Case Study 1 — The Diary That Nobody Wants to Keep

A clinical case. Theo Vasquez is an illustrative composite; the pattern is not.

Setup

I want to slow down the scene from the chapter opening, because what happens in that appointment is the most consequential twenty minutes in most people's nutrition care, and it is routinely handled badly — including, for the first several years of my career, by me.

Theo had guessed 2,000–2,200 kcal/day. His diary averaged 3,160. That's a gap of about 1,000 calories, or roughly a third of his actual intake — squarely in the range that doubly-labelled-water validation studies find in the general population.

The clinical question is not "how do I tell him?" It's "what do I do so that this number becomes useful information rather than evidence for the prosecution?"

Because here is what happens if you get it wrong. The patient hears you eat more than you said, which they hear as you lied, which they hear as you are the kind of person who lies about food, which — for someone who has already tried five diets and concluded the problem is him — confirms the thing he already believed and is the reason he's been avoiding this appointment for two years.

And then he stops keeping diaries. Forever. And you have destroyed the single most useful assessment tool you had, in exchange for being right about a number.

The three days

Here is what Theo actually recorded, condensed. I'm including it because most people have never seen what a real food diary from a real person looks like, and the striking thing is how unremarkable it is.

Wednesday

Thursday — similar, 2,940.

Saturday

Three-day average: 3,160 kcal/day.

What I actually said

Not "you're eating a thousand calories more than you thought."

I said: "This is a really good diary. Most people don't record the ice cream."

Which was true, and which was the most useful thing available to say, because the accuracy of the diary was the thing at risk and the thing I needed him to keep doing.

Then: "Before we look at totals — what surprised you while you were writing it down?"

He said three things, unprompted:

1. "I didn't know I was drinking that much coffee with stuff in it."
2. "The pretzel thing. I genuinely didn't know I did that twice a day."
3. "I measured the oil because you told me to and I've been using about three times what I thought."

He found three of the seven sources himself. Before seeing a single total. That's not a coincidence — it's the mechanism by which food diaries work, and it's why the diary is worth keeping even if you never add it up.

Only then did we total it, and by that point the number was confirmation of things he'd already noticed rather than an accusation.

Analysis

Why the gap exists, and why it isn't lying

Look at where the calories hid:

Not one of these categories is food that a person would deny eating. They're categories that don't get encoded as eating events in the first place. Ask Theo at 4 p.m. what he'd eaten and he'd have listed toast, a sandwich, and told you honestly that was it.

The weekend problem

Notice Saturday: 3,435 kcal, against weekday averages around 3,000.

Now generalize. If someone maintains a modest deficit Monday to Friday and eats 800–1,000 above maintenance on both weekend days, the arithmetic frequently nets to zero across the week. This is the most common reason for "I'm doing everything right and nothing is happening," and a diary that covers only weekdays will never reveal it.

That's why the checkpoint specifies two weekdays and one weekend day. It isn't arbitrary.

What the number is and isn't

Theo's 3,160 is an estimate with an error bar of roughly ±20% — and, per §4.7, it's probably still an underestimate, because everyone's is, including his careful one.

So we did not build a plan around 3,160. We built it around the structure: seven identified sources, of which three could be changed without him cooking anything differently, negotiating with his wife, or downloading anything.

The plan that came out of it was not a diet. It was:

• Beer to two nights a week instead of five (≈−135 kcal/day averaged)
• Coffee: keep four, but two of them black (≈−90)
• Measure the oil (≈−150)
• Don't sit near the bowl (≈−200)

Roughly −575 kcal/day, from four changes, none of which involves a food group, an app, or the word "diet." Against his surplus of +490, that's the whole problem solved on paper.

On paper. §4.6 says it won't deliver the full amount, and it didn't — but it moved, and it kept moving, and eighteen months later three of the four changes had survived. (The oil measuring did not. Nobody measures oil forever.)

Discussion Questions

1. I opened with "this is a really good diary" rather than the number. Was that manipulative? What's the difference between building rapport and managing someone? Where's the line?

2. Theo identified three of seven sources himself, before any totalling. What does that suggest about the mechanism by which food diaries work — measurement, or attention? What would follow for how you'd prescribe one?

3. The four-change plan targets ~575 kcal but §4.6 predicts it will deliver less. Should Theo be told that up front? Argue both sides — consider what happens when the plateau arrives and he wasn't warned, versus what happens to his motivation if he's told at the start that it won't fully work.

4. Not one of the seven categories is food a person would deny eating. What does that imply about the common clinical assumption that patients under-report because they're embarrassed?

5. The oil-measuring didn't survive. Design a replacement intervention that achieves a similar effect without requiring a measuring spoon in perpetuity.

Your Turn

If you've done the three-day diary from the Project Checkpoint, do this analysis on your own data.

1. Before totalling anything, write down three things that surprised you while recording. This is the highest-value output and it evaporates once you see the total.
2. Now total it, and compare to what you'd have guessed before starting.
3. Sort your calories into Theo's seven categories: drinks · communal/ambient food · cooking fat · portion error · other people's leftovers · non-meal eating · weekend inflation. Which is your largest?
4. Identify four changes you could make that don't require cooking differently, buying anything, or telling anyone. Estimate the calories for each, and rate honestly — 1 to 5 — how likely you are to still be doing it in six months.
5. Multiply the two columns. The total of that product, not the total of the calorie column, is your actual expected result.

Most people find that their highest-calorie change scores lowest on likelihood, and that the plan worth doing is not the plan that looked best.