Overheard: a medical ethics expert
On a train to Watford (at about Bushey) I overheard a conversation between an older grey haired gentleman and his female junior as they caught up on the issues of the day.
They were discussing (favourably) a colleague who was passionate about data and how it an be linked together to generate new conclusions. Understandable. As an IT guy I am often defining, detecting or deducing patterns or extracting data according to patterns and this is indeed very useful in all sorts of circumstances. Naturally their conversation moved on to the challenges of doing this with medical data, not least for reasons of confidentiality. I was familiar with the outline of the anonymisation methods that senior was explaining to junior, having read articles about them in the Communications of the ACM, but I made an effort to stop listening as I realised I was earwigging rather atrociously.
It’s a funny situation around a table on a train. One assumes that the people around you don’t give a fig about your business and you certainly have an expectation that people won’t earwig because your business is generally none of theirs. At the same time you wouldn’t attempt to seduce someone, discuss an embarrassing medical complaint or hatch a plot around the table on a train, because people cannot actually prevent themselves from hearing what you say and could start listening at any moment whether it’s their business or not. Depending on your plot, of course, your hatching of it could very well be the business of the person next to you so – one expects – most people interested in hatching plots would not do it on a train.
Our expert on medical ethics, on the other hand, was dealing with a topic that effects precisely every British citizen and perhaps he could have kept his opinion to himself, rather than impressing it upon his younger female companion in a public place.
This gentleman was harking back to his chairmanship of the “BMA” or some committee of it and in context I assume he meant the British Medical Association. He expressed regret that large amounts of useful data could not be examined by researchers just because the subjects of that data – patients – could not feasibly be contacted to give cast iron consent. He was wondering, entirely openly, about whether “most people” would “agree to some level of data” being shared with researchers. Of course he meant anonymously but he had already described the problem that reverse engineering the identity of that man with AIDS from Milton Keynes is fairly easy because if you know enough about a group of people and it’s a very small group then it’s easy to work out who you mean and though it’s very “complicated” to anonymise the data while keeping it useful some good people are working on squaring that circle. So, jolly well then, perhaps we should set up the NHS records system so that people had to opt out of his data processing scheme rather than opting into it, just like with organ donation. He felt that it was such a shame that his profession was not being treated like adults, that we don’t trust that he does in fact know better than patients about what to do with their medical records.
Oh why, oh why, did I let him get away with saying such a disgusting thing in my earshot. It may in fact have been politer to call him a monster to his face than write a blog post now explaining why I feel that way, but I’m not perfect.
At least with organ harvesting one might make the case that the owner of the organs doesn’t need them any more and this takes some of the emotion out of the idea that claiming communal ownership of something so uniquely yours is a grim violation of the principle of self-ownership. This man was advocating, in public, the idea that deeply personal information about your life be communally owned and processed for the common good before you are done living, unless you have applied in writing to be left in alone in privacy. The hubris of taking on that kind of responsibility, by default, is bad enough without adding casually that you should be allowed to do so because you know best.
My take is that this is another example of why need, rather than value, is the wrong quantity to measure when making decisions. Just as organs are to be appropriated according to need from people that need not and the unique value and love that might be felt for a persons body is not measured, the data about a life is valued according to the needs of researchers not the value of the data itself.
Money is supposed to be base and materialistic but one can easily imagine a positive role for it here. A free post-internet society might choose a system where patients owned their medical data and were able to select an IT service provider from a market of interoperable services, some of which allowed them to share it if they wanted to. In such a system, I guarantee the price of medical data would be well known and there would be no question at all about how much people would want to share with researchers – it would be precisely the amount they actively selected to share and the system would suit more or less perfectly the needs of every individual patient. Such a system would treat the patients, not the medical bureaucrats as adults, and place it’s trust in them.
I thought twice about publishing a post sharing the private conversation of a stranger. I would not want anyone to do the same to me, but it seems that what was discussed was a form of a plot. It was the idea to hatch a plot of highly controversial ethical nature and it was everyone’s business including mine. If this person feels that we now know too much about him and someone might work out who he is, perhaps he might reconsider his opinion, or at least get the message that his idea is not okay.
Simon Gibbs
Originally published at LibertarianHome
-
1
February 17, 2012 at 09:41 -
Declaration of interest: I am a retired orthopaedic surgeon and am currently researching in a totally unrelated area that requires anonymisation of personal data.
I see your concerns, but do not agree. Let’s take a silly example. Do I ‘own’ the fact that I am 6ft 2ins tall? Can I withhold that information from people who are compiling national average height data for those engineers designing the next generation of train and want to be sure it is big enough?
In fact pooling of clinical data has been around for a very long time. In the old days blood taken from pregnant women as part of their clinical care after the tests for their care were done, was stripped of all personal data in such a way it could not be tracked back, and used for epidemiological testing, that is to see what percentages of the female population had various abnormalities. This was invaluable information, but it is my understanding that AIDS activism stopped it stone dead, precisely when the information about how many pregnant women carried HIV would have been useful. Done properly with proper safeguards (which basically means such large numbers that it is impossible to work out the identity of an individual) I think it would be of benefit.-
2
February 17, 2012 at 10:01 -
How on earth did the Victorians ever create railway carriages without all those computers to tell them what size seats to use? And you’re supposedly a clever bloke.
-
3
February 17, 2012 at 11:07 -
Now you get into the interesting bits about design. The Victorians knew what it was that they didn’t know. The result was that they over engineered just in case, a Victorian train was far heavier than it needed to be. That meant it was by modern standards very inefficient. It paid off in some respects, I live in a Victorian house with two foot thick walls which is great from the heat loss point of view-pah I spit on your cavity wall insulation!
PS As for being clever, orthopaedic surgeons are held by the rest of the medical profession to be as strong as horses and almost as intelligent.-
4
February 17, 2012 at 12:01 -
“….a Victorian train was far heavier than it needed to be.”
Wrong.
The Victorians worked within the limitations of the materials they had available (basically, cast iron, wrought iron, brass, bronze and several species of timber). Had they had access to alloy steels, aluminium alloys and composite fibres, they would have used them. The vehicles the Victorians built were as effecient as those materials allowed them to be – and in many cases they also turned out to be lasting, as you point out.
The Victorians didn’t always know what they didn’t know. There is a string of railway disasters attributable to lack of knowledge of the properties and limitations of materials of construction, the lessons of which informed later engineers. (My rather imperfect understanding of Victorian medical history suggests that it advaced along a rather similar path.)
As a method of advancing human understanding, making mistakes may not be the ideal way of learning lessons, but it has been a very fruitful source of knowledge. That is not in any way a reason to allow people to deliberately make mistakes – in any field of human endeavour. See Chernobyl for details.
-
5
February 17, 2012 at 14:47 -
Quite right, Engineer. It’s a pity the name of David Kirkaldy is almost forgotten nowadays. And ,a href=”http://en.wikipedia.org/wiki/Joseph_Whitworth”>Joseph Whitworth who devised the optimum screw thread by measuring thousands of screws and set the standard. If the Victorians over-engineered things, why did Whitworth need to invent the thou or thousandth of an inch for manufacturing tolerances in 1844?
-
-
6
February 17, 2012 at 15:44 -
I still don’t see how the size of the seats within a railway carriage relates to it’s strength unless they were stressed members of the frame. Would be quite interesting in a crash though.
Yes, I know people who live in Victorian terraces who rely on their neighbours’ central heating to keep their house warm. As an ex-surgeon, I doubt you’d be living in a terrace though
-
7
February 17, 2012 at 15:56 -
With regard to the last, a semi on a main road. I didn’t do private practice.
The Victorian era was quite long, yes they mad mistakes and learnt from them, but yes, when in doubt they over engineered, which is why a lot of their stuff lasted. As for the development of medicine it was by empirical trial and error, it wasn’t until John Charnley that we started to learn from engineers, and the current problems with metal on metal hips suggest we still have a way to go. -
8
February 17, 2012 at 16:26 -
As for the ‘lasting’ Victorian stuff – only the stuff that lasted, lasted. They built a lot off stuff that didn’t. The history of Victorian boiler development is littered with loud bangs and widows….
I gather that with hip joints, the current flavour is the use of ceramics for one of the wear surfaces. Lubrication is (as I’m sure you are better aware than I) something of a problem – fitting people with grease nipples may not be a viable solution.
-
-
9
February 17, 2012 at 17:42 -
“I live in a Victorian house with two foot thick walls which is great from the heat loss point of view-pah I spit on your cavity wall insulation!”
Whilst I wouldn’t question your expertise about orthopaedic matters, I fear your knowledge about space heating, thermal inertia and the insulation values of various materials leaves a lot to be desired. Consider why the the string-vest, with its pockets of air, is such a good insulator when worn beneath an outer shirt.
Your solid walls conduct the heat to outside. Very effectively.
Your house requires a considerably greater-output boiler to provide comfort, than a modern house of similar size. By a factor of 3x – 4x greater, at a guess.
-
-
-
10
February 17, 2012 at 12:18 -
“Done properly with proper safeguards (which basically means such large numbers that it is impossible to work out the identity of an individual) I think it would be of benefit.”
And you are, therefore, free to give your data to whomever you wish. You are NOT free to appropriate mine without my express consent.
If you’re a clever man, you’ll understand the difference, I’m sure.
-
11
February 17, 2012 at 15:57 -
To which I can only make one answer ‘Neigh’
-
12
February 17, 2012 at 18:59 -
And this attitude is the reason why I have insisted that the medical practice with which I am registered do not share my information electronically with any other organsiation.
-
13
February 19, 2012 at 13:10 -
Don’t believe it. I used to do drugs analysis for the National Treatment Agency and the data available to me was horrific. Analysis has to add value and i’m not yet convinced that public health data isn’t just a bullshit boondoggle. But then again I wanna legalize drugs so maybe i’m myopically biased.
-
-
-
-
-
14
February 17, 2012 at 09:43 -
No plot, just the usual patronising ‘we know best what’s good for you’, or ‘you can’t be expected to understand’.
-
15
February 17, 2012 at 10:37 -
I think it is clear that people should be expected to share thier data to a limit. It is clearly the right thing to do if it can help others and in a society with a shared culture and values, one would expect a certain level of societal pressure in this direction.
However, it should never be compulsory or assumed (with an opt out). Choice is the key. We are all free to be a B*stard if we want.
-
17
February 17, 2012 at 12:09 -
I regard any medical notes or details relating to me and my medical conditions to be confidential between myself and the physicians treating me. Any use of that data – which I regard as personal to me – for any other purpose without my express written consent is at best arrogant, dowright rude and a breach of confidence, and at worst theft of personal data. However ‘anonymised’.
-
18
February 19, 2012 at 13:18 -
Yup, in an ideal world, maybe but err..no, it’s not. If the analyst or coder (or more usually tosser of a manager) hasn’t been trained and don’t know what they’re doing – it’s public. Cross reference it with the electoral register, mosaic data, experian info and you’re located. A lot of effort sure but quite easy.
-
-
19
February 17, 2012 at 12:15 -
“If this person feels that we now know too much about him and someone might work out who he is, perhaps he might reconsider his opinion, or at least get the message that his idea is not okay.”
Spot on! There has to be consequences for these people.
-
20
February 17, 2012 at 13:04 -
I agree with Mike on this and there seems to be a touch of hysteria creeping into the debate. If the data is anonymised (and it can be) it is no more an infringement on your freedom that having your movements ‘monitored’ by one of those people on station platforms with counters.
-
21
February 17, 2012 at 13:23 -
Before the doctors appropriate personal information given to them in confidence for the sole purpose of medical treatment for another purpose, they might at least have the courtesy to ask first.
-
22
February 17, 2012 at 13:28 -
I don’t see why. In some instances they would have a duty to make the information public, if you had a highly contagious disease for example, and I think that that principle is correct. It seems to me absurd that the S should be forced to keep secret important information such as the rate of prostrate cancer in men over 40 without first asking each of those men if it is all right to mention it.
-
23
February 17, 2012 at 14:36 -
We seem to be perfectly capable of quantifying the amount of tax gathered through various types of tax levy without compromising anybody’s personal tax data. The same principle can surely apply to medical data gathering. GPs and hospitals can gather data on how many cases of particular conditions pass their portals without compromising personal records – the two should remain seperate.
No bank manager would reveal their customers’ financial affairs except if required to do so by police with a warrant. What bank managers seem perfectly capable of doing, doctors can do also.
To disclose another person’s personal information without asking their permission is downright arrogant.
-
24
February 17, 2012 at 17:51 -
What’s wrong with a bank manager revealing that he has, say, two customers whose accounts exceed £xxxx, and 10 customers whose accounts never go into the red?
What if you go to the doctor with an illness. If that’s the 1st set of symptoms that doctor’s ever experienced, then it might take ages to diagnose & treat you.
But if that doctor recognises the symptoms from data shared by previous patients, he may be able to swiftly cure you.
-
-
-
-
-
25
February 17, 2012 at 14:17 -
26
February 17, 2012 at 14:37 -
A sceptical attitude to the value of statistics is fine SadButMad, but epidemiology really does rely on this sort of info and it really is very useful. I think the idea that your privacy has been invaded because you were one of 30,000 reported, anonymised cases of chickenpox is daft and the idea that we don’t need to understand patterns of disease not much less so.
-
27
February 17, 2012 at 15:37 -
It’s not about being one of 30,000 cases of chicken pox, it’s about being male, 6’1″ tall, having blue eyes, living in Milton Keynes, catching chicken pox as a boy, and an embarrassing medical conditional as an adult. As the expert conceded, once you know enough about someone the group of similar people shrinks to a group of one. It’s easy to work out who you mean, and if you mean to embarrass someone specific or verify a rumour then it gets easier.
Then of course, disclosing the data for anonymation implies some extra functionary or IT process must have access to the raw unanonymised data to start with and the /process/ of anonymisation must also be protected, a quite separate problem.
Without marke competition and self interest being factors for the people concerned, I don’t trust them.
-
-
28
February 17, 2012 at 15:50 -
“No bank manager would reveal their customers’ financial affairs except if required to do so by police with a warrant”
You think that, do you?
Go and read the money-laundering regulations that were brought in by (I think) the so-called “Serious Organised Crime” act(s).
-
29
February 17, 2012 at 15:54 -
“As the expert conceded, once you know enough about someone the group of similar people shrinks to a group of one. ”
That seems to me to concede the point that there is no principled problem, just a practical one of how much anonymisation needs to be applied.
In fact, the sort of detailed info you suggest would never be needed. and it is very difficult to imagine a situation where anyone would be motivated enough to try to find one anonymous individual among a massive data set on, for example, heart disease on the NE of England. Why would you, even if it were possible?
-
30
February 17, 2012 at 15:57 -
“GPs and hospitals can gather data on how many cases of particular conditions pass their portals without compromising personal records – the two should remain separate.”
‘Cases of particular conditions’ is just a long-winded way of saying ‘patients’. If they can use this information unproblematically, without ethical principles being violated, the fundamental point has been conceded and all we are left with is a technical discussion about how much anonymisation is enough.
-
31
February 17, 2012 at 16:20 -
Wrong. There is a clear difference.
A hospital will know how many hip operations it has performed without accessing any patient data.
I repeat – and I am genuinely amazed that anybody should think otherwise – personal data given to a doctor in confidence should remain in confidence. To reveal any such data about any patient without that patient’s express permission is a breach of confidence. Full stop.
-
-
32
February 17, 2012 at 16:25 -
Mike: “Let’s take a silly example. Do I ‘own’ the fact that I am 6ft 2ins tall? Can I withhold that information from people who are compiling national average height data for those engineers designing the next generation of train and want to be sure it is big enough?”
Under the current system, no you can’t; or rather you can’t get information that has been recorded by the NHS out of you medical records. I have no option but to submit to that information being used in any way that the NHS sees fit, however intrusive, inappropriate etc.
With the admittedly trivial example you gave, there is relatively little harm in letting that information out, but suppose your records contain instances when you were worried about your mental health (justifiably or not) – you have to rely upon the probity and decency of your doctor. The organ retention scandal at Alderhay put paid to public trust in doctors (rightly so in my view). We should all engage in civil disobedience and refuse to divulge to our NHS doctors, information that is not necessary for the treatment of the specific complaint about which we are consulting.
A proper system, one which was not a monstrous nationalised edifice would allow me to determine what information I wished to allow others to see; every patient would own their own information, it would not be deemed to be public property.
Hands off my medical records, hands off my organs, hands off my taxes.-
33
February 18, 2012 at 18:54 -
‘We should all engage in civil disobedience and refuse to divulge to our NHS doctors, information that is not necessary for the treatment of the specific complaint about which we are consulting.’
Surely it is in the interests of the patient to answer honestly any questions a doctors asks them. If the patient already knows what information is relevant to their particular complaint, eliminating all potential complaints with similar symptoms, I have to ask what they need the doctor for?
-
-
35
February 17, 2012 at 17:51 -
There is another point to be made. If Sir Richard Doll (as he became) hadn’t abstracted the data from patients records linking the fact that those who smoked got lung cancer, the health of the mass of the population would be significantly worse. Some mention has been made of hip replacements. Which ones failed was identified because of anonymised data from patients records. Should we go on putting in (name omitted for legal reasons) joints just because we can’t look at individual patients data? Anonymised data tells us which performs best long term (and no, so far, it isn’t ceramic, it is the good old metal on high molecular weight polyethylene, currently the best performers are, to the best of my knowledge, the Charnley and the Exeter hip replacement. We wouldn’t know that without anonymised data from patient records. Frankly no-one is interested if was Joe Snooks who was 27 stone, smoked got lung cancer and whose hip replacement failed. We do want to know that smoking gives you lung cancer and that hip replacements in overweight patients don’t do so well.
-
36
February 17, 2012 at 19:04 -
Please refer to answer to Torquil Macneil above.
(Just as a minor point, I wasn’t suggesting that ceramic inserts were ‘the best’, merely that that is where research seems to be currently directed. There are still a couple of drop-stamps in Sheffield banging out titanium implants, I gather; one assumes they’ve had a bit of a clean-up before they reach your colleagues.)
-
37
February 17, 2012 at 23:03 -
Not the discredited Sir Richard Doll?
And regard smoking and cancer, isn’t there more evidence being gathered that it’s got quite a lot to do with genetics as to whether you get cancer or not.
-
-
39
February 17, 2012 at 19:04 -
MIke, what if you are an insurance company?
-
40
February 17, 2012 at 19:43 -
No, I am not an insurance company, but I see where you are coming from. The reason I blanked out the name of an implant in an earlier post was because of a confidentiality agreement regarding a failed implant. I could wax loud and long about manufacturers who want to introduce a new joint replacement, on the basis of very little experience. Funny old thing, good follow up from anonymised patient records eventually sees them off, otherwise they would still be selling them. (ANNA, do you want a post on this?)
In retirement after thirty plus years in medicine, I am a student doing what I always wanted to study.
Turning to titanium, it is an unbelievable metal; bone accepts it like no other and actually seems to integrate with it.
-
-
41
February 18, 2012 at 00:26 -
Do most people care at all?
-
43
February 19, 2012 at 00:53 -
http://www.economist.com/debate/days/view/806
On the same theme.
This house believes that society benefits when we share personal information online.
{ 43 comments… read them below or add one }