Patients Know Best - The Power of Access to Health Records

Last week, I came across this post by Patients Know Best CEO, Mohammed Al-Ubaydli, responding to some strong feelings from the medical community about his Personal Health Records organisation. The intensity of the feelings on all sides struck me. As a patient, I’ve found immense value in Patients Know Best; it’s been a crucial tool for my health management. As a technologist, I see the enormous potential of technology to transform healthcare and improve the patient experience. However, even without any medical training, I can also understand the concerns from a medical professional’s perspective.

The main theme of the debate, which played out on Reddit, revolved around the name: “Patients Know Best.” From the doctors’ perspective, the primary argument seemed to be, “No, they don’t!” And, in many ways, they are right. Very few patients can perform their own brain surgery, remove their own spleen, or reach a reliable diagnosis given a set of observations. But this misses the point.

No one ever claimed that Patients Know Best in the operating theatre or GP surgery. The service’s name reflects a subtly different angle. Mohammed Al-Ubaydli himself emphasized that the name is meant to signify the importance of patients having access to their own health data, not that they are more knowledgeable than doctors in medical expertise...

Consider the analogy of my experience with my iPhone. A few years ago when it started overheating after an iOS update, I immediately sensed something was wrong. Apple initially denied any issues, but eventually, the overheating bug was confirmed, and it was global. My familiarity with my phone allowed me to identify the problem - I use it for long periods of time, every day, and have done so for many years; even barely perceptible change is immediately noticeable to me. Did I know more than Apple’s engineers? No. But I knew something was off because I knew my device. This analogy extends to health. Patients might not know more than their doctors, but they do know their bodies.

A similar situation occurred with an electrical certification for my house. The electrician provided detailed test results, most of which I didn’t fully understand - what's an Ohm? is my circuit frequency good or not? Of the three possible wiring settings, do I have a good or bad one? I didn't know, nor really did I care too much to ask, however, seeing the data reassured me that the job was done properly - the electrician was saying "here's the evidence, and if you want to get a second opinion, you have an immutable document that can be audited". This transparency is what builds trust and confidence, much like in healthcare - trust comes from empathy, and empathy comes from truly understanding and adapting to the perspective and needs of the other person, it doesn't come from "hey, just trust me".

As my maths teacher always used to say, "if you make mistakes (and you will) its easier to find where you went wrong when you show your workings".

My Experiences with Patients Know Best

Access to personal health records empowers patients to take an active role in their health. It doesn’t mean they know more than doctors, but it allows them to spot issues and ask informed questions. This is so important, because ultimately, if I'm entrusting my life into someone's more experienced hands, then it's only right that I get to ask "what does this mean"... when I had my central heating boiler fixed last year, my plumber/gas guy not only showed me his work, but explained how the boiler was working; I was reassured.

There are big debates around patient-centricity vs patient-led, and these debates are bigger than I'm qualified to wade into, with the exception to understand that patient-centric ("knowing your patient, looking for the hidden agenda, meeting the patient at a mutual comfortable middle ground") is generally seen as a better approach than patient-led ("the patient tells me their spleen has a [whatever] so I'm going to remove it", to use an extreme).

It must be really difficult to decide whereabouts patients accessing their files sits - on the one hand, patient-centricity says be open, transparent, and answer the questions... even where those questions might be seen as dumb. Seeing test results that are hard for a patient to interpret, could, in theory, lead to increased anxiety: "what does it mean that my HBA1C is in the middle of the reference range? Do I need do eat more HBA1C?!"

On the other hand, I can see how patients could be really powerful care navigators, taking an active role in the management of their own health, by understanding the basics, and by being comfortable with seeing test results.

Here are three detailed examples from my own experiences where having access to my health records made a significant difference.

Sinus Surgery

Earlier this year, I had sinus surgery. The day after the surgery, I found it really difficult to breathe. I'd been expecting difficulty with nose-breathing (such is the nature of sinus surgery), but this was more serious than that - I was gasping for breath. When I called 111, they asked me for plenty of details and promptly sent me to A&E. On arrival at A&E, I explained my symptoms and medical history repeatedly to a string of medical staff. My blood pressure was taken: 165/109. "Is that normal for you?" I was asked.

Obviously, it's not, but it's not far off my normal. The doctor explained that it's probably just some white coat syndrome because people don't really like to be in hospitals. My blood oxygen was taken: ~95%. It's within range. I got discharged with the instruction to let them know if it got any worse, getting an ECG on the way out. I just couldn't get through that this wasn't right. But the doctors simply weren't listening... their machine was telling them I'm okay, and I must be wrong.

I got discharged, and I got in the car and drove to A&E in the next town. This time, I pulled my data from Patients Know Best (PKB), my discharge letter from the surgery, and my Withings data. I used Copilot to pull some of this data together and wrote all of my symptoms down.

Finally, I met with a doctor who sat and read through what I'd written. She asked me if I'd been intubated, and I said I wasn't sure. As it turned out, I had been, and during the insertion, they'd slightly bruised my uvula (the dangly bit at the back of the throat). She took a look and saw that my throat was blocked by this swollen uvula, and she asked an ENT specialist to take a look. About ten minutes later, I was on intravenous meds, steroids, and antibiotics to help reduce the swelling, and everything returned to normal. My blood oxygen went back to the top end of the reference range, my blood pressure came back down to my low-end hypertensive state, and most importantly, I felt like it was fixed.

Now, I'm not saying that I could've known to prescribe steroids to myself, nor could I really explain why I got the treatments I got. At that point, I felt like we'd reached mutual understanding. I couldn't have got there without the data.

Blood Tests

We figured out I needed surgery in the first place because of a long-term series of oversights. I know now that I have chronic sinusitis, and it turned out I also had a deviated septum (the cartilage bit in the nose was bunched up like an S). The issue was causing poor sleep, and because of an oversight years ago, I'd been referred to a sleep clinic, which decided that I had sleep apnoea (albeit very mild - so mild, in fact, that the diagnosis was withdrawn at my next consultation a year later), and then got diagnosed with non-REM parasomnia.

Just none of this felt right at all, and each time I'd had blood tests - which I have to because of my ADHD medication and blood pressure medication - I noticed that I consistently had numbers that were slightly above the reference ranges. The doctor would always say, "Don't worry, it's not far out." Fair enough.

At my test in December, my doctor told me what he'd be ordering; he wanted to take a look at my kidneys (I think), and it was an add-on to the generic tests. I had the bloods taken and was told, "If you don't hear, then that's good news." At the same time, I had to drop off a "sample."

By mid-January, I hadn't heard, and I checked Patients Know Best for my results. I could see the basic ones, but not the extra one my doc had arranged. I called my GP. "If he hasn't called you, you can assume everything's fine."

I don't like to assume a result by omission. When I was a programmer, I knew better than to assume - that's a fast route way to find you're storing integers (numbers) in a string (text) variable, only for your calculations to fail later.

If you're about to work on mains electricity and you shout to your mate, "Is the mains power switched off?" and you hear no response, you don't go and grab hold of the wire on the basis "if there's a risk, he'd have shouted back." You wait until he says, "Yes, it's off," then you'd probably flip a light switch to make sure, and you'd probably use a multimeter too, just to check.

So I called back again, "Sorry no, I've checked Patients Know Best, I can see the results for the [basic tests] but not the [other blood test and urine test], please could you check?"

"Okay, can you hold?"

"Sure."

[...]

"Err, do you know if you had three bottles taken or just two?"

"I'm not really sure... why?"

"Okay, it looks like the extra bottle didn't happen—the results have been processed, but we're going to need to redo the other one, and can you bring a new sample too?"

If it hadn't been for me knowing what I was expecting to see in the results and then having access to the results, I wouldn't have known that the results were missing, and I'd have assumed everything was fine. Maybe my (very overworked) GP would have spotted the issue, maybe not.

Discovering Sinus Issues

I discovered that I needed sinus surgery in the first place, completely by accident. For years, I've been complaining about postnasal drip. I've had serious headaches. I had all the symptoms. Generally, those symptoms were attributed to something else - hay fever, medication side effects, being run-down or stressed.

By luck, I ended up with meningitis a couple of years ago. I say "luck" because I was diagnosed quickly - I couldn't open my eyes without searing pain from the light, move my chin to my chest, and I had a rash that didn't disappear when I rolled a glass over it. Plus, I had a spinal tap which found indicators like elevated Cerebrospinal Fluid (CSF) protein - and I was treated quickly.

Meningitis' etymology is from the Greek "meninx" (membranes) and "+itis" (inflammation), so the most basic definition is that it's an inflammation of the meninges (the protective layers around the brain and spinal cord). It's pretty hard to check the brain from the outside, so I had a couple of brain scans to look for where the pressure was happening and to check that the inflammation was reducing. The good news was that the treatments worked.

When I reviewed the notes afterwards on PKB and the paperwork I'd been given, my eye kept being drawn to an additional comment made by the radiographer, who mentioned an "opacified sinus."

I wrote to my GP, who said he couldn't interpret those results from a radiographer, but that the consultant charged with my care on the meningitis case will have reviewed those words and decided if they were relevant or not - and that in the absence of other feedback from that consultant, we should assume the feedback is not significant.

Here's where it got interesting, because I got connected with a new GP a few days later. I pulled the data from PKB, including the observations and trends around the inflammation markers, including the CSF detail, and including the observation from the radiographer, and I asked to be referred to ENT. That GP did so, and several months later, I sat with an ENT who was keen to review my data and my own views. The GP also referred me for a follow-up spinal fluid exam, which also found unusually high CSF protein, which remains unexplained to this day.  

It turns out that the radiographers comments from earlier were actually an additional helpful comment - not related to the meningitis, but a useful "by the way I saw this whilst I was here", something that the critical care consultant ought not to have cared about, but something being flagged for follow-up later.

The ENT asked me if I had any theories, and I sheepishly explained that I'd been using an AI model to identify trends (this was a few months before ChatGPT came online), and that I'd noticed a strong correlation with the symptoms of chronic sinusitis. I remember when ChatGPT came online, and I was able to convert some of the words from the doctors' letters into standard English, and it totally transformed the capability of the learning model I'd been playing with.

Anyway, you're guessing where this went. The consultant was already there and was way ahead of me. He had his nose camera (nasoendoscope) to look inside my nose, and he found two things - a blocked passage and infection. I got sent for a scan to confirm, and it found a whole load of grim inside my nose. I got scheduled for surgery, which happened eight weeks later.

In this case, I did know best. I'm not saying I knew the medicine - I'm untrained, and I have no license to practice. I felt like when you have to call a first-line "IT help desk", and you tell them your computer is running super slow and feels wrong, and they come back later saying that "we've checked everything and it's all fine" before closing the ticket with "no fault found."

There was an issue, but my difficulty was in describing the symptoms in the right order, with the right words, and with an awareness of the significance of the symptom (e.g., I'd always assumed postnasal drip was something everyone had), in a way that could help the medical professionals to give me the help I needed... indeed when I'd visited ENT probably a decade ago, I realise now that how I described the issue led to the lower part of my nose being checked and cauterised, rather than the inside... we were 1 inch away from finding the problem, but when it was missed, it stayed missed because "that's already been checked".

My complaints recently had been that this daft CPAP machine was uncomfortable because of the interaction with the postnasal drip (I'll spare you the gory detail), which led us down the path of trying to make CPAP work rather than to deal with the postnasal drip. I now know that I should've been mentioning that as the lead symptom, but to me, the headaches were the main symptom (and rather than solve that, I was prescribed amitriptyline on the basis "some people get ocular migraines, and we can't explain it but the meds will help").

So whats the answer?

Seeing the potentially tricky situation that PKB are facing, I could see a load of outcomes.

1. Say "hey, actually Patients Know Best is our mission, and we're aiming for that", and to soften the line around patients actually knowing best. To me, this would look like a compromise.
   

2. Say "actually, Patients do Know Best, and thats what we believe, and we'll accept challenge on that when we can see that Patients don't know Best about their own circumstances"
   

3. Buckle under the pressure, to stop allowing patients access to their results. But what about those patients like me, who found that tests hadn't run properly, or who linked separate results received over an extended timeframe to find a trend that was otherwise missed?

4. Partially buckle, and to prevent the release of results until the GP has authorised the release (E.g. a removal of the automated release)... maybe mandating that the results release must be accompanied with interpretation notes.

5. Double down on patients as a first line resolver. Patients already self-treat for minor illness ("rest, paracetamol, plenty of fluids", "clean the wound and apply a bandage")... why not educate patients about what their results mean? Why not give patients the words to describe different types of pain, different types of headache? Why not just explain what lymphocytes are? If there's going to be patient letters with words like "lobular minor polypoidal thickening in the maxillary antra-bilaterally with ostital occlusion", then why not use technology to solve the obvious next question?

In the new world, where consumerised large language models can translate those words in to human-accessible words, a world where the general public can obtain high quality home-medical equipment from companies like Withings and Dexcom, why not embrace that?

I get that doctors will be getting some really ill-informed questions "my LDL cholesterol is at the bottom of the normal range, how can I increase it", "how can I get my creatinine put back in" might be annoying, especially to a GP who has a finite budget to deal with bigger issues. Are these not easy issues to solve through education? Could doctors and PKB work together to deliver 30-second video training clips, to be played to the patient before they can see their result? Could LLM's help to explain the terminology, even if they don't explain the result? The answer is "yes, they can", and I demoed such an approach over on LinkedIn last year (read more about that here) - and I think the question has moved past the argument of "maybe they can, but should they", and jumped straight into "what's a safe and proportionate way to implement this?".

I get that surgeries are under increasing pressure because the population appears to be eating lower quality food, moving less, getting generally unhealthier, as well as the average age increasing to the point where more complex issues are likely to appear more frequently. But does this need to mean lower quality service? (e.g. a finite amount of butter, being spread over an ever increasing number of slices of bread). Does it necessitate the empowerment of the end-patient? To what extent could this exponentially increasing issue be solved by exponentially responsive technology?

When we consider the exponential value add of cloud services, IOT products, deep learning / machine learnings / large language model LLMs, we could empower patients to take more control of their own health... to get automated explanations of the concepts behind blood test results from large language models (saved to their record in case they need to escalate the conversation to a GP, to avoid repeating everything).

We could (with the users' consent) capture all that great data stored in Apple Health - long term blood pressure readings, pulse rate over time, activity over time, spo2 over time, estimated vo2 max, respiratory rate... and aggregate that with the point-in-time records created adhoc at the GP surgery.

We could empower patients to carry out home test: imagine a self-test scenario using the old system of jellybabies and finger sticks, or even continuous glucose monitors, to run rudimentary home-oral glucose tolerance test, where the patient could buy those things for themselves without bothering the GP if they chose, and where a simple algorithm could run a diabetes risk assessment. (I did demo this about a month ago, and the total cost was £50.99 per test, with the added bonus of 10 additional days of blood sugar monitoring, which taught me some interesting things about my eating habits! You can read more about that here on Linkedin)

We could do all of those things, but that requires a collaborative approach between medical professionals (which I'm not), patient advocates like PKB, and a whole load of trust. It requires deep engagement, between the doctors, the PHR companies, and patients, to make sure not to unnecessarily increase workloads: but to accept increasing workloads if it means finding and solving root causes of issues before they become chronic... a sort of spend-to-save approach. Most of all, it requires a continuation of the demonstration of the benefits of these incredible innovations, iteratively adding features to PKB, continuing discussion with Doctors forums, and continually assessing the impact on the end-users experience and end-users health, because ultimately, that end user, the patient, is who this is really all about.

Conclusion

My examples highlight the value of patients having access to their own health data. Granted, on a sample of one, and granted without any really useful medical background, but nonetheless a perspective, which you may agree with or disagree with.

While patients may not have the medical expertise to diagnose or treat complex conditions, they often know when something is wrong with their bodies, and in my case, having access to this data through PKB totally changed my life trajectory.

By leveraging their own health data, patients can play a more active role in their healthcare, ask informed questions, and ensure that important details are not overlooked. The transparency and accessibility provided by platforms like Patients Know Best empower patients to collaborate more effectively with their healthcare providers, ultimately leading to better health outcomes, surely a core pillar in the future of providing empathetic healthcare.

And if we can get comfortable with users providing their own data, interacting with a deep-learning model fronted by a large-language model AI... imagine a 24x7 medical bot that provides immediate reassurance, timely escalation triage, on a text-chat for folks who prefer not to pick up the phone, which can capture and probe the patients symptoms ahead of a GP or A&E visit? That could change our experience for the better, forever.

What do you think?