COMU1130 Chap.7 Health Data and Platform Promises
Health Data and Platform Promises
Health platforms commonly join a personal promise with a collective one. Individuals receive monitoring, insight or support, while shared data are said to advance research or public health. This chapter traces information from bodies and self-reports into patterns, recommendations and secondary use.
It asks who benefits, which populations are represented, how uncertainty is communicated and what governance turns data donation into a defensible public contribution. Sensors, diaries and interaction traces translate bodies and routines into comparable variables. That translation may reveal a useful pattern, but the score is a proxy rather than the whole condition.
The user often sees a recommendation while the platform controls the categories, model and secondary access. Analyse the diagnosed problem, collected data, inference, recommended action and beneficiary as separate steps. The same data flow changes meaning when access shifts from an individual to an employer, insurer, researcher or commercial partner.
Data donation is organised participation: devices, eligibility, persistence and analytic capacity influence who is represented. A cause-driven project can depend on corporate infrastructure, and a commercial service can still support research, so ownership, funding and governance are better evidence than labels. Prediction should be described cautiously.
Statistical usefulness does not establish a clinical explanation, and validation does not automatically justify a high-stakes action. Ask how uncertainty is communicated, whether users can correct data, whether errors fall unevenly and whether the recommendation is reversible. Public benefit must name the public, the benefit and the accountable institution.
Continuing notice, purpose limits and meaningful choices matter because later partnerships or ownership changes can create uses that were not salient when information was collected.
What this chapter covers
- 01
personal and collective benefit
- 02
sensing and self-reporting
- 03
health inference and uncertainty
- 04
data donation and consent
- 05
secondary use and beneficiaries
Worked example · free
Audit a fatigue-management wearable
- 1Name the diagnosed problem.
- 1Identify sensed and reported data.
- 1Explain the inference.
- 1State the recommended action.
- 1Judge access and secondary use.
Key terms
- Data donation
- The contribution of personal information to a claimed research or collective purpose.
- Information asymmetry
- Unequal access to the data and reasoning needed to evaluate a decision.
- Secondary use
- Use of collected data for a purpose beyond the immediate service interaction.
Health Data and Platform Promises FAQ
What is the double promise of health platforms?
The service promises an individual benefit such as insight or monitoring and a collective benefit such as research. Critical analysis asks whether the same data flow actually serves both and who captures later value.
Why is consent more than a single click?
Users may depend on the service, face changing terms or misunderstand secondary use. Meaningful consent requires understandable purposes, limits, realistic choice and governance across the continuing relationship.
How should a health prediction be described?
Describe the selected inputs, the inferred category, uncertainty and the action attached to it. A pattern can be useful without explaining an individual's condition or replacing professional and contextual judgement.
What makes a collective-benefit claim credible?
Name the public, the benefit and the institution accountable for delivering it. Trace information from collection through inference, recommendation and secondary access, distinguishing what the individual receives from what researchers, employers, insurers or commercial partners receive.
Check who is represented by the devices and categories, how uncertainty is communicated and whether participants can correct data or refuse a later use. Consent at collection is not enough when purposes, partners or ownership can change over time. The safeguard should match the mechanism: purpose limits for secondary access, separation of employer use, validation across affected groups or a correction and appeal route.
This approach can recognise genuine research or service value while refusing to treat more data as automatic proof of public benefit.
Assessment move
Use the five-part health-platform audit: problem, data, inference, advice and beneficiary. Add one representation question and one secondary-use question. When proposing a safeguard, aim it at the mechanism you identified, such as purpose limits for access, correction for bad data or clearer uncertainty around a prediction.
Practise by drawing two arrows after the recommendation: one to the immediate user outcome and one to later research, product or institutional use. Label who controls each arrow. Create an observation-and-inference table for a wearable, keeping sleep duration or movement separate from readiness, stress or risk categories. Beside every category, write what context is missing and what action follows.
Test representation by asking who owns the device, remains engaged, supplies clean data and is absent from the training population. Test consent across time by changing the partner or purpose and asking whether the original choice still has force. For a short response, preserve the personal benefit before explaining a governance condition; this avoids treating all measurement as deception.
For a recommendation, select one control point and defend it. A generic request for ethical AI is weaker than a correction route, uncertainty display, separation of employer access or limit on secondary use. Finish by checking that no sentence offers medical advice: the course task is to analyse mediation, institutions and public claims.