Health hypotheses: what they are, how to build them and what to do with them

A health hypothesis is a possible explanation for something you notice in your body, framed so that data can confirm or rule it out. “I get very tired, maybe I sleep less than I think” is a hypothesis. “I have something serious” is not, because it does not say what to measure or what result would confirm it. This text explains how medicine reasons about a symptom, why a hypothesis is not a diagnosis, which reasoning errors we all make, how to build a hypothesis from your measurements and your tests, and how to bring it to your doctor so they can review it.^[1]

Before going on, what cannot wait. If you have chest pain or discomfort lasting two minutes or more, trouble breathing, sudden numbness or weakness of the face, arm or leg, especially on one side, trouble speaking or understanding, confusion or a sudden, severe headache with no known cause, this is not the time for hypotheses or logs: call emergency services now (911 in Mexico and the US).^[8][9]

What a health hypothesis is and how it differs from a diagnosis

In medicine, the diagnosis is the explanation of a person’s health problem. The US National Academies of Sciences, Engineering, and Medicine define diagnostic error as the failure to establish an accurate and timely explanation of the patient’s health problem, or to communicate that explanation to the patient.^[1] That definition makes clear what a diagnosis requires: accuracy, timeliness and someone with the training and responsibility to establish it, which is your doctor.

A hypothesis sits one step earlier. It is one of several possible explanations, not yet confirmed, that helps decide what information to look for. The same report describes the diagnostic process as a complex and collaborative activity that unfolds over time, and as an iterative process: as information is gathered, the goal is to reduce uncertainty and narrow down the possibilities.^[1] The hypothesis is the tool of that process; the diagnosis, when it comes, is its conclusion.

How a doctor reasons about a symptom

Faced with a symptom, a doctor does not look for a single answer from the start. They use mental shortcuts learned through experience to reach a provisional explanation, and they test it with what they ask, what they examine and the tests they order.^[2] The order looks like this: a symptom, several possible explanations, data that support some and weaken others, and a decision that can change if new information appears.^[1]

That is why the National Academies report insists that patients and families are partners in the diagnostic process: they contribute valuable input that makes it easier, and reviewing your own health records for accuracy is one way to take part.^[1] Arriving with a well-framed hypothesis and an orderly log is a concrete way to do it, without stepping into what belongs to your doctor.

That teamwork matters because diagnostic error is not rare. The same committee concluded that most people will experience at least one diagnostic error in their lifetime,^[1] and among hospitalized patients, studies found that diagnostic error accounted for 17 % of preventable errors.^[2] It is not a reason to distrust medicine, but a reason to bring the clearest possible information to the appointment.

Why a series of measurements says more than a single number

The American Heart Association explains it with a simple image: one blood pressure measurement is like a snapshot, and a record of readings over time gives a much more complete picture.^[6] The same applies to weight, home glucose, resting heart rate or sleep. A single number can come out high because of coffee, stress or a bad night; a series shows whether that number is your usual or an exception.^[6]

The same guidance points out that a single high reading is not an immediate cause for alarm: you repeat the measurement and write down both results.^[6] That idea is at the heart of a well-built hypothesis. It is not built on the number that scared you, but on what repeats. And when you compare your data, the useful reference is not only the lab range, which describes a population, but your own history, as we explain in why the reference range is not your personal norm.

The World Health Organization includes self-monitoring devices, such as those used for diabetes and hypertension, among self-care interventions, and defines self-care as the ability of people to promote and maintain their own health, with or without the support of a health worker.^[7] But the WHO also makes clear that self-care does not replace the health system or health workers: it complements them.^[7] Measuring at home is there to give your doctor better data, not to replace their judgment.

There is also a practical reason. An appointment is short, and in that time your doctor has to understand what is happening to you, since when, and what changed. A timeline with your measurements and your tests answers in one minute what takes ten from memory. If you want a method for reviewing it, in what to review in your weekly health summary we explain how to read a week of data without getting lost in the details.

Table: from a symptom to hypotheses you can check

The diagnostic process moves forward by narrowing possibilities with new information.^[1] The table carries that logic over to six common symptoms: for each one, some explanations that can be checked with measurements or tests, and what is worth logging. It is not a complete list or a ranking by probability, and it has no cut-off values: each row summarizes a blog article where the sources and reference values are.

Notice two things about the table. First, each symptom has several hypotheses, and some repeat across rows: short sleep shows up in almost all of them. That is normal, because the body does not separate its systems like a table of contents. Second, the column of what to log almost always mixes home data with lab tests. A hypothesis is best checked when both sources tell the same story, and it stays in doubt when they contradict each other.^[1]

If a symptom is not in the table, the logic is the same: write down what you notice, think of two or three possible explanations and ask yourself what data would tell one from another. If the symptom is slow recovery after a cold or a trip, in how many days the body takes to recover we explain what heart rate and sleep show during those days.

How to build a hypothesis step by step

A useful hypothesis fits in one sentence and has three parts: what you notice, what could explain it and what data would support it or rule it out. The rest is method. This order helps you avoid jumping to conclusions, which is exactly what the sources on clinical reasoning recommend avoiding.^[3]

1. Describe what you notice with facts, not labels. “I get very sleepy an hour after eating, almost every day since August” is more useful than “I have a sugar problem”.
2. Write down at least two possible explanations. A single hypothesis quickly turns into certainty. The question “what else could this be?” is the one the sources propose to slow that jump.^[3]
3. For each one, think about what data would support it and what would contradict it. If sleepiness after eating comes from the food, it should change with the size or type of meal; if it comes from sleeping little, it should follow short nights.
4. Choose the simplest log that tells them apart. Sometimes a notebook with time and meal is enough; other times you need home glucose or sleep data from your watch.
5. Set a period. A few weeks with the same method are worth more than months of scattered entries. Measure at the same time and under the same conditions, as recommended for blood pressure.^[6]
6. Review honestly. If the data do not support your hypothesis, that is valuable information too. Bring it to your doctor anyway, along with what you did find.

A complete example: “For a month I have been waking up tired. Hypothesis A: I sleep less than I think. Hypothesis B: something in my tests changed. Log: hours of sleep from my watch for two weeks and energy from 1 to 10 each afternoon; I upload my latest blood test. Question for my doctor: which test would help tell the two apart?” That sentence gives the appointment a concrete starting point and leaves the decision where it belongs.^[1]

Reasoning errors we also make at home

Mental shortcuts are useful because they allow quick decisions, but they have well-known traps, and doctors’ cognitive biases play a role in many diagnostic errors.^[2] What matters for you is that the same traps show up when a person interprets their own symptoms. Knowing them by name helps you recognize them.

Anchoring and premature closure: sticking with the first idea

Anchoring, also called premature closure, is relying on the first impression even when information that comes later contradicts it.^[2] A close variant is to stop looking as soon as the first plausible explanation appears.^[3] At home it looks like this: you read that tiredness can come from the thyroid, you become convinced, and when your TSH comes back normal you keep thinking it is the thyroid instead of looking at sleep or iron. The remedy the sources describe is to consider the opposite: ask yourself what would have to happen for your idea to be false.^[3]

Confirmation bias: seeing only what fits

Confirmation bias is interpreting new information so that it fits the idea you already had.^[3] If you believe your blood pressure goes up with stress, you will remember the day of the meeting when it came out at 145 and forget the three weeks of normal readings. A log with date and time, without choosing when to measure, is the best protection: it makes you look at every day, not only the ones that confirm what you thought.^[6]

Availability: the last thing you heard seems most likely

The availability bias makes recent or easy-to-recall explanations seem more likely than they are.^[3] In medicine it is described as letting experience with past cases color the current one.^[2] At home, it is the relative who had something similar or the video you watched last week. That does not make it wrong, but its place on your list of hypotheses should depend on your data, not on how vivid the memory is.

Relying too much on a single result

Among the shortcuts the sources describe is placing undue reliance on test results or on expert opinion.^[2] A result outside the range is not a verdict, and one within the range does not rule everything out. That is why it is worth reading each result alongside the previous ones and alongside what was going on that day, as we explain in how to read your lab results without panicking. Slowing down and thinking deliberately is, according to the sources, another way to protect yourself from these biases.^[3]

How to keep a diary your doctor can review

A health diary does not have to be complicated, but it does have to be orderly. The US Agency for Healthcare Research and Quality (AHRQ) has materials designed so that patients prepare for their visit and tell their health story in a clear, concise way, and it proposes giving the patient one minute to tell it without interruption.^[4] That minute goes much further if what you are going to say is already written down.

• Date and time of each measurement or symptom. Without the time, a data point loses half its value.^[6]
• The same method every time: same device, same time of day, same conditions. For blood pressure, two readings one minute apart each time.^[6]
• Context in a few words: what you ate, how you slept, coffee, alcohol, stress, exercise, illness, travel, cycle phase.
• Both results when you repeat a measurement, not just the one that looked right.^[6]
• Symptoms with intensity, for example from 1 to 10, so your doctor can see whether they improve or get worse.
• Medicines and dose changes, with the date they started.

If your device stores readings in memory, bring it to the appointment.^[6] If you use a watch, the trend over several weeks usually says more than one day’s number. And if at the end of the period the data show nothing striking, do not throw them away: they are your baseline, the reference you will compare with next time. That is, at heart, the idea behind preventive medicine: knowing your starting point before something changes.

How to raise the hypothesis with your doctor

How you say it matters. “I think I have such-and-such” closes the conversation; “I noticed this, I logged that, and I wonder whether it could be A or B” opens it. The National Academies report points out that a correct explanation is of no use if it is not communicated to the patient, and communication goes both ways.^[1] A simple structure:

1. What you notice and since when, in one or two sentences.
2. What you logged, for how long and with what method.
3. The hypotheses you thought of, presented as questions, not conclusions.
4. What really worries you, even if it feels awkward to say.
5. What you would like to know: which test could tell the explanations apart, or what is worth continuing to measure.

AHRQ publishes a starter list of questions that fits this method well: what is the test for?, when will I get the results? and are there any alternatives?^[5] And if you take medicines, asking whether a new one will interact with the ones you already take is also on that list.^[5] Writing them down beforehand and noting the answers during the visit saves you the feeling of having forgotten what mattered once you leave.

If your doctor rules out your hypothesis, ask what data rule it out and which explanation they consider more likely. It is not an argument about who is right, but the way to make the next log more useful. The diagnostic process is iterative: each round reduces uncertainty, even if not always in the direction you expected.^[1]

Warning signs

Hypotheses are for what can wait days or weeks. There are situations where there is nothing to log: timely diagnosis, in some urgent cases, is measured in minutes or hours.^[1] Call emergency services (911 in Mexico and the US) without waiting if you or someone else has:

• Sudden numbness or weakness of the face, arm or leg, especially on one side of the body, or one side of the face drooping when smiling.^[8]
• Sudden confusion, trouble speaking or trouble understanding what is said.^[8]
• Sudden trouble seeing, dizziness, loss of balance or coordination, or sudden trouble walking.^[8]
• A sudden, severe headache with no known cause.^[8]
• Chest pain or discomfort lasting two minutes or more, or trouble breathing.^[9]
• Fainting, loss of consciousness, changes in mental status or difficulty waking up.^[9]
• Bleeding that will not stop, coughing or vomiting blood, or sudden, severe pain anywhere in the body.^[9]

With signs of a stroke, every minute counts, and the instruction is to call emergency services right away, not to wait and see whether it passes.^[8] When in doubt, the general rule is to call emergency services when the situation is life-threatening or could become so on the way to the hospital.^[9] Outside these situations, if a symptom worries you a lot or gets worse quickly, book an appointment soon, even if you do not have a log yet.

What to bring to the doctor

For an appointment where you want to review a hypothesis, this is what helps most. The goal is for you to tell your story in one minute and for your doctor to have the data at hand to check it.^[4]

1. Your hypothesis in one sentence, with the two or three explanations you thought of.
2. The log: dates, times, method and the period it covers. If the device stores readings, the device itself.^[6]
3. Your latest lab tests and the earlier ones, so the trend shows and not only the latest value.^[1]
4. A list of medicines and supplements, with doses and since when.^[5]
5. Your questions in writing, including what each proposed test is for and when you will get the results.^[5]
6. What you have already ruled out and why, so what you already did is not repeated.

If it is your annual check-up and there is no specific symptom, the list looks similar: your home measurements, your previous tests and what you want to keep an eye on. In what to include in your annual preventive check-up we detail which tests are usually considered by age and sex.