Heath Tech: Avoid or Mitigate Violations with Advocacy Solutions

Hang on…. this is a long post.

BUT! It’s full of “how-to’s” you can implement with clients right away. It’s worth your time, I promise!

Picture this:

You or your client are undergoing your annual OB-GYN checkup, or your mammogram, or your colonoscopy, or….  and in the middle of the exam you realize that your doctor, or the tech helping out, or someone in the room is wearing “smart glasses” – you know: those glasses that have teeny-tiny cameras and microphones implanted in them. 

You realize your exam is being video-recorded. No one mentioned this as a possibility. No one asked for your permission. It’s a HUGE violation of your client’s or your privacy. Now what do you do?

Here’s another one:

An older patient who has upcoming surgery scheduled, is sent to a cardiologist to get approval for the surgery. (This is required prior to any surgery for many Medicare patients.) The cardiologist scheduled the patient for a stress test, which she passed easily after running on the treadmill for 7 minutes.  Later she learned that her results got mixed up with someone else’s, and were reported incorrectly in her record, indicating she had failed. That put her surgery in jeopardy. She calls you for help.

Or this one:

Your client has been diagnosed with something new, potentially life-ending, and must make treatment decisions. All of the options seem difficult; none of them even seem worth living for. So he jumps onto AI, inputs some of his medical test results, his diagnosis, his opinions, and asks AI to help him choose the treatment option best for him. AI’s reply is opposite of what his doctor has told him, and now he’s asking you what to do. What do you tell him?

One more:

You get a call from a previous client, distressed that he has been informed that his insurance will not cover the treatment he recently received, citing decisions outlined in his medical record.

Your work with him had covered only the medical navigation itself – not his billing needs – but you tell him you’d be happy to take a quick look at his records. He provides you with access to his MyChart where you pull up records that don’t sound familiar to the experience you know he had. There’s a good possibility that what you’re reading is also affecting payment. 

So what happens next? 

Every one of these scenarios happens every day.

In the example of the woman who went for a medical procedure, she realized that the provider was wearing AI enabled glasses – but not a word had been mentioned to her. What about her privacy? Wasn’t her permission required? Isn’t that illegal?

In the second example where the stress test results were swapped between patients…  That happened to me. In the summer of 2024, I was diagnosed with renal cell carcinoma and was told I needed surgery to remove the tumor from my kidney. Because I’m on Medicare, I needed a pre-surgery physical. It included a stress test (heart), and I was able to run the full (required) 7 minutes. Yet, a few days later, I received a call that my surgery was on hold because I had failed the stress test. WHAT? That started my journey of getting my test results corrected and my surgery approved. The cardiologist did correct the mistake, and I ultimately had the surgery. But it took 3 months to get the correction made to My Chart, which I believe, even today, may become important sometime in the future.

The third example, a client who relies on AI for making medical decisions; this happens every day, too often to the patient’s detriment.  According to the New York Times, there are currently NO AI models that are truly ready to provide accurate information – much less good decisions- to patients. 

And finally, the fourth example:  technology may be costing you or your patients plenty. What most of us don’t realize is that many health systems and medical networks are now utilizing AI to draft appointment notes, discharge summaries, lab result analyses, prescription writing, and much more. Importantly, and as was the case described above, AI is used to assign codes for medical billing (DRGs and CPTs). There is no way to tell whether any of that work is being reviewed or even whether prompts have been written to skew the resulting codes to higher levels of billing, etc. This makes it imperative that someone – the patient, patient’s billing manager, or you, the advocate, review the results step by step.

How You Can Help Your Clients or Yourself

Don’t Panic. Prepare Instead. 

There are so many problems that can get in the way of your healthcare, or healthcare for your loved ones OR your clients. In fact, I’ve lived in the world of medical errors now for 22+ years, and not a week goes by I don’t hear about another safety or privacy violation.

What is unique about THESE problems is that they are all caused by technology. Every one of them, including the swapped medical records. (Because, yes, it’s always been possible records could have errors or be assigned to the wrong patient. But the fact that they were swapped in MyChart made it SO much more difficult to get them corrected.)

None of them are – yet – illegal. There is a general understanding about technology that states that the law can never keep up. New technology has always been used before it can be regulated because regulation is always REACTIVE. We have to identify the problem before it can be regulated, repaired, adjusted, or whatever it will require.

So since the law isn’t there to help us, our most effective approach will be to avoid negative possibilities before they happen, at least to the extent we can.

How to Help or Mitigate Such Violations Before They Happen

The overall goal of these questions is to manage one’s own expectations; knowing full well that it’s rare when a provider will address them on their own. They may not think to, or they may be much more concerned with spending too much time with any one patient.

These can get you started with expectation management. You’ll probably (and should) think of more:

• Ask Questions

Provide a list of questions for your client to ask their doctor or surgeon prior to any upcoming procedure/test, treatment, or surgery.  You can include some generic questions such as, “Is this test required to make a diagnosis?” or “would you recommend this surgery for your own mother (or father)?”, or “what are the side effects”, or ?  Those questions will make a great handout for you to supply on which you will, of course, include your branding.

Make sure you include on that list any questions you think are important or relevant specific to the reason for the test, procedure, or surgery. These should include questions about the possible use of technology, such as the meta glasses. Questions like “What technology might you use that could potentially lead to privacy violations?”  And “is the use of [the meta glasses or other] covered by HIPAA?” (We all have a right to know!)

• Review Records 

As soon as results, notes, or documents are available, they need to be carefully reviewed. This is true for any interaction with the system: prescriptions, tests, procedures, surgeries, therapies – anything. 

If they are your own records, that’s easy to do. If they are a client’s records, you can offer to review them FOR the client, or WITH the client – their choice. 

The minute any data seems off (as in “GIGO = garbage in, garbage out), or if the information isn’t received in an acceptable amount of time, or if the follow up required either isn’t specified or doesn’t seem right, contact the medical records department and ask for a review. Contact the doctor for a review as well. If possible, alert the primary who sent you to the doctor who ordered the test or other item…. In effect, you’ll need to put on notice all people, departments, medical records personnel, billing personnel, or anyone who might have been affected, or will be affected by the results you have concern about. (Example: if there is an error in the medical records, there could both an error in treatment recommendations, and a resulting error in the billing, too. There can be a huge ripple effect!)

If the review does turn up an error, then the real work begins and corrections will need to be made, both medically, and with the records in question. You’ll need to be the squeaky wheel to get the information corrected, explaining to your client that even though the immediate problem has been solved, that if it doesn’t get corrected in the records, then the ripple effect could create even bigger problems down the road.

• Provide Some Simple AI Education

Seems like an oxymoron, right?  “Simple AI Education” – really?

Yes – in this case, really. Here’s my advice for using AI for health advocacy-related inquiries:

1.  Any information shared with AI must be as private as possible. No matter what AI tells you, do not provide any personal information such as your surname, address, phone number, social security or Medicare number – none of it. I repeat: do NOT provide personal information! BTW – AI engines might ask you to scan in any test results you’ve received. If that seems necessary, be sure all personal information has been redacted, including the names, addresses, etc of the provider, AND any insurance information. A thick, black, sharpie will do the trick.

2. Ask whatever questions you want to, but expect them only to yield IDEAS. Do not expect your questions to yield ANSWERS. There is NO AI tool that can accurately diagnose or prescribe for YOU. But they certainly can outline considerations and possibilities. Because…

3.  The very best outcome you can achieve from asking AI is to formulate questions for whatever providers you are working with: prescribers, treaters, surgeons, even psychologists and psychiatrists – and billers. In my personal experience, there are always ideas I never would have known to consider. I write up a few questions for my doctor and tell them that my questions came as a result of an AI review. Every time, those doctors have appreciated my approach of researching – then confirming with them. This is how a good partnership works.

That said – the other side of this approach will help determine whether another opinion (second? third?) is required. At least a second opinion FROM A QUALIFIED HUMAN MEDICAL PROVIDER is always required when any sort of invasive or long-term treatment is prescribed or recommended. AI might also produce considerations your current choice of doctors isn’t willing to discuss. Such a case is a red light flashing “Get another opinion!”

I’ve been a techno-phile for a very long time. My extensive experience has led me to believe that tech is something most of us think we prefer not to live with – and prefer not to live without, too. 

Do yourself a favor by learning to manage technology, rather than letting technology manage you.