A shame to lose this
The primary intention of care.data was to enable the huge volume of information in doctor's systems to be made available for researchers and others. I understand one of the first business cases for this data is actually billing - doctors charge the nhs for appointments and treatments and care.data would replace the existing interfaces.
It has to be mentioned that the quality of data is very variable and often, especially with old data, very poor. For example, the "postcode" may be "Swindon", the patient date of birth 01-01-1900 and the gender not given. Such data would likely be ignored. A bigger problem is with the clinical codes used for describing illnesses (and treatment if I remember correctly). For example, the code for diabetes will have different values on different supplier systems, changes over time and may not be correctly assigned anyway.
That aside, there is real medical value in this data. It would enable national scale queries such as how many appointments are made for an age range with diabetes versus one with copd. It would enable us to know where our money goes potentially leading to better targetting of at risk groups. Another query would be to create a cohort based on those that have copd and have taken drug x long term. There is the potential to identify common side affects (by determining other apparently unrelated medications or treatments given).
As always huge debate is raised by a minority when the majority doesn't give a toss. If I apply for a life insurance policy, it is already based on statistical assumptions; potentially those assumptions will be better in the future.