Why memory is not sufficient
Recall of symptoms is systematically distorted, and the distortions are well characterised. Recent experience is weighted far more heavily than earlier experience. The most intense episode dominates the memory of an entire period. And once you form a belief about whether something is working, recall reorganises around that belief — you remember the instances that fit and lose the ones that do not.
The result is that a symptom history reconstructed at a clinical appointment and one recorded contemporaneously routinely disagree, and the contemporaneous one is right. This matters most in exactly the situation where it is hardest to fix: a consultation where you are trying to describe eight weeks in five minutes.
Making a subjective record consistent
The requirement is not precision — it is comparability with itself. A scale that means the same thing to you in March as it did in January is more valuable than a finer scale you apply differently over time.
- Use a fixed scale and define the ends once. A 0–10 with an explicit anchor for what a 7 means beats an undefined one you drift on.
- Record at a consistent time of day. Symptoms fluctuate diurnally, so a log sampled at random times mixes that variation into everything else.
- Record absence as well as presence. A log containing only bad days looks like an unbroken run of bad days.
- Note the interval since your last dose. This is what makes timing patterns visible at all.
- Keep the symptom list stable. Adding a new one mid-way means you cannot tell whether it started then or you only started looking then.
Timing relative to dose is the useful signal
A symptom log without dose timing tells you what happened. With dose timing it can tell you whether the two are related, which is the actual question.
Patterns that emerge are usually simple once visible: something clustering in the first day or two after each administration points somewhere quite different from something that appeared once and persisted, or something that tracks the trough rather than the peak. None of these are visible in a list of dated symptoms without the doses on the same timeline.
The caution is the obvious one: co-occurrence in time is not causation, and there is no control group in a personal log. What the record does is generate a specific, dated, checkable observation to bring to someone qualified — which is considerably more than a general impression that you have been feeling off.
What to escalate rather than log
Tracking is for patterns over time. Some things are not pattern material, and treating them as a data point to observe is the wrong response.
Severe or rapidly worsening symptoms, difficulty breathing, swelling of the face or throat, chest pain, signs of a systemic allergic reaction, or an injection site that becomes hot, spreading and painful over days all warrant medical attention rather than a log entry. The record is a tool for the ordinary case; it is not a substitute for judgement about the unusual one.
Key takeaways
- Recalled symptom history is systematically distorted; a contemporaneous log and a recalled one disagree, and the log is right.
- Consistency with yourself matters more than precision — same scale, same time of day, stable symptom list.
- Record absence too, or your log will read as an unbroken run of bad days.
- Timing relative to dose is what turns a symptom list into something a clinician can act on.
Stop keeping this in your head.
Pep AI keeps your compounds, vials, schedule, injection sites and history in one place — and does the reconstitution math for you. Free on iOS and Android.
Frequently asked questions
Keep reading
This guide is general information for people already organising their own protocol. It is not medical advice, it does not recommend any compound or dose, and Pep AI is not a medical device. Talk to a qualified healthcare professional about anything you inject.
Published by the Pep AI team · Updated August 2026