Unfortunately, I simply don't trust those responsible to treat my data so that it is used for the common good.

"Assumed consent", as I understand things, means giving your consent for your records to be shared and around & sold on to pretty much anyone at all who wants to take a look, for whatever reason.

I started to trust physicians here as much as politicians. They are more interested in their gains and status than patients health. The experience with my late father in hospitals was not good at all.

I won't trust them with my data, nor the the industry behind them which has showed to be greedier than Scrooge.

I have no problem to help others to heal - but I'm afraid data would be used to make physicians even more selfish and arrogant, and the industry even greedier. Death may not be the worst outcome, in such situation.

Indeed, came here to say the same thing: "Sharing your medical records with researchers" sounds so serious and positive: Help the world, cure cancer, bright tomorrows, cue bird and whale song... But that's just sugarcoating, the "think of the children" routine: Actually who you'll be sharing your information with is Big Pharma marketing, so they can better concentrate on the most profit-making affections and better ignore the less commercially interesting ones.

So, apparently that glorious "informed consent" starts with weasel words and half-truths. That's encouraging...

This isn't straight forward at all. I presume most of us would agree to sharing aggregated anonymous data to cure disease and illness.

But, certainly in the UK, there is a already a significant history of people either stretching these definitions or making a mess of them.

It's extremely difficult to row back once this has happened.

As such, I'm a no.

Quote: ".....securely anonymised......"

Key phrase in this article. The proposed NHS data slurp is stated to be "pseudonymised". This weasel word means that the data will be stripped of personally identifiable information EXCEPT FOR DATE-OF-BIRTH AND POSTCODE.

*

With people like Palantir and DeepMind available to match the slurp with many OTHER DATABASES, the data will only be anonymous for a minute or two.

*

As with many well-meaning commentators, the article states an obvious truth -- namely that aggregated health data will IN PRINCIPLE be very useful for managing the health care of large populations. Unfortunately, that other principle -- personal privacy -- is violated when the anonymity of individuals cannot be guaranteed.....and indeed it cannot be guaranteed.

*

If you think this risk is small, take a read through these links:

- https://www.theguardian.com/technology/2019/jul/23/anonymised-data-never-be-anonymous-enough-study-finds

- https://www.bloomberg.com/features/2018-palantir-peter-thiel/

- https://www.theregister.com/2021/09/30/royal_free_deepmind_representative_action_uk/

Yes. Very much this.

And anonymising data is extremely difficult, if you want to do it properly.

And if done improperly, it can be deanonymised.

I'd say that provided the "securely anonymised" holds, then yes, the sharing of such information is potentially very beneficial. Unfortunately I am too old and cynic to believe in that premissa.

I voted yes but you're right, in too many cases the promised (and required by law) anonymization has fallen short, enabled the possibility of de anonymization, and with palantir and their kind involved.... The details of anonymization are extremely important to get right and those details are not regulated well enough (researchers demonstrate they are routinely able to break it, which leads to... zero change) and too complicated for general public discourse to understand. Gonna change my vote to no.

My not-a-doctor take is that birthdate (age) and postcode (location --> environmental factors) are useful information for interpreting medicinal outcomes, the former especially.

Shirley there must be an acceptable third option (happy medium) instead of pass/fail (use or don't use). Is there a level where the data is sufficiently "blurred" to provide a higher level of anonymity without unduly sacrificing correlation with medical findings?

What if birthdate was reduced to birth month? Or even calendar quarter?

What if postcode was combined with neighboring postcodes to reduce location granularity?

In theory there is. Do not share the data at all. Keep it on an air gapped system. Run the queries on the system and return a graph of number ill versus age or a low res heat map of disease incidence.

In practice the UK government (blue or red) leaps at every opportunity to become even more untrustworthy. This sort of project should be kept on hold at least until they grow up.

This.

The aim is to allow research questions to be answered - the method proposed it to essentially sell the raw data, suitably tagged so as to prevent the same individual from being double counted.

The other way to do it is to run an "air-gapped" service which will perform the analysis and present the results, but that's likely to be far more expensive. Those interested in buying the data will no doubt complain that it's too much, and appeal to the powers of capitalism.

"And as for assumed consent for sharing data, it’s important to remember that assumed consent is still informed consent: patients are told that they are assumed to have consented to the sharing of their data for use in metadata analysis and, should they wish to opt out, how to do so. "

Right there is one of the bigger reasons why assumed consent is wrong.

If the assumption is made then there are so many reasons why the effort made to inform fully will be less than perfect.

With assumed consent TPTB already have what they want and will view anything that could reduce their treasure trove as undesirable and make as little effort as possible to do anything that may aid that reduction.

Assumed consent is never acceptable and ethically wrong.

"Assumed consent is never acceptable and ethically wrong."

Except for organ donation perhaps? Before the law change, 80% of the public said they supported the practice in principle, but only 38% of people had formally opted into organ donation, with the majority saying they simply hadn't got round to registering as a donor.

With organ donation the donor is dead and the interests of the still-living outweigh theirs.

I realize the OP is Australian, but here in the UK the NHS has done things like transfer data to Google without proper consent, and given that despicable firm's equally despicable data practices, that means control over the data is forever lost.

As for her contention that data can easily be "anonymized and securely aggregated", it is simply laughable, as all anonymization schemes this far have proven easy to defeat by a determined data scientist.

The way to get that data for important purposes like epidemiology is by earning trust through transparency and accountability, and shady practices like implied consent are the opposite of that.

[1]Did you sign an organ donor card?

[1] https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwipnLiG18LzAhVjlFwKHaw7A-YQwqsBegQIBRAB&url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DSp-pU8TFsg0&usg=AOvVaw09lUvagZayPm5xXpQeDkXm

That's not consent regarding sharing of data - that's something which was pretty well discussed, and has pretty wide ranging benefits and limited downside.

No-one really needs their organs once they are deceased.

There are various groups who would like their bodies to remain intact post mortem - and that's fine, that's why there is an opt out.

This is consent to sharing the most personal of data with completely unknown and unspecified companies for unknown and unspecified purposes.

Given the attempt to shift the NHS to the US model of "care" that data becomes *really* dangerous.

" assumed consent is still informed consent: patients are told that they are assumed to have consented to the sharing of their data "

Unfortunately that's not what 'informed consent' means. It's not enough to be told open ended sharing will or may take place. In the context of consent, 'informed' means (strictly) informed about exactly what will be done - i.e. about the details of the sharing. Under the GDPR at least, if these details haven't yet been defined, it's not possible for a data subject to consent as they don't yet know what they will be consenting to.

The same applies if the details haven't been fully disclosed, even if already defined. But in the three years since the GDPR came into force I haven't seen more than a couple of 'privacy policies' that described data sharing adequately. However, despite this almost universal non-compliance (or perhaps because of it) it appears that the UK is aiming to relax the requirement to provide full information.

Adequately explained or otherwise, I think she's entirely wrong that assumed consent is informed consent, for precisely the reasons you've articulated.

I don't care whether it's lazy thinking, malicious or well-intentioned-but-wrong, the ICO need to assure that organisations holding private data don't make the mistake she's made.

We, unfortunately, live in a country where people hear a handful of very specific petrol stations are running low on fuel and immediately fill up every container they have to hand with fuel they don't actually need. Worse, they put it in to containers that are woefully unsuited to storing it. The information given was quite clear - specific petrol stations in a specific part of the country ran by specific companies were struggling with deliveries . The public took that to be something else, and agreed to accept the falsehood they came to, and we were left without petrol for the whole country, impacting on everyone else.

To simply say that it's absolutely fine to take this data and for people to opt out because they'll be told what the data is used for is just pure bollocks. We have demonstrated time and again in this country that we don't understand what we're told, and often what we're told doesn't align with what we're given. Furthermore, you are assuming there is an understanding of what's being told. I think we've all been in conversations with people who aren't as tech savvy as some here, and they're absolutely amazed that their Faceache news feed is full of adverts for something that they were just discussing with a friend (verbally, not over the app). How are they allowed to do that? They have absolutely no idea that they consented to it when they signed up years ago, and have no idea it's quite clearly pointed out in the T's & C's that they scroll right down the bottom of just to accept to get access to the thing they've signed up to.

Further more, for as beautiful as the NHS is, there is a history of missteps resulting in catastrophic failures for people who rely on it which results in nothing happening. Right now I'm thinking of the Factor VIII/Factor XI scandal that has resulted in too many people dying, or living with a stigma, that is only now being investigated. We're 30 years down the road from politicians knowing the risks, medical bosses knowing the risks, and yet none of them are being held to account for it.

If it's allowed to carry on, we'll be 30 years on from this decision and none of the people promoting this, in charge of it, will ever have to answer to their own mistakes (whether willingly made or not). That is the kicker for me more than anything. We will have to live with the consequences of this, while those promoting this fallacy will be happy in their retirement knowing they won't have to clean up the shit they've helped create.

assumed consent is still informed consent

I fear for her patients...

Why even publish this garbage? To let people know the bar to become a GP has been set too low?

"I fear for her patients"

Given that this item is published as part of a debate it is unfortunate that you have chosen to make such a statement rather than address the weaknesses and failings of the argument. This type of ad hominen attack is the sort of thing that hampers open discussion.

As someone who has offered to open the debate for the motion it is hardly her responsibility to point out weaknesses in her argument even if she is aware of them. Other commentators have already highlighted some of these (assumed consent <> informed consent, securely anonymized is dubious) - I anticipate that the respondent "against the motion" will also make these points.

In the meantime it would be nice if the commentary could focus on the points rather than atack the people who happen to be presenting the arguments.

"it is both appropriate and extremely beneficial for the progress of medical research to know that 50 year old men with prostate cancer who take medication A live for longer than those on medication B."

Extremely useful but should be part of a properly conducted randomised trial. The NHS has made great strides with outcome analysis (in covid treatments) without needing to give data to 3rd parties. The properly anonymised data may at best indicate that a trial should be performed.

A big issue is that once the data has been given out there is no mechanism to later withdraw consent. (unlike organ donation for example)

Why not have internal NHS researchers?

The trials of which you speak (RCTs) which study very small amounts of people are very different from mass data research in a generalised population.

RCTs to develop new treatments cost millions and then to bring new drugs to market from those trials cost billions (typically). Universities cannot afford to run such trials and big pharma is not going to gamble on them for anything other than close to dead certs. Rare diseases don't get a look in.

A mass data analysis is quick, cheap, focussed and doesn't put patients at risk by exposing them either to untested drugs or denying them those potentially life-saving drugs (at the efficacy stage - Phase IIa/b and Phase II of the trial process) if they are on the placebo wing of the trial. This is not comparing like with like. Data research for mass disease is a totally different paradigm.

Assumed consent is not consent, never mind informed consent.

What the hell even is "assumed consent"? It sounds like something a rapist would use as an excuse after putting a roofie in a woman's drink. "Well, she came home with me, so I assumed she consented to sex."

This is right up there with the "Legitimate Interest" buttons on opt-outs. If you ask for an exception to my refusal for legitimate interests, then you admit your regular interests are not legitimate. If you have to assume my consent, that is because *I have not given my consent*.

"She didn't keep an eye on her drink, so I assumed she consented"

Once the data is generated it will be "lost" or "abused" at some stage. The more data is gathered, the more likely it is for loss or abuse to happen. The only secure data is the data not generated and stored.

The whole discussion of data (aggregation) with(out) consent is a distraction. You have lost control once the data is generated and stored by a second or third party. That is the point. You, as a person, are an afterthought when you are not the sole proprietor of your own data on your own systems with your own exclusive access and rules.

These discussions of opt-in/opt-out are post-factum rationalizations. They are only to make one feel good and have ultimately no effect.

If we'd really be talking about the user's choice of opt-in/opt-out, then a) opt-out as initial choice would no exists and b) we'd be asked every time, for each and every thing, to opt-in. We'd also have total control of where any data is stored and accessed, all the time.

Unfortunately, most people simply do not care or do not understand.

If you actively consent (and it is possible to actively consent even within the aegis of 'assumed' consent) because you see a personal benefit or a benefit to the commonwealth then it is less a matter of losing control, rather actively taking control. It is perhaps like stepping on an aircraft. One makes an active decision to handover control in one's personal interest. That is, in itself, a controlled decision.

Automatically opting people in is NOT informed consent. The UK government tried this before. Just forgot to tell us what they were doing. I for one wouldn’t trust them not to try the same again. Those actions call into question the motives behind data sharing. For sure it seems as if the data isn’t really seen as the patients, more like a resource that the medical profession thinks they should control Doing ‘informed consent’ this way is just a way to take something away from the uninformed or those who never get round to it opting out.

Wanting to share our data with everyone is dangerous. Individual patients have no idea who will end up with their data. As for randomising data, HA! At best it’s pseudo random and can be backtracked in the UK. And there’s always the mission creep factor. It’s for medical trials now. But just imagine, that Colostomy you’ve got, how neat would it be if we could offer the patient cheaper stoma care products?

Singapore has many reasons for having better Covid outcomes than other place’s besides data sharing, so comparing Covid outcomes between authoritarian regimes and more liberal regimes isn’t an argument that would persuade me. In fact it strikes me as almost emotional blackmail being used to support a pretty weak proposition.

I have no problem with:

"The NHS cannot analyse all information on its own, so we safely and securely share some with researchers, analysts and organisations who are experts in making sense of complex information. We only share what’s needed for each piece of research, and wherever possible, information is removed so that you cannot be identified." From: https://www.nhs.uk/using-the-nhs/about-the-nhs/sharing-your-health-records/

Why? Well:

I'm nearly finished treatment for Prostate Cancer (They say the treatment should 'cure' me). I had a friend (half my age) who died of cancer last week leaving a wife and two young children.

If my medical history could help others I'd be more than glad. They do need make sure the safeguards are strong and there will probably be breaches, but to help my fellow citizens it seems a no brainer to me.

"societies whose people are willing to widely participate in public health measures like social distancing and wearing masks for the collective good have better outcomes. This is a lesson that translates to health data sharing."

In what way?

In an opt-out system there is an incentive to design the system well, in an opt-out system you have a huge captive audience, so don't have to care.

And anonymisation isn't always possible. 50 is fairly young for prostate cancer. If he's a new case then, combine that with Brian living in London and we've narrowed him down to around 3000 people (1700 new cases per 100k in people aged 50-54. 9 million people in London). Currently there are around 100 Brians born a year in the uk - it was probably higher in 1961, but I still think we have enough data to uniquely identify him.

Leaving aside the assumed/informed consent, this part is key to me.

I don't actually believe the NHS are capable of properly anonymising data, and I'm far from convinced that they'll even try. Which is why I've opted out.

No: even if this is explained on the front pages of all newspapers most people will not understand what it is about. Their eyes will glaze over after a few words and they will look for the latest story about $celebrity.

1. The scheme is opt-in, not opt-out. As mentioned by many earlier commentards, 'opt-out with assumed consent' is a fox-guarding-the-chicken-coop disaster.

2. Absolute transparency about how data are aggregated and anonymised.

3. Absolute transparency about how data are encrypted and secured.

4. Aggregated and anonymised datasets are not able to be / intended to be downloaded or distributed, and that the doing of such is a serious criminal offence.

5. A written public policy (not 'guidelines') clearly defining the boundaries of acceptable use (which does not - and never will - include for commercial purposes.)

6. Absolute transparency about who has access to the data, and for what specified purposes, in granular detail. For example:

- Name of the research group

- Name of the research group leader

- Organisation the research group belongs to

- All sources of funding, both direct and indirect

- Purpose of research, written clearly in plain english

- The name and position / job title of the Civil Servant or Health Service staffmember directly responsible for granting permission for the research group to access the data

7. Audits to ensure that research teams / orgs are behaving themselves.

8. Clearly-defined protections for whistleblowers.

Set the system up with all that and I'm in!

Dr Katherine Hanks main arguments are that the data can be anonymised and is for the public good.

The problem with anonymised data is that it can always be deanonymized. You will not be able to identify an individual from a single data set but once combined with all the other data set out there that an individual falls into the easier it becomes to deanonymize and identify an individual.

What I really have a problem with the public good argument. It is for the greater good, it will reduce health care costs etc. Why would individuals not want to give up their private medical data for research and the benefit of society?

Well, I will when the medical industry also works for the benefit of society and not for profit. That is what will happen from the research on our medical data. From private health care to big pharma huge profits will be made from research on our medical data.

The best way to reduce health care costs for the benefit of society is to reduce the bloody huge profits that are being made by the medical industry.

If you want to use our data for the good of society then for the good of society you should not be making a profit.