Chronic condition logs: recording without interpreting

A log for an ongoing condition exists to do one thing: preserve dated observations accurately, so that at the next consultation there is a sequence to look at rather than an impression to describe.

That is genuinely valuable, and it is worth being precise about why. A consultation sees one moment. A condition managed over months is a series of moments, and the series contains information that no single visit can contain — whether something is steady or moving, whether a change coincided with something else, what the range looked like rather than what today's figure is. Only a contemporaneous record can supply that.

The same document has an obvious second use, and this is the one to refuse. Once a clean, dated column of numbers exists, it invites a conclusion from whoever is keeping it. It becomes very easy to add a note that things are getting worse, to mark an entry as high, or to write down that something did not seem to be working.

Every one of those is a clinical judgement, made by someone not in a position to make it, and inserted into a document that will be read as a record. In India, offering a clinical opinion on a health finding is restricted to practitioners enrolled under the National Medical Commission Act, 2019.

So the discipline is: record, do not assess. Everything below follows from that.

The most common alternative to a log is answering from memory at the appointment, and it fails in a specific way worth naming.

Asked whether medication has been taken as prescribed, most people answer with a summary: mostly, more or less, I think I missed a few. That is not a small imprecision. It compresses a pattern into an impression, and it discards exactly the information that a pattern contains — how many, which days, whether the gaps cluster, whether they coincided with anything.

The same compression happens with readings. A recollection reports a general sense of where things have been, which is drawn disproportionately from the most recent and most alarming values, because that is how recall works. A log reports what the values actually were, in order.

There is also a directional problem. Remembering is not neutral: people under-report gaps they feel awkward about and over-report ones they feel were justified. A record made on the day carries none of that, because it was written before it became an answer to a question.

So the argument for keeping a log is not diligence. It is that a dated record and a recollection are different kinds of evidence, and only one of them can show a sequence. What the log gives a clinician is the raw material; what a summary gives them is somebody's already-processed version of it.

This is the question most templates get backwards. A log's columns should be determined by what the treating clinician has asked to be recorded, not by what a general format offers.

That matters because the choice of what to measure, how often, and under what conditions is itself a clinical decision. A log that adds columns because they are available — recording additional measurements nobody asked for, at times nobody specified — is not more thorough. It is a household deciding what is worth measuring, which is the same category of judgement as deciding what a value means.

So the practical sequence is to ask, at the appointment, what should be recorded and how often, and to write that instruction at the top of the log along with the date it was given. When the instruction changes, the new one is noted with its own date rather than replacing the old one, so the log itself shows what it was recording during each period.

Beyond the clinical columns, a few structural ones always help. The date, and the time where the instruction specifies timing. A column for the value exactly as displayed by whatever produced it. A column for context if the clinician asked for it — before or after something, at a particular time of day. And a notes column, used as described below.

A plain paper table or a spreadsheet is entirely sufficient. Clinics frequently supply their own format, and using theirs is better than a generic one.

Each entry should be a transcription, and there are four specific ways people unintentionally process values instead.

Rounding. A figure written as approximately something has lost precision that was present in the original, and the log is the only place that precision existed.

Converting. Changing units, or restating a figure in a form that feels more familiar, introduces a conversion the reader cannot see or check.

Selecting. Taking a reading twice and recording the one that looks more representative is a judgement about which measurement is valid. If an instruction covers repeat measurements, both go in, each with its time.

Characterising. Writing that a value was high, normal, better or improved is an assessment rather than a value, and it is the most consequential of the four because it is the one a reader will take as information.

The medication column has its own version of this. What it records is a fact with a date: taken as prescribed, or not taken, on that day. Not an explanation, not a justification, and not an adjustment. Where a dose was changed or stopped, the log records that it was and the date, and who instructed it if anybody did — which is a factual entry, distinct from a note about whether the change seemed to help.

The test for any entry is simple: could a different person, reading only this row, know exactly what was observed and when?

A log with missing entries is normal. A log with invented entries is worse than no log, and the difference between the two is a decision made when catching up.

The temptation is understandable. A week has been missed, the values were probably much like the surrounding ones, and a complete-looking log seems more useful than one with holes. But a filled-in estimate is indistinguishable in appearance from a measured value, and it will be read as one. A single back-filled week can make a sequence look steadier or more variable than it was, and nobody reading the log can tell which entries were measured.

So a gap is left as a gap, and it is marked as one. A blank row with the date and a note that no measurement was taken is a truthful entry and carries real information — including, sometimes, the information that a period was difficult.

The same applies to a value recorded later from memory. If it has to be entered after the fact, it is marked as entered from recollection with the date it was written, so its status is visible.

And the notes column follows the same rule. Events and symptoms are recorded as plain factual description — what happened, when, how long — without diagnostic vocabulary and without an attempt to attribute a cause. Writing what occurred is a record. Writing why it occurred is a conclusion, and it is one the clinician is there to reach.

The log's whole purpose is realised in a few minutes at the next consultation, and how it is presented affects whether it gets used.

Bring the log itself, not a summary of it. A summary is a processed version, and processing is what the log exists to avoid — handing over a page reading that things have generally been stable substitutes an impression for the sequence.

Bring it in a form that can be read quickly: the instruction and its date at the top, entries in date order, gaps visible as gaps, and any recollection-based entries marked. A log that has to be deciphered will not be looked at in a short appointment.

Bring the impression separately, and as a question. Anyone maintaining a log for months will form a view, and that view is worth raising — written on the question list as a question about what the sequence shows, rather than written into the log as a finding. That way the observation reaches the person who can assess it without contaminating the record it came from.

And write down the answer, along with any change to what should be recorded, dated, at the top of the log.

The division of labour is the point. The household's job is to produce an accurate, dated, complete, unprocessed record and to ask specific questions about it. Reading the sequence and deciding what it means is the clinician's job, and a log stays valuable precisely by not attempting it.