Health Install #055: It Answered You. Did It Check?

You ask a health question online. The answer sounds clear. It even sounds kind.

For a moment, you feel better. Someone has explained it.

Then comes the question worth asking: what did that answer actually check?

A clear answer can still be missing facts about you.

Today, turn one answer into a short note you can use with a clinician or pharmacist. Keep the useful questions. Check the claims before you act.

Time: 15 minutes. Cost: Use what you own. One online claim has three written questions, a source status, and a person who can help check it.

Print this issue for your household binder. A saved phone note works, too.

Notice: Use online answers to prepare for a conversation. Before changing a medicine or delaying care, check with the clinician or pharmacist who can review the details. For urgent symptoms, seek prompt care instead of doing this exercise first.

Build: The Three-Question Check.

Keep: One online claim has three written questions, a source status, and a person who can help check it.

On September 10, the National Institutes of Health posted a research topic about chatbots and their use. It calls for study of both benefits and harms.

Among the questions: how do people use these tools to read symptoms, make health choices, or decide whether to seek care?

This is a call for research, not a finding that every chatbot helps or harms people.

The household issue is the step after the answer. If a wrong answer makes someone put off care or change a medicine, the words can lead to a real problem.

There is a better way to use a helpful explanation: bring it to a person who can check your case. A chatbot can help you form a question. It has not examined you.

In 1966, Joseph Weizenbaum described a computer program at MIT called ELIZA.

A person typed a message. The machine typed back. That alone could feel startling in a room full of early computer equipment.

One well-known script made the exchange sound like a conversation with a therapist. Yet the program was following rules for words.

It looked for certain words in the message. Then it used a rule to turn parts of that message into a reply. When it found no useful match, it could give a more general response.

Imagine telling a machine about a worry and seeing part of your own sentence come back as a question. It might feel as if someone had listened closely.

But a personal-sounding reply did not mean the machine knew your life.

The Computer History Museum describes ELIZA as an early program that could imitate this kind of conversation. Its place in history comes partly from how much people could read into a simple exchange.

Today's chatbots are far more advanced. They are not just ELIZA with a new screen. The old program still gives us a useful question to carry forward.

Which part of my trust comes from the facts? Which part comes from how the answer sounds?

You can ask that without being rude, fearful, or against new tools.

A calm voice can make hard information easier to use. That is a real benefit. It just does not tell you whether an answer fits your medicines, your health history, or what is happening now.

Before a clear answer becomes a health decision, it needs a check that goes beyond clear words.

The sound of being understood is not the same as having your case checked.

An old Egyptian medical text gives us a very different starting point.

It is known today as the Edwin Smith Papyrus. The name belongs to the man who bought it in the 1800s, not the person who first wrote the medical lessons.

The surviving roll dates to around 1600 BCE. Much of it deals with wounds and injuries.

What makes it useful for this story is the order of the cases. The text describes an injury, an exam, a judgment about the problem, and what might be done next.

The patient came before the verdict.

Imagine a reader following one of those cases. A page gives guidance, but the next step is to look at the person. What signs are present? What can the person do? What does the exam show?

That is different from letting a neat sentence stand in for the missing check.

The text did not treat every problem as easy to fix. Some cases received a hopeful judgment. Others were described as difficult or beyond what the writer could treat.

The old methods belong to their time. What we can carry forward is the value of asking for evidence before choosing the next step.

Your health question may involve facts that never made it into the search box. A medicine name. A recent change. A symptom that began at a certain time.

A good note can bring those facts into the conversation. It helps a clinician or pharmacist see what you are asking and what you have already read.

That does not require a long report. Three clear questions can be enough to start.

First, make the missing facts visible. Then ask what the answer means for you.

An answer becomes more useful when someone checks what it is based on and whether it fits the person asking.

Use online answers to prepare for a conversation. Before changing a medicine or delaying care, check with the clinician or pharmacist who can review the details. For urgent symptoms, seek prompt care instead of doing this exercise first.

Done: One online claim has three written questions, a source status, and a person who can help check it.

A short note for your next health conversation — Paper or your phone’s Notes app will do. Bring the original wording so the person helping you can see exactly what you read.

Keep the clear explanation. Check the facts before you turn it into a health choice.

— James Williamson
Freedom Health Daily

Know what you are being told. Know what to ask next.

P.S. Have you ever read a health answer that sounded clear but left you with more questions? Hit reply and tell me. Forward this to someone who likes to look things up before an appointment.

P.P.S. Keep building: