oracle - health data oracle
Problem: Insurance buyers omit risk-/pricing-relevant information and insurance companies lose money. For instance, private health insurance or disability insurance requires disclosure of diagnoses.
Idea: build a service that gathers and analyses health records from health insurance companies.
Product:
- An insurance company offers an expedited onboarding service to a customer. A customer has to provide existing diagnoses and his policy starts preliminarily.
- A customer provides a letter of attorney and a list of current or previous health insurance companies.
- Using the letter of attorney, the service asks the customers's health insurers to send all existing health records to an email, fax or physical address specified by the service.
- Once the data is received, the data is transformed into a structured format and analysed for risk-/pricing-relevant data. The service informs the insurance company whether the service may leave preliminary status or should be adjusted.
Benefits:
- faster underwriting for the customer
- lower claims payments or better risk-adjusted pricing for the insurer
Data privacy concerns:
- a customer needs to entrust his/her health records to the service
- the service needs to provide proof to insurers to prevent liabilitiy charges
Motivation:
- Observation #1: A relevant buyer of private health insurance stays with his insurer for more than 30 years. A policyholder that did not disclose information at the time of underwriting may cost more than half a million EUR per year. Insurance companies are only allowed to change a customers policy because of non-disclosed diagnoses within the first 3 years.
- Observation #2: German health insurance companies must share all of a customer's data if a customer request such information.
- Observation #3: Trusted data oracles will grow exponentially in relevance over the next few years. Data feed infrastructure companies have been being actively built over the last 3-5 years (e.g. Chainlink, valuation in 06/22 of 3 bn EUR).
Why will it not work?
- Data privacy issues are too great.
- Service must have a customer-centric data privacy policy (customer owns everything, service does not save data), while preventing liability charges.
- Sales might go down unexpectedly because many diagnoses are undisclosed in Status Quo.