Health Data Awareness

This is part of the Pre-MOOC track and requires only basic digital literacy.

Steps

Step 1: Types of Health Data (20 minutes)

1. Study the health data spectrum:

   Type	Definition	Example	Risk Level
   Identifiable	Can be linked to a specific person	"John Kamara, DOB 15/03/1990, HIV positive"	Highest
   Pseudonymized	Identifying info replaced with codes	"Patient #4472, HIV positive"	High (can be re-identified)
   De-identified	All identifiers removed	"45-year-old male, HIV positive, Freetown"	Medium (small populations risk re-identification)
   Anonymized	Impossible to re-identify	"Urban male, 40-49, HIV positive"	Low
   Aggregate	Group-level statistics	"HIV prevalence in Freetown: 2.1%"	Lowest

2. Exercise: Classify these 5 examples into the correct type:

   • "Mrs. Aminata Koroma tested positive for malaria on 12 February"
   • "Patient MK-2024-0847 tested positive for malaria"
   • "A 35-year-old woman in Western Area tested positive for malaria"
   • "Malaria positivity rate among women aged 30-39: 12%"
   • "35-year-old woman, only female teacher at [school name], tested positive"

3. Key insight: The last example shows that de-identification can fail in small populations. Context matters.

Step 2: Patient Rights (30 minutes)

1. Study the core patient data rights (common across GDPR, Malabo Convention, and most national laws):

   Right	What It Means	Health Example
   Right to be informed	Know what data is collected and why	Patient told their blood test results will be entered into DHIS2
   Right of access	See your own data	Patient requests their full medical record
   Right to rectification	Correct inaccurate data	Patient's blood type was entered incorrectly
   Right to erasure	Request deletion (with limitations)	Patient wants STI test result removed from system
   Right to restrict processing	Limit how data is used	Patient consents to care but not research use
   Right to data portability	Move data between providers	Patient transferring from one hospital to another
   Right to object	Refuse certain data uses	Patient refuses data to be used for AI training

2. Research your country's data protection law:

   • Does your country have a data protection act? (Check: UNCTAD tracker)
   • If yes: What is it called? When was it enacted? Which body enforces it?
   • If no: What international framework applies? (Malabo Convention, GDPR for EU partners)

3. Write a 1-paragraph summary of the data protection landscape in your country.

Step 3: Informed Consent Design (30 minutes)

1. Study the 7 elements of valid informed consent for health data:

   • Purpose: Why is data being collected?
   • Data types: What specific data will be collected?
   • Duration: How long will data be stored?
   • Access: Who will see the data?
   • Use: Will data be used for care only, or also research/reporting?
   • Rights: How can the patient access, correct, or delete their data?
   • Voluntary: Patient can refuse without affecting their care

2. Design a consent form for this scenario:

   A medical school is conducting a pilot project to digitalize student health records. Students will have their immunization history, blood type, and chronic conditions entered into a new electronic system. The data will be used for student health services and may be used in anonymized form for a research paper on medical student health patterns.

3. Your consent form must:

   • Be written at a reading level accessible to all students
   • Include all 7 elements above
   • Include a clear opt-in/opt-out mechanism
   • Be no longer than 1 page

Step 4: Spot the Violations (20 minutes)

1. Read these 5 scenarios and identify the data protection violation in each:

   Scenario A: A nurse takes a photo of a patient's wound on her personal phone to show a dermatologist colleague on WhatsApp.

   Scenario B: A hospital IT department migrates patient records to a new system. The old hard drives are thrown in the general waste bin.

   Scenario C: A research team publishes a paper about a rare disease case at a specific rural hospital. The paper says "a 23-year-old female patient" but the hospital only had one female patient with that condition that year.

   Scenario D: A community health worker enters patient data into a Google Sheet shared with the entire project team, including administrative staff.

   Scenario E: A medical school uses student patient encounter logs (with patient names) as training data for an AI model, citing the institution's general research consent.

2. For each scenario, document:

   • What right or principle was violated
   • What should have been done instead
   • What the potential harm could be

Step 5: Care vs Research vs Reporting (20 minutes)

1. Understand the three purposes of health data collection:

   Purpose	Legal Basis	Consent Needed	Example
   Clinical care	Necessity for treatment	Implied (emergency) or explicit	Recording vitals in patient chart
   Research	Informed consent or ethics board approval	Explicit, specific, documented	Using patient data in a clinical trial
   Public health reporting	Legal mandate	Not always required (statutory reporting)	Reporting cholera cases to district health office

2. The gray areas:

   • Quality improvement: Is it care or research? (Usually care, but check your institution)
   • Secondary use: Data collected for care, now wanted for AI training (needs new consent)
   • Aggregate analytics: Usually acceptable if truly anonymized, but "anonymized" is harder than you think

3. Write a 200-word reflection: Think of a situation in your clinical experience or education where health data was collected. Was the purpose clear? Was consent obtained? What would you do differently now?

Artifacts

You must produce all 4 artifacts to complete this skill:

1. Data Classification Exercise — Correct classification of all 5 examples from Step 1 with explanation of why the last example is problematic
2. Country Data Protection Summary — 1-paragraph summary of your country's data protection landscape (law name, year, enforcing body, or absence of law and applicable frameworks)
3. Informed Consent Form — 1-page consent form for the medical school digitalization scenario, including all 7 required elements
4. Violation Analysis + Reflection — All 5 scenarios analyzed (violation identified, correct approach, potential harm) PLUS 200-word reflection on data use in your own experience

Assessment Criteria

Passing score: 8/12 (at least "Adequate" on all criteria)

Common Mistakes

1. Confusing de-identified with anonymized — In small populations (rural clinics, rare diseases), "de-identified" data can still identify someone. True anonymization is harder than removing names.
2. WhatsApp as clinical tool — Widely used, rarely compliant. If your institution uses WhatsApp for clinical communication, that's a policy discussion worth having.
3. Assuming consent is one-time — Consent for clinical care does not automatically extend to research, AI training, or publication.
4. Ignoring paper records — Data protection applies to paper too. An unlocked filing cabinet is a data breach waiting to happen.
5. "We've always done it this way" — Historical practice is not legal justification. Many common health data practices violate data protection principles.

Related Skills

• Prerequisite: Digital Literacy — Basic digital skills foundation
• Parallel: Gen AI Basics for Health — Understanding AI privacy implications
• Leads to: FHIR Resource Basics — Structured health data standards
• Leads to: Digitalize Paper Records — Apply data protection when digitizing records
• Leads to: Regulatory Landscape Analysis — Map the full regulatory environment

References

1. WHO (2021). Global Strategy on Digital Health 2020-2025. World Health Organization.
2. African Union (2014). Convention on Cyber Security and Personal Data Protection (Malabo Convention).
3. UNCTAD (2024). Data Protection and Privacy Legislation Worldwide. https://unctad.org/page/data-protection-and-privacy-legislation-worldwide
4. Vayena, E. et al. (2018). Machine learning in medicine: Addressing ethical challenges. PLoS Medicine, 15(11), e1002689.
5. Abouelmehdi, K. et al. (2018). Big healthcare data: preserving security and privacy. Journal of Big Data, 5(1), 1-18.
6. El Emam, K. et al. (2011). A systematic review of re-identification attacks on health data. PLoS ONE, 6(12), e28071.