the WHO's catalogue of diseases; every illness has a code. LOINC catalogues lab tests. SNOMED catalogues clinical concepts in fine detail. Use the official codes and your data is instantly comparable with everyone else's. And the identity standard, one person one record, is what stops the duplicate-file chaos we saw on slide two. Bangladesh's MCI already enforces it at 14 million scale, so you inherit it for free.

11 of 18

08 · GLOBAL STANDARDS · PART 2 OF 2

From Prototype to National-Grade

Three more standards. These separate a class project from something a hospital could actually adopt.

INTEROPERABILITY

OpenHIE & HL7

The reference architecture for wiring many health systems into one national network. Kenya, Rwanda, and India follow it, and so does Bangladesh's SHR. Design within it and your project fits the national plan by default.

INSURANCE & UHC

FHIR Claims

The standard way to submit and process insurance claims digitally. This matters because Bangladesh is moving toward universal health coverage, and someone in this room could build the claims layer.

APP PLATFORM

SMART on FHIR

A way to build health apps that run inside any FHIR-compliant hospital system, the way one app runs on any phone. Its companion, CDS Hooks, shows doctors live alerts during a consultation.

DNA Hack for Health · dnahealth.co/hackathon

11 / 18

Three more standards, slightly more advanced. OpenHIE is an architecture, a blueprint that says: put your patient index here, your record store there, connect them like this. Kenya, Rwanda, and India built their national systems from this blueprint and Bangladesh did the same. If you arrange your project's components the OpenHIE way, it slots into the national system without redesign. FHIR Claims is about money. Universal health coverage means the government or insurers reimburse care, and reimbursement means claims: who treated whom, for what, at what cost. There's a FHIR standard for exactly this, and Bangladesh doesn't have a claims layer yet. That's a genuine gap. SMART on FHIR is my favorite to explain: it's an app store model for hospitals. Build your app to the SMART standard and it can launch inside any compliant hospital system worldwide. CDS Hooks goes one step further and lets your app pop up advice while the doctor is mid-consultation, for example warning about a drug interaction.

12 of 18

09 · YOUR MISSION

The 10 Official Hackathon Themes

Here's the official list. Pick the theme that excites you. Your submission has to fit at least one.

Health FinTech & Wallets

Digital health wallets, save-now-pay-later schemes, and micro-insurance.

Smart Pharmacy & E-Meds

E-commerce for medicine, OCR for handwritten prescriptions, and pill reminders.

Next-Gen Telehealth & AI

AI chatbots as the first line of defense before connecting to real doctors.

Mental Health & Wellness

Mindfulness apps, burnout trackers, and gamified mental health journals.

FemTech & Women's Health

Period trackers, PCOS management apps, and maternal nutrition guides.

Aging & Elder Care Tech

Fall detection wearables, dementia aids, and simplified care interfaces.

Fitness & Food as Medicine

Gamified fitness challenges and AI that scans local food to count calories.

Medical EdTech & Tools

VR anatomy viewers, flashcard apps, and AI medical textbook summarizers.

The Smart ICU

AI-driven monitoring, sensor fusion, and predictive alerting for critical care.

Healthcare OS & Infrastructure

Unified, interoperable systems connecting hospital admin, EHRs, and patient data.

The judging factors are the same for every theme, so pick the one your team can demo best.

12 / 18

This is the slide to photograph. Ten official themes, and your submission must fit at least one. Let me give you a feel for the range. Theme 1, Health FinTech: how do people pay for care? Wallets, save-now-pay-later, micro-insurance. Theme 2, Smart Pharmacy: medicine e-commerce, reading handwritten prescriptions with OCR, pill reminders. Theme 3, Telehealth and AI: a chatbot that handles the first conversation before a real doctor joins. Theme 4 is mental health, which is hugely underserved here. Theme 5, FemTech: period tracking, PCOS management, maternal nutrition. Theme 6, elder care: fall detection, dementia aids, interfaces simple enough for a 75-year-old. Theme 7 treats fitness and food as medicine, including AI that counts calories from photos of local dishes. Theme 8 is education tools for medical students. Theme 9, the Smart ICU, is the most hardware-flavored: sensors and predictive alerts for critical care. And theme 10 is infrastructure itself, connecting hospital admin, records, and patient data into one system. You can blend two themes. A Bangla mental-health chatbot is themes 3 and 4 together, and that's perfectly fine.

13 of 18

10 · MARKET CONTEXT · PART 1 OF 2

The Money Behind the Themes

A quick reality check on money. These themes connect to real markets, and the gaps in Bangladesh are still wide open.

$96 Bn

EMR / EHR Systems

Connects to theme 10. More than 3,500 private hospitals in Bangladesh still keep paper records. Digital records are step one for everything else: analytics, insurance, AI.

$62 Bn

Community & mHealth

Connects to themes 3, 4 and 7. Eighteen crore people, most reachable only by phone. Community health workers are asking for digital tools right now.

$18 Bn

Insurance & UHC Claims

Connects to theme 1. The clinical data already exists in FHIR format, but the country has no standard digital claims layer yet. The government wants one.

DNA Hack for Health · dnahealth.co/hackathon

13 / 18

Why show you market sizes at a student hackathon? Because the best projects here become startups, and it helps to know which problems the world pays for. The gold numbers are global market sizes. EMR and EHR software is a 96 billion dollar market worldwide, and here at home more than 3,500 private hospitals still run on paper. Every one of them is a future customer for somebody. Mobile health is a 62 billion dollar market, and Bangladesh is unusually well suited to it: eighteen crore people, and for most of them the phone is the only realistic channel for healthcare. Insurance claims is an 18 billion dollar market, and this one has a special twist: the clinical data already sits in the national system in FHIR format, but nobody has built the claims layer on top. The government has said publicly it wants one.

14 of 18

10 · MARKET CONTEXT · PART 2 OF 2

Smaller Markets, Faster Wins

Three smaller markets where a focused student team can stand out fast.

$12 Bn

Labs & Diagnostics

Connects to themes 9 and 10. Labs produce the most structured data in healthcare, yet most private labs in Bangladesh aren't connected to anything national.

$4.5 Bn

Patient Portals

Connects to themes 4, 5 and 6. There are 14.2 million Health IDs and still no app where citizens can see their own records. Someone will build the first one.

$9 Bn

Terminology Services

Connects to themes 2 and 8. Shared codes are what make health data usable. The national registry exists but needs friendly APIs built around it.

For scale: the global digital health market is projected at $659 billion by 2030, and Bangladesh is the fastest-growing mobile health market in South Asia.

DNA Hack for Health · dnahealth.co/hackathon

14 / 18

Three smaller markets now, and smaller can be better for a student team because the competition is thinner. Labs first. A laboratory is the most data-rich place in healthcare; every test produces clean, structured numbers. Yet most private labs in Bangladesh aren't connected to any national system, so all that data goes nowhere. Connecting even one lab type, with LOINC codes, would be a strong demo. Patient portals are the strangest gap of all. Fourteen million people have Health IDs, the records exist, and there is still no app where a citizen can open their own medical history. The first good one will spread by word of mouth. Terminology services sound dry but think of it as selling dictionaries to every health app in the country. The national registry of codes exists; what's missing is a developer-friendly API around it. And the bottom line gives you the overall scale: 659 billion dollars globally by 2030, with Bangladesh the fastest-growing mobile health market in the region.

15 of 18

11 · THE FRONTIER

AI in Healthcare: You're in the Right Country

The hard part of health AI is getting clean training data. Bangladesh has it, at national scale. Three directions worth your weekend:

Clinical AI & Diagnostic Support

An AI that reads patient records and flags the urgent cases first. The same idea works for reading chest X-rays or retinal photos. Studies put the potential cut in avoidable hospital stays at 10 to 30 percent.

FEWER AVOIDABLE ADMISSIONS

Predictive & Population Health

Use the 99 lakh existing records to catch outbreaks early or spot high-risk patients before they get worse. The data is sitting there, waiting for someone to ask it good questions.

NATIONAL-SCALE TRAINING DATA

Bangla Health Assistants

Most health AI speaks English. Most patients here don't. A Bangla assistant that checks symptoms, reminds people about medicine, and books appointments would be genuinely new.

$5.5 BN VOICE-AI MARKET

One warning: an AI is only as good as the data you train it on. Get your data into FHIR shape before you touch a model.

DNA Hack for Health · dnahealth.co/hackathon

15 / 18

If your team wants to do AI, here's the honest picture. The model is not the hard part anymore; you can call an API for that. The hard part is clean training data, and that's exactly what Bangladesh has been quietly accumulating: nearly a crore of structured FHIR records. Direction one is clinical AI. An AI that reads incoming records and sorts the queue, sickest first. The same approach reads chest X-rays or retinal photos. Published studies estimate ten to thirty percent of hospital admissions could be avoided with earlier flags. Direction two is prediction at population scale. If fever cases tick upward across fifty clinics in one district, an outbreak model can raise its hand in week one. Direction three is my personal favorite for this room: a health assistant that actually speaks Bangla. Nearly every health chatbot on earth assumes English. Build symptom checking, medicine reminders, and appointment booking in Bangla, connected to real FHIR records, and you have something that doesn't exist yet. The warning at the bottom is serious: structure your data as FHIR before you train anything. Messy data in, useless model out.

16 of 18

12 · BUILD LIKE A PRO

Six Habits of Winning Teams

Teams that do well at hackathons tend to do the same few things. Steal these.

Build FHIR-first

Design your data as FHIR resources from the start. It costs you an hour early and saves a day later.

Use standard codes

Don't invent your own disease or test codes. ICD-11, LOINC, and SNOMED already exist. Use them.

Plug into SHR, don't rebuild it

Patient registration is a solved problem. Call the MCI API and build on 14.2 million verified records.

Consent before features

Write the permission flow first. The Cybersecurity Act 2023 requires it, and judges will check.

Clean data beats clever models

A simple model trained on tidy FHIR data will beat a fancy one trained on a mess.

Think OpenHIE from day one

Fit the standard architecture and your project can plug into a national system, here or abroad.

DNA Hack for Health · dnahealth.co/hackathon

16 / 18

Six habits, learned by watching teams win and lose. One: model your data as FHIR resources before you write features. It feels slow for the first hour. By Saturday night, when your AI needs structured input and your demo needs a realistic record, you'll be glad. Two: never invent codes. If you find yourself typing disease equals quote diabetes quote, stop and look up the ICD-11 code instead. Three: don't build patient registration. It's done. The MCI API exists. Spend your weekend on the part that's new. Four: consent first. Sketch the permission screen before the feature screens. The Cybersecurity Act 2023 makes consent mandatory, and the judges know it. Five: when choosing between a cleverer model and cleaner data, choose cleaner data. Boring, but it wins. Six: arrange your components the OpenHIE way so your demo maps onto the real national architecture. It's the difference between a toy and a pilot.

17 of 18

13 · HOW YOU WIN

The Judging Scorecard: Four Factors

Four factors decide everything. Each judge scores them 1 to 5, and the weights turn that into a total out of 100. Full rubric at dnahealth.co/hackathon

35% WEIGHT

Cost-Effectiveness

The "Budget Genius" Factor

How cleverly did your team stretch a minimal budget for maximum impact? Smart use of open-source tools counts in your favour.

30% WEIGHT

Medical Impact

The "Life-Saving" Factor

Does this solve a genuine, high-priority healthcare problem in Bangladesh? Show the judges a clear unmet clinical need.

20% WEIGHT

Feasibility & Scalability

The "Rural Readiness" Factor

Could this be deployed tomorrow in an Upazila health complex, and keep running after you go home? Sustainability is part of the score.

15% WEIGHT

Technical Execution

The "Professional" Factor

Does the prototype actually work? Was your 3-minute pitch persuasive, and was your deck easy to follow?

35%

30%

20%

15%

17 / 18

Read this slide twice, because it's how you win. Four factors, weighted. The biggest is cost-effectiveness at 35 percent, what we call the Budget Genius factor. Judges want to see maximum impact from minimum money. Built it entirely on open-source tools? Say so in the pitch; it scores. Medical impact is 30 percent, the Life-Saving factor. The question is whether you're solving a real, high-priority problem in Bangladesh. The strongest pitches name the specific clinical need and who suffers when it goes unmet. Feasibility and scalability is 20 percent, the Rural Readiness factor, and the test is wonderfully concrete: could this run in an Upazila health complex tomorrow, with the staff, power, and connectivity that actually exist there? And would it keep running after the hackathon? Technical execution is 15 percent, the Professional factor. Notice it's the smallest weight. A working simple prototype with a clear 3-minute pitch beats an ambitious broken one. The bar at the bottom shows the proportions. Plan your weekend hours the same way: spend them where the weight is.

18 of 18

BUILD FOR IMPACT

Your Code Can Reach

18 Crore People

"The best time to build Bangladesh's health tech future was 10 years ago. The second-best time is today, at this hackathon."

BEFORE YOU START HACKING

Pick a theme and a real Bangladesh gap

Sketch your FHIR data model before you code

Read the four judging factors again. Plan around the weights.

DNA Hack for Health · dnahealth.co/hackathon · Powered by Bangladesh SHR (DGHS · MIS)