Selecting a Meaningful Outcome

What an outcome is

An outcome is the thing you are watching to see whether a change made a difference. Picking it is the second step after defining a clear question. A poor choice can make an experiment look successful when nothing useful happened, or look like a failure when something real changed.

This is information only. It does not tell you what to measure for a health condition, and it never replaces advice from a clinician.

Direct, estimated and calculated

Before choosing, know what a number is:

Scores calculated by a company have no universal clinical meaning. See consumer data vs clinical measurements.

A number is not the same as a benefit

The FDA describes a biomarker as a measured characteristic, and states that a biomarker is not an assessment of how an individual feels, functions, or survives. [1] The same caution applies to consumer metrics. A tracker reading that changes is not proof that you feel or function better. For more, see what counts as a meaningful outcome.

Numbers can also pull attention away from what matters. Clinicians have described patients who pursue ideal tracker data and seek treatment for self-diagnosed sleep problems that the tracker seems to show. [2]

Choosing well

• Prefer what you actually care about. For a bedtime routine, that may be how rested you feel in the morning, rated in a simple way you decide in advance. For a walking habit, it may be how your energy or mood feels in the afternoon.
• Keep it simple and consistent. One short daily rating you can keep up is better than a long form you will stop filling in.
• Write down how you will rate it before you start. Changing the rating scale midway makes the comparison unreliable.
• Pick one main outcome. Watching many numbers and then highlighting the ones that moved makes chance look like a finding.
• Remember that expectation matters. In experiments on perceived sleep, people told they had slept well or poorly performed differently on cognitive tasks according to what they were told. [3] Ratings of how you feel are real, but they can be shaped by what you hope to see.

Choosing a sensor reading as the outcome

If you use a device reading, ask whether it measures the thing directly, how it compares with a reference method, and how much it varies from day to day for you. Day-to-day variation is covered in establishing a baseline. One study of daily resting heart rate found that people can have a normal level that differs from another person's normal by as much as 70 beats per minute. [4] That is why comparing yourself with a stranger's number is rarely useful.

Outcomes to avoid

• Anything that requires a risky change to move it.
• Anything that stands in for a diagnosis. If you are worried about a symptom, see a clinician; see talking with a health professional.
• An outcome you cannot stop thinking about. If tracking makes you anxious, consider pausing it.