My bread'n'butter is patient systems. I cannot think of a way to say this that is not patronising or condescending on some level. Your suggestion will not work. At all. On any level. If we implemented a system like this people would die. Hospitals and Doctors would be sued to oblivion, and rightfully so, but only if they don't succumb to bankruptcy due to malfeasance and incompetence first.
Ethically building a patient administration system requires that you collect relevant information, that is patient-centric, and that identifies fully the patient, their needs, their location and their past episodes of care.
> At this point, if you care about your privacy, you really only have the option of not engaging with the system. Or engaging as little as possible.
The spectrum between privacy and effectiveness with patient and clinical systems is simple and stark. If you want absolute privacy you will receive care that will be truncated, untailored, unfit and unlinked.
I don't think the system can work in the ideal way I would like - the system's priority is the system itself, regardless of the wishes of the individual. As you say, doctors would be sued. I would say that suing, etc relates to the culture that has grown up around the system. And perhaps the medical system was always this way - ie system-centric, rather than patient-centric. With the onset of the 'IT age', we get to see that system-centric fruit fully 'ripen' such that everyone's data can be seen by anyone.
You might say - as I do - that it is for the person who is being treated to be the collector of their information. A different culture would have grown had we had the patient at the center of our considerations.
I'm sure people have been warning all the way down the line, about the issues over privacy. It is frustrating that at every turn the expediency over the system overrides all other concerns. After, what 100 years(?) we find ourselves where we are no, with the patient having no control (and often no knowledge) of the issues that relate to them, while the system has it all. I would go further and say that this outcome is by design, as those running the system - those who govern us - would prefer that we were ignorant while they are informed. Its pretty dystopian anyway.
I'd argue this, but I recognise I am in a privileged position of knowing how the systems work. HIPAA (in the US) and similar legislation in other countries actually puts a lot of power in the hands of patients - if they know how to use it.
You are never going to be in the situation where the patient controls their data completely (what happens if they are non-responsive, drunk or psychotic and unable to legally consent to an intervention they desperately want). I think we shall move towards access to data, and possibly interpretation of that data, but not ownership. Control of the data will always come with caveats, as the body paying for the treatment will always assert their right to how the data is transmitted to minimise costs.
Are you sure your expertise in this area is up to date?
Even with "unlinked": Signal can do discovery without the server storing a social graph. There's nothing stored on a server for a hacker to ransom.
Rank speculation, but I'd imagine there are current designs even at Apple that protect user's health data, and those designs contradict your claim of a shrug-while-we-work dichotomy between "privacy and effectiveness."
Edit: clarification, added a barb :)
It's a constantly changing field, but Health data systems are not the fastest changing field. Change is bound to lower pace layers (governance, usually) and so the capabilities of these systems are ground slowly but incredibly fine.
> Rank speculation, but I'd imagine there are current designs even at Apple that protect user's health data, and those designs contradict your claim of a shrug-while-we-work dichotomy between "privacy and effectiveness."
Cool. I always enjoy new ideas for how to do things. If you can solve these issues, you would upend a HUUUUUGE industry overnight and would be richer than Bezos.
Ppl will likely die because of this leak, I don't think there was any informed decision weighing the risks of both approaches before making the choice.
Every year I see leak after leak in medical data, I am only an developer not in the industry, but the nature of the leak has always indicated massive incompetentence and poor design or ops, rather than than real need for patience care.
Storing data in public S3 bucket has nothing to do with patient care, unwilling to invest to upgrade from windows XP has nothing to do with abiltiy to give care. Ransomware attacks because of aging stack has nought to do with privacy or care.
There is very very little need to aggregate data given the risk of leaks, and very little value to any single patient when his data gets aggregatated. A system not influenced by pharma or software industry would make it incredibly expensive or hard to collect aggregate data, while keeping individual records usuable
The harsh reality is hospital administration and regulation tends to be skewed with doctors, it is the same with legislators who tend to be lawyers, or professional management tend to be sales than tech. None of these groups understand the nuances of tech always to manage or regulate them better, sometimes they are smart or trust their advisors and get it right, but it hit and miss.
Data is money today , so industry is not interested in better safety of medical data either, preventing aggregate collection will loose a lot of value for med tech companies.
I'm reading this as a header declaration.
You are arguing a range of items without actually touching on my point - that health data patterns require a link from encounter and observations back to a patient record. I think that you are arguing a number of these points with incomplete information (i.e. that you have little understanding of the pressures which result in the continued existence of Windows XP within the health ecosystem), and on other points you are completely missing the point (storing patient data on the cloud is a whole different argument).
> Data is money today , so industry is not interested in better safety of medical data either, preventing aggregate collection will loose a lot of value for med tech companies.
The data has relatively minimal value for the makers of patient record systems. The value of these systems are in the licensing and ancillary/secondary services.
I don't need to have an understanding, I am just a regular Joe who's data security your industry does not give a thought about and his data leaked constantly[1].
After so many leaks, I do not see an acknowledgement of systemic issues. I see no discourse on how the industry can step up and improve their security practices. I see no significant update to regulation in response to any of this.
If your industry cannot move fast enough with the consumer tech industry then you should not be basing your tech stack on a OS like XP in the first place , XP always was going to have limited window of mainline support.
Core banking is perfectly happy using COBOL and zOS for the last 50 years. [2]
If you don't believe the architectural choices the industry makes and attitude towards tech has anything to do with the data security problems there is nothing I can say to convince you.
[1] I am focusing on one thing only security not privacy or anything else, i.e. unintentionally sharing data.
[2] it is not that fin tech is without its problems either, however in general fin tech has responded to security concerns.
Effective care requires knowledge of patient history, co-ordination between providers, etc. Your general practitioner needs to identify you to a downstream specialist using a strong-identifiable source. Your hospital requires knowledge of your allergies, of your heritage, of your past medical history, etc, in order to ensure you receive the best therapies and medicines. This does not even touch of things such as funding models or holistic healthcare plans.
The best any care provider would be able to do, in lieu of a trustworthy patient data source, is to offer general medicines and stabilisation.
At least in this mental health practice, you charted on paper / locally to the providers systems. We had a billing team (in old model) that only uploaded key patient info (MRN / DOB / Service date / time / billing code) to a billing system. This was done using billing slips that went to someone trained on it (who could handle all the weird hassles around eligibility etc etc).
Provider would actually talk to patients etc etc. The notes were pretty focused afterwards. And yes - pretty secure - it was actually hard to just browse around - providers locked up their notes in their office, and their office was locked (we had a three lock rule -> front door with alarm (metal gate), office locked, file drawer locked.
What people don't tell you is with the electronic health record, they also required that all these required elements be "charted" - on EVERY visit. So in new system pretty soon charts become a pages of boilerplate per visit - copy and paste stuff. It was not useful anymore to do a case history from this chart.
So we made the admin folks happy, but what a trash experience - the EHR folks are selling bills of goods and the admin folks implementing and the compliance folks mandating endless boilerplate and macro entries to get anything approved are wasting all our time and money.
I got out, never felt better - the people were great, the ordered from on high decade old java app through a double VPN that worked 50% of the time with no user experience thinking - a nightmare.
And in this new system, every sysadmin could basically snoop on anyone (all in the name of patient coordination) as could lots of others who just ended up with the extra permissions because the system was so locked down we all basically had to be superusers. To run billing reports you needed access to every record - so they had to give you these ridiculous permission levels. And to avoid QC issues billing staff are reading all the chart notes to make sure all the required boilerplate is in them. Such BS. This system was integrated with a lot of providers, so you could also see everything related to that person everywhere (all the office / admin / billing staff at literally every provider included). While a bit helpful if you had a billing issue (same unit type billed at an unapproved frequency) all the benefits focused on that type stuff (admin ease) and privacy was zero across that backend system.
My organisation has done many EHR/EMR, PAS and EMPI implementations using systems from third party vendors (i.e. I don't have any dog in the which-system-is-best fight). With modern systems (i.e. applications developed during and after the HIPAA act) we have seen significant reductions in error rates relating to drug interactions and dosing.
> What people don't tell you is with the electronic health record, they also required that all these required elements be "charted" - on EVERY visit. So in new system pretty soon charts become a pages of boilerplate per visit - copy and paste stuff. It was not useful anymore to do a case history from this chart.
I did not know we were meant to keep this a secret. Yes, it is important that you record every encounter with "boilerplate" code. There are funding impacts (as case complexity rises with these lifestyle and prior history codes). If a patient presents with a cough and blood in the sputum, it is important to know that they were once a smoker. If a patient presents with a broken leg and they have a history of diabetes, their treatment will be considerably more difficult and expensive.
> And in this new system, every sysadmin could basically snoop on anyone (all in the name of patient coordination) as could lots of others who just ended up with the extra permissions because the system was so locked down we all basically had to be superusers.
I have worked on these systems as well. I find more recent systems to be far better better at accountability with regards to patient record access tracking. It has been a long evolutionary path, however.
The patient wants time with their clinician. So in a 20-30 minute visit either the clinician has to be busy charting away (we called them progress notes) or you get actual patient time and more summarized to the point notes. Why can't charts include decision useful info.
For example, if a patient has 5 objectives in their plan, why make EVERY 20 minute encounter go over all these objectives yadda yadda - is it unfair to expect the clincian will be working on these issues during their visit. Now every progress note has to have. We discussed patients success and potential tools to help them improve in their ability to control X, including various strategies and evidence tested techniques shown to demonstrate good results in this area. We discussed patients success and potential tools to help them improve in their ability to control Y, including various strategies and evidence tested techniques shown to demonstrate good results in this area.
This was govt delivered / managed care BTW - so in reality, the EHR made clear that they gave to craps about patients or outcomes and just wanted to check boxes, so everything was both insanely detailed and offered no decision usefulness.
The tell is that most of these systems have been total top down forced (from national level). That's a good good sign there is not much natural desire for them.
In other fields you get bottom up adoption of tools - a proof in the pudding point of evidence that folks find value in them (beyond the admins).
Yeah, MH is an odd beast. I think the slowness of health data change is exacerbated by the complete lack of funding to this area (at least in countries with public MH services).
I'm sure that it makes a difference to some people, but so does the vulnerability of privacy. How many people will simply not see a doctor because they're worried about this?
Also your healthcare providers kept their fully-identified records.
And, yes, it was signficantly less effective with higher error rates, etc.
> How many people will simply not see a doctor because they're worried about this?
Very few, I suspect. None that I have ever heard of outside of people with significant anxiety disorders (this is literally true, and not my implying anything).
I.E. software engineers can become ... real engineers!
Is it though?
Its only a conundrum because of the way we have organised data, with NO regard for the people the data comes from. We could have had data collected WITH consent only.
The privacy boat left the port long ago - personal privacy was intentionally eroded, and this is just one example.
Personally, on ethical grounds, I could never have designed a system like that, with so little regard for others, externalising private information so that a permanent record is kept. Once private data is exteralised, who knows where it ends up, even if it is in the name of some good cause (eg an improvement in therapy). Plenty of people warned of these sorts of outcomes.
At this point, if you care about your privacy, you really only have the option of not engaging with the system. Or engaging as little as possible.