Does more patient data improve care, or add to information overload in healthcare?

Does more patient data improve care?

Up to a point. Data helps when it is tied to a decision. Past that point, it adds review time, false alarms, and missed results.

The evidence cuts both ways. In a meta-analysis of studies published from 1997 to 2012, Zhi and colleagues estimated an average lab test overuse rate of 20.6% and an underuse rate of 44.8%. Underuse was the larger problem. Ordering more tests doesn't fix it if they are the wrong tests.

What does information overload in healthcare look like?

It looks like more data than a person can hold in mind. A 2024 narrative review in JMIR Medical Informatics, by Asgari and colleagues, brought together several measures:

• Each pediatric ICU patient generated "an average of 1460 new data points" in 24 hours, in one study the review cites.
• Emergency physicians used "4000 mouse clicks in a 10-hour shift," in another.
• Working memory "can keep 3 to 5 items in mind at a time."

A longevity visit is not an ICU shift. But a 100-marker panel plus a year of device data is still far more than 3 to 5 items.

Is there evidence that overload leads to missed results?

Yes. In a survey of VA primary care practitioners published in JAMA Internal Medicine in 2013, Singh and colleagues reported a median of 63 EHR notifications a day. Of those surveyed, 69.6% said they received more than they could manage, and 29.8% said they had missed abnormal results that delayed care.

A 2022 review in the Journal of Patient Safety, by Nijor and colleagues, concluded that EHR usability problems, including information overload, can "threaten patient safety."

What makes a data point a signal worth acting on?

Scriptura's working rule has four tests. A data point is a signal when it:

• Answers a question you asked before you looked.
• Is larger than the variation you would expect from the test and from the patient's normal fluctuation.
• Persists, or is confirmed on repeat.
• Fits the rest of the clinical picture and comes from a source you trust.

If a result fails the first test, write it down and move on unless it is dangerous. If it fails the second or third, repeat it before you act on it.

Why does a 100-marker panel produce false alarms?

Because reference ranges are built to flag some healthy people. Most ranges are "reported for the middle 95% of the reference population," as Testing.com explains, which "means 5% of people in that same population will have results that fall outside the limits."

Run that across many markers, and the numbers add up. Assuming each marker independently has a 5% chance of falling outside its range in a healthy person, the chance of at least one out-of-range result is:

• About 40% with 10 markers (1 minus 0.95 to the 10th power).
• About 64% with 20 markers.
• More than 99% with 100 markers.

That is simple arithmetic, not a study result. Real markers are correlated, so the exact figures differ. The direction doesn't change: a large panel on a healthy person will almost always show something "abnormal."

How much change between two results is real?

More than most clinicians assume. Two results from the same person differ for two reasons besides real change: the imprecision of the test itself, and the person's normal day-to-day biological variation. Laboratory medicine uses a "reference change value" for this. As Fraser explains in the Journal of Laboratory and Precision Medicine, it is built from within-subject biological variation and examination imprecision "to assess the statistical significance of differences in serial results from an individual."

In practice, a small rise between two draws may be noise. Ask your lab for its analytical imprecision for the tests you trend most, and treat changes smaller than the expected variation as unconfirmed.

When is a wearable trend a signal?

When it departs from the patient's own baseline and stays there. Dr. Steinhubl's example in the AMA report is a patient whose usual resting heart rate is 64 and who suddenly reads 80. That is worth flagging even though 80 would be normal in a clinic.

Quality improvement offers a useful test for "stays there." The run chart rules set out by Perla, Provost, and Murray in BMJ Quality & Safety in 2011 treat "six or more consecutive points either all above or all below the median" as a shift, and "five or more consecutive points all going up or all going down" as a trend.

Those rules were written for care processes, not individual patients. Scriptura's suggestion is to borrow them as a rough screen, not a diagnostic threshold. One bad night is noise. A week of nights above baseline is worth a question.

When does ordering one more test stop being useful?

When no plausible result would change what you do. Professional societies apply this principle directly. The American Society for Clinical Pathology's Choosing Wisely list advises against ordering broad respiratory pathogen panels "unless the result will affect patient management."

Imaging shows the cost of testing without a question. In its April 2023 statement on screening total body MRI, the American College of Radiology said: "there is no documented evidence that total body screening is cost-efficient or effective in prolonging life." It warned that such screening finds "numerous non-specific findings that will not ultimately improve patients' health" and can lead to "unnecessary follow-up testing and procedures."

Scriptura's suggestion before any test: write down the result that would change your plan. If you can't name one, don't order it yet.

How do you triage a 100-marker report and a year of sleep scores in one visit?

Start from the question, not the data. This sequence is Scriptura's suggestion:

1. Write down the patient's main question and your clinical question in one line each.
2. Pull only the markers that bear on those questions. List the rest as "reviewed, not acted on" with the date.
3. Scan the rest for dangerous values. Those need action regardless of the question. A value that is only slightly outside its range usually doesn't.
4. For each relevant value, check the lab, the unit, the reference range, and the date.
5. For wearable data, compare the most recent weeks with the patient's own baseline. Ignore single nights and proprietary scores you can't explain.
6. Decide what to repeat to confirm, and when.
7. Document what you reviewed, what you acted on, and what you deferred, with your reasoning.

The last step matters as much as the rest. A note that shows what you reviewed and why you set some of it aside is easier to defend than one that lists everything without a decision.

How do you explain to a patient why more data is not always better care?

Honestly, and in their terms. These talking points are Scriptura's suggestions:

• "Each test has a small chance of looking abnormal when nothing is wrong. With enough tests, that becomes almost certain."
• "Looking abnormal can lead to repeat tests, scans, and worry, sometimes for nothing."
• "I'd rather test for a reason. Tell me what you're trying to learn, and we'll choose the tests that answer it."
• "Your device is most useful compared with your own normal, not with someone else's."

Patients who pay out of pocket often expect more testing to mean more care. A clear explanation of false alarms usually lands better than a refusal.

Why do more dashboards and alerts make this worse?

Because people stop responding when everything is flagged. Clinicians override drug safety alerts "in 49% to 96% of cases," according to a 2006 review of 17 papers by van der Sijs and colleagues in the Journal of the American Medical Informatics Association.

The number of alerts per patient matters. In a 2017 study of primary care clinicians, Ancker and colleagues found that clinicians "were more likely to override alerts when there were multiple alerts per patient." The authors recommend reducing the number of alerts each patient triggers.

The same applies to patient dashboards. A screen with forty colored tiles asks the clinician to do the filtering. Scriptura's view is that a tool should start from the clinical question and show only the data that answers it, with the source of each value.

What is already on record, and what should you watch?

On record as of October 31, 2026:

• Reference ranges usually cover the middle 95% of a healthy reference population, so 5% of healthy results fall outside them.
• In a VA survey published in 2013, 29.8% of primary care practitioners reported missing abnormal results that delayed care.
• Drug safety alerts are overridden in 49% to 96% of cases, and more alerts per patient are linked to more overrides.
• The ACR does not recommend screening total body MRI for people without symptoms, risk factors, or relevant family history (April 2023).
• The AMA's August 2026 wearable study found a gap between the value physicians see in wearable data and its use in practice.

Watch:

• Further publications from the AMA's multi-country wearable study.
• Updates to the ACR position as screening MRI studies are published.
• New Choosing Wisely recommendations on laboratory panels.
• Biological variation estimates for the markers you trend most.

Frequently asked questions

About the author

How can Scriptura help longevity clinicians sort signal from noise?

If your patients bring more labs and device data than a visit can absorb, book a demo to see how Scriptura's clinical decision support starts from your clinical question and shows the data that answers it, with each source shown, for you to review and sign. Clinicians verify their license before platform access.

Book a Scriptura demo.

Related articles

• Most practices do not need another staff member. They need better systems.
• The visit should start before the patient walks in
• Why regulations are changing faster than clinics realize

Get FDA and state board changes by email: subscribe to Scriptura Clinical Intelligence