Don't bake errors into your nursing practice

This is an answer I provided when a nursing student asked if they should pay for an AI subscription through nursing school. 

No

I know people may disagree with me and I know there might be places where AI will not do harm - but I have been working with and training AI for some time now and I find its constraints every day. I have worked in healthcare for 35 years and been a professional nurse for almost 12 years. I am front an center training AI daily to be used in a healthcare realm and working amongst medical devices. I am hopeful and in favor of the the use of AI in areas such as healthcare, but I am realistic about its limitations because I see them every day.

My rationale: I only trust AI to a point. At times, AI will confidently be wrong and you have to know enough to question that it is wrong and correct it. As a nursing student, you are too busy building a sound scientific foundation of knowledge and skills. You do not want to utilize a tool that can sound correct and be incorrect.

AI knows TECH and ENGINEERING things mostly well (those are the people who created it) but AI does not have as good a grasp on non-technical things (e.g. medical diagnosis, how to apply the nursing process, understanding the ambiguity in a clinical workflow and learning to take in all variables about a patient and truly prioritizing and making an appropriate nursing decision or medical decision) Part of my work is to help train both engineers and AI on some of these things and the AI makes the same mistakes the engineers make (again, this makes sense - AI has the sensibilities of an engineer - not a healthcare professional). The AI is making logical errors, but it is still making errors. I personally, don't want my professional nursing knowledge to be hinged upon a tool that lacks humility and asserts incorrectness with confidence. In nursing school that behavior is considered dangerous. As a student, you may not know that so in your heart you don't intend to be dangerous, but the outcome is the same - unnecessary risk. AI will get there, but right now my personal opinion is that it is not ready to help cultivate professional nurses who need to build critical thinking skills as it can introduce flaws in your thinking that you may not realize. Do not bake errors into your nursing practice.

I look back to my nursing days and I think about what I needed to know and why. In my opinion, there are better resources at this time (back then and definitely now) that are more peer-reviewed and established.

For practicing test questions - the older version of the NCLEX questions (e.g. NCLEX 10000). NCLEX style questions have not changed in years. Choosing all the correct answers or choosing the most correct answer is a skill but it merely requires practice.

For individual specialties, body systems - refer to your class curricula, text books and when you are struggling with a concept you find a Children's book that covers the concept. Rationale: My mother in law - a retired nursing professor said "Text books are written to impress people, children's books are written to inform". I struggled with the cardiac system and the abdomen. I got a 4th grade book with clear pages that layered the body and explained the basics so I could get my head around what the textbook assumed I knew and I didn't. I am not too proud to call out my own lack of knowledge. That same book is one I used to teach my patients about their own body when they had a chronic illness that they didn't understand and landed them in the hospital because they could not manage it. 4th grade materials, in these cases, saved lives and improved outcomes.

There are many ways you can get all of the knowledge you need. Use flash cards. Tech that helps with flash cards is better than AI because the flash cards aren't trying to draw conclusions, or be smart - they are helping you practice.

Advice I still hear in my head. We call it a Nursing Practice, not a Nursing Destination. We are always practicing and learning. Increase your practice and you will do better.

If It Isn’t Documented, It Didn’t Happen. (Yes, Even With AI.)

AI Is Confident. I’m Accountable

AI can help me move faster with confidence. It can also be confidently wrong.

And here’s the part people may miss: whatever AI produces becomes my responsibility the second I use it. Not the model’s. Not the tool’s. Mine.

That mindset isn’t new. In high-reliability work, you don’t get to outsource accountability.

So this is my “professional cover-yourself” version of AI governance—practical, lightweight, and designed for real work.

Garbage in, garbage out

AI does not fix messy inputs. It scales them. 

Decisions pile up. One small assumption can get lost in a long string of decisions.

Rationale: If you don’t document what you did and why, those stacked assumptions eventually implode—usually at the worst possible time.

Guardrails:

Programmatic guardrails

• global guardrails (always true)

• project guardrails (context-specific)

Rationale: This prevents “helpful shortcuts” from quietly becoming bad defaults. 

Guardrails in action: Trust, but verify

AI outputs are hypotheses until proven, because confidence is not correctness 

I challenge every notion: Prove it. What could disprove it?

I periodically test the AI against the guardrails to make sure it still conforms. If it drifts, I correct the prompt, tighten the format, or reduce the scope.

Rationale: AI is not perfect (just like humans) and small tests are investments in avoiding larger future failures

Category guardrails: 

I use AI to:

• summarize

• identify patterns (especially in large data sets)

• drafting / formatting documents 

• discuss, suggest options

I do NOT use AI to:

• make final calls

• declare root cause

• “certify” anything without evidence

Rationale: Decision support is not decision ownership. 

Standardize the communication

I don’t want paragraphs. I want a clean hand-off.

So I standardize the format:

what happened → evidence → next action

Rationale: Closed-loop communication beats interpretive storytelling every time.

Keep a document trail as you go (because you will sleep between now and remediation)

“If it isn’t documented, it didn’t happen.”

And in AI-assisted work, documentation is not the conclusion.

I document the rationale in a lightweight way—because I have slept since then is a real thing.

My preferred framework documentation trail:

• what I asked

• evidence I reviewed

• decision

• rationale for above decision

• what might change the decision

• next steps

• confidence: high / medium / low

Rationale: This makes remediation faster, hand-offs cleaner, and mistakes easier to unwind.

The cost of doing this business this way: 

At first, this might feel like it slows you down.

Truth told - briefly.

But it builds something you can trust. And once the guardrails + trail + standard hand-off are in place, you stop re-learning the same lessons and start scaling outcomes.

Rationale: Fast is fine. Defensible is better.

Spoon-feed the judge

 

"Spoon-feed the judge" The best personal advice I ever got about high-stakes work. It became part of my professional practice: don’t make the decision-maker swivel their chair, open another book, or go hunting for context. Put the proof next to the claim. Make it one-and-done.

In regulated healthcare software, the “judge” might be an auditor, security reviewer, change approver, incident responder, future-you at 2 a.m., or future-you three years later in a deposition—but the rule stays the same:

If it isn’t easy to verify, it isn’t easy to defend.

I didn’t fully appreciate how universal this principle was until years later, when my roles included conducting audits and being subject to them.

As an auditor, “spoon-feed the judge” is exactly what I want. I trust, but verify. This is transparency at its best.

As the person under audit, it makes my position easier to defend because the evidence travels with the claim.

Bonus (worth its weight in gold): over time, this approach has kept me out of many audits entirely - because the work was naturally verifiable without a live explanation. More than once I’ve been told: “No—this answered all our questions. We don’t need to call you.”

What “spoon-feeding” looks like in systems work:

1. Make it readable

   Short, structured, and complete. If verification requires a scavenger hunt, you did not spoon-feed your judge and your risk rises.

2. State assumptions upfront

   Identifiers, timing windows, source of truth, and what “acceptable” looks like.

3. Locate proof next to the claim

   Your claim was clear. Now provide the proof right next to it. If proof cannot be adjacent, keep the ordering parallel and easy to follow.

4. Make cross-system truth explicit. For transitions, prove each of the following:

   • Sent
   • Received
   • Applied

   Acknowledged is not the same as true.

5. Leave an audit trail that survives time

   Links, IDs, timestamps, and artifacts that let anyone re-verify without tribal knowledge.

This is not “more documentation.” It’s high-quality evidence—delivered where it’s needed, when it’s needed.

Where do you still see scavenger hunts in your high-stakes work? Consider how you, too, can "Spoon-feed the judge".

The most critical defects often exist between systems rather than within them

 

 

Would you like to quickly find where some of your most critical system defects? 

Take a lesson from healthcare: transitions in care are well-documented high-risk times. This is not a one-off issue but a repeatable pattern.

Software systems exhibit the same pattern during transitions:

• You might say, “System A did its job,” but “System B holds the truth.” It's essential to prove it.

Consider this real-world example:

Your doctor says they sent the prescription. The pharmacy claims they don’t have it. Ultimately, you still don’t have your medication. The gap lies in this hand-off: sent ≠ received ≠ applied.

To address this, lean into testing transitions:

• Define what “done” means on both sides of the transition.
• Trust, but verify: don’t stop at “sent.” Confirm it was received and actually applied.

Test both acceptable and unacceptable outcomes:

• Acceptable: slowdowns, retries, short delays.
• Not acceptable: missing, wrong, late, duplicates.
• If you exceed a limit, log it as beyond tolerance - don’t mask issues behind a green check.

Why is this important? Testing these hand-offs provides evidence where it matters most and can make the rest of your test suite feel surprisingly simpler. Your customers will thank you for it.

I now do accessibility testing. I tested my own blog. I am not impressed.

 

My blog has health care information but it will begin to include what I learn as an accessibility tester. I think I can set the content up with proper heading levels and in a way that is consumable if I have a tool that conforms to WCAG standards.

Unfortunately, I keyboard tested a piece I published and I had too many findings despite the fact that I used heading levels, alt text, etc.

I looked at the company's help pages to see if there was information on making it accessible, a VPAT or ACR or anything about WCAG conformance and got nothing. I am now looking to change the tool I use to publish my website, and remediate my content with a more inclusive design.

The favor I am asking for is which mediums might be able to do that, so I can get down to business and remediate my content, because I can do better.

[Congestive Heart Failure] When the ol' pump ain't what it used to be

This is a repost of my response on Reddit about how someone was going deal with a new sodium and fluid restricted diet after being hospitalized for heart failure. They loved milkshakes and had met with a dietician who really did not have a good therapeutic connection.

Trigger Warning - Below I'm going to talk about what I saw taking care of heart failure patients on a hospital cardiac unit. Someone who has had a heart failure episode may become anxious when I describe the early parts of their admission.

 

This discussion barely scratches the surface of the complexity of this condition. Most people are not admitted every week or two - but the advice I learned to reduce admissions for these patients helped to curate the discussion and education I provided to patients in the clinic and helped them have much more control over their own health and disease.

 ---------------

Day 1: I have had patients who were admitted about every 10-14 days like clock work. The ER gave them diuretics, the ER would take off 1-2 liters of fluid. By the time they get to my unit, they were still afraid from being short of breath and it being hard to breathe and this was the new and improved version. There is a look in a scared person's eye that doesn't leave you. I know that look from across the room.

Day 2: Less panic, still feeling very drained. We continue to remove fluid.

Day 3-4: Start to perk up, feeling a bit better. Better enough to be irritated with us, and complain about being hungry and don't like the food we're bringing because not enough salt/fluids. Sometimes they ask family to bring food - like fried chicken.

I cannot believe my anecdotes came about because I had so many patients who would come in and say how we didn't take care of their heart failure, but by day 3-4 they were having family bring fried chicken. #ffs #nurselife #we'veseensomeshit

Now let me get to the advice I found worked the best to stop repeat admissions:

Of all the dietary restrictions, salt+fluid can be one of the hardest.

Accept this is a difficult thing. I don't love people who treat it like it is easy like dieting and losing weight. *just* change your diet and the things you eat multiple times a day, every day for the rest of your life. Yeah bud, *just* is a 4-letter word. It is okay to have grief over this.

This person discussed a having dietitian who did not connect well with them: 

If you feel like you and the dietician aren't connecting, consider finding one who has empathy. I have seen great dieticians (they are part of your healthy team and they find ways to help you feel satisfied while having to change something you've been doing your whole life) and ones who just tell a grown adult patient what to do. I have no use for someone telling adults what to do like they are children.

Intake is all choices. You can absolutely eat all the fried chicken and drink all the ice cold milk shakes (or whatever items you love). Be prepared to spend an increasing amount of time in the ED and in the hospital as that taxes a heart that has had an injury.

Basically, the heart is a pump that is with you your whole life. That pump got hurt somehow. Maybe a heart attack or some other type of an injury but at the end of the day, the ol' pump ain't what it used to be. It can't pump like it used to. Your body has tried for a long time to help it - but it just can't do all it used to do. 

What happens when you try to use the pump the same way and it can't do it? These are the chains of events that bring some people to the hospital with a worsening of their heart failure enough to be admitted.

If you are committed to liquid, consider eating foods with less fluid - since the liquids add up you can trade off. Now if less liquid-y foods are dried (e.g. beef jerky) - mind the sodium.

I don't know how you are managing your HF, but daily weight is the gold standard.

Wake up, eliminate any stuff from your body and weigh in the "same clothes" (e.g. naked = same clothes) every day. Write the weight down. If you gain more than 3 pounds in a day or 5 pounds in a week, it is time to call the doctor as that is the best indicator of water weight.

What are the other upsides to daily weights?

Just for a bit of context - our patient often didn't get admitted with a 3 pound weight gain. They were frequently admitted with a 20-30 pound weight gain and often didn't detect it until they were short of breath (height and weight can change these numbers). Watching these numbers lets you know in short order what is going to cause more fluid retention. Daily weights give you insight about your intake from the day before and if it affected you, you can make adjustments.

This is great info for the doctor too. Your doctor can have a much more informed discussion with you over time when you can bring that one bit of data for daily weights to your PCP or Cardiologist.

I don't know your situation, but if you are a Veteran, they actually have a heart failure line where Veterans can call. The line has clinicians who can advise on medication adjustments if needed on the fly. I personally believe it should be an industry standard. The concept is simple, they can manage you at home to reduce admissions. Less cost for them and more convenience and quality of life for you.

The unappreciated choice

 

You have a decision to make. You weigh the options in hopes of making an informed decision (or informed refusal). 

• The decision can be difficult
  
• Maybe you don't like the answer or outcome
  
• Maybe you feel like you don't have enough information to render a decision yet
  

Enter analysis paralysis

Consider this: *Not* deciding is also a decision. 

Why? There will always be an outcome. The absence of a decision is often the unappreciated choice in a scenario.

Consider this when you might come across time when analysis paralysis sets in:

No decision is a decision in and of itself

Why you want to interview a nurse, whether or not you are in the healthcare sector

 

Nurses are limited in their career only by their own imaginations. 

If you are a nurse thinking about changing roles or if you are an employer looking for that next employee to round out your team, let's discuss the things a nurse brings to the table:

1. Nursing is a profession - we teach each other

2. Safety first - Our license is to demonstrate that we know how to keep folks safe - it is a privilege not a right

3. Prioritizing - we have to get ahead of the thing that may kill you first; then we'll get to the second thing that might kill you (see # 2)

4. Learning - evidence based medicine is ever changing, and we have to stay up to date, so we are expert learners (also see # 2)

5. Delegation - we are trained to lead because we know we can't do it all. We learn to leverage the power of a team and the strengths of the many (See # 3)

6. Resourceful - we don't always have everything we need on hand, yet we still manage to get things done (see # 3, # 5)

7. Teachers - we have to teach medications, treatments and healthcare to folks who may not know much about medications, treatments or healthcare (also see # 1, # 3)

8. All comers - we treat everyone because we are everyone <3

9. Compassion - We see people at their absolute worst and we learn to help them maintain their dignity and autonomy as much as possible (see # 2, # 3)

10. Practice - We know how to work toward improving things "That's why we call nursing a practice, not a destination" (See # 2)

11. Pattern recognition - Nursing and healthcare are algorithmic in nature. There is math in humans and we still treat them like humans

12. Soft skills - we are consistently the most trusted profession while doing all of the above things in any given situation (See # 1, # 2, # 3, # 4, # 5, # 6, # 7, # 8, # 9, # 10, # 11)

Universal Truths

 The Distillation Process

 

These things. They happen over and over. They prove themselves at every opportunity. They are Universal Truths.

How do you recognize a Universal Truth? Well - it isn't always easy. Think of distilling a fine spirit. You boil off the impurities and what is left may be the Universal Truth you are seeking.

If something happens over and over - it may be true, from your perspective. Consider though, you are one of over 7 billion people on the planet. What is true for you, may not be true for everyone else. Try to see someone else's perspective. Challenge your own thoughts on the thing. Observe the world around you. Does it change when you changes these variables? If it does, it may not be a Universal Truth. Does it stand the test of challenge? Then you may have a good candidate!

I encourage you when you think it is a Universal Truth - don't declare it, you still may be wrong or the thought may still need further distilling and removing of impurities. Instead, try floating the thought out there, "Wow, this may be a Universal Truth" or "Do you think this could be a Universal Truth?" Have this discussion with folks. Welcome the discussion and challenge. You aren't married to it, you are working on it. 

Some Universal Truths have taken me minutes to find and years to challenge. Others take me years to find and minutes (or seconds) to say...yup...that may be a Universal Truth. I like them because these truths are the things we share with others in life. How we can connect and know each other. How we see eye to eye. How we learn and get better. Remove the impurities, retain the true spirit.

Sit with me by the fire

Imagine you and I are camping. We are sitting around the campfire, you on one side - me on the other. We are discussing things. I am telling you what I see from my perspective. I see the fire, you and the stars. It is beautiful. You, on the other hand see the fire, me and a grizzly bear approaching. 

 Stop time

If you came over to my side of the fire, you would see my perspective and you would not appreciate the danger. 

If I came to your side I would see your perspective and I would see the danger. 

There are at least three hundred and sixty seats around this fire. We may share views, we may not. They are all a little different. 

When we interact with folks, we are at a campfire. I always strive to mentally slide around to their side and try to see where they are coming from so I can understand their perspective.

I also work to help them slide to my side and see what I see. 

This brief stopping of time and mentally putting yourself on the other side of the campfire is the easiest way I know how to explain both perspective and how to promote empathy. 

During a pandemic we are both isolated and we interact in ways that are both social and anti-social. I still remember when working with folks - we all share this campfire.