How to Spot Patterns Between Treatments and Symptoms

You open the cabinet and there are fourteen bottles staring back at you. Doxy, azithromycin, two tinctures, a binder, a methylation B-complex, magnesium, low-dose naltrexone, fish oil, vitamin D, a probiotic, an adrenal something-or-other, and two things your LLMD added last month that you can't remember the purpose of. Somewhere in there, two of them are probably doing most of the work. The rest might be neutral. One might be making you worse. You have no idea which.

That's the bottle-fatigue moment. It's also when most patients try to figure out what's helping by feel, panic when they can't, and either keep everything forever "just in case" or drop the whole stack at once and crash. Neither is a good answer.

There is a saner method. It's slower than you want, and it doesn't give you a clean dashboard with a checkmark next to the winners. But it works.

Why this is genuinely hard

A few things conspire against you when you try to figure out what's helping:

Treatments overlap. You started doxy and the herbal protocol the same week. Three weeks in you feel different. Was it the doxy? The herbs? The fact that your sister stopped staying with you and you finally slept? You can't tell, because you changed three variables at once.

Symptoms vary on their own. Lyme symptoms migrate and cycle even when nothing about your treatment changed. Some patients describe roughly four-week flare patterns that some clinicians link to the bacteria's life cycle. So a "good week" three weeks in might be the treatment, or it might be where you happen to be in your own cycle.

Memory lies. Recall bias cuts both ways. If you decide doxy is helping, you'll remember the good days on it. If you decide it's making you worse, you'll remember the bad ones. Without a written record, you're not a reliable witness about how you felt three weeks ago.

The clock is slow. A new antibiotic or herbal needs at least two weeks to show whether it's doing anything, and often four. By the time you'd see a real signal, you've already added two more things and dropped one.

Placebo and nocebo run in the background. Starting something new with hope changes how you feel. Starting it while reading horror stories on a forum also changes how you feel. Both effects are real and temporary, and both can fool you for a couple of weeks.

The formal version of "figure out what's helping a single patient" has its own name in clinical research. It's called an n-of-1 trial, and it exists because group averages don't tell you what works for one specific person.

The honest method

You can't run a blinded crossover trial in your kitchen. But you can borrow the parts that matter.

Log the treatment. Name, dose, time of day, taken or skipped. Every day, not from memory at the end of the week. The skipped days matter as much as the taken days.

Log the symptoms. Pick three to five that bother you most: fatigue, joint pain, brain fog, sleep, headache, mood. Score each one 1 to 5 at roughly the same time every day. Morning works for most people because it's before the day's noise piles up. Consistency matters more than precision. A 3 today and a 3 next Tuesday should mean roughly the same thing to you.

Change one thing at a time when you can. Nobody wants to hear this because chronic Lyme patients are usually trying to claw back their lives and waiting feels like losing. But if you start two new things on the same Monday, you've cooked the data. Stagger by even a week if you can.

Give each change two to four weeks before judging it. Anything shorter and you're reading noise. The first ten days of a new antimicrobial often look bad because of Herxheimer reactions, then settle. The first two weeks of a new supplement often look good because of placebo, then settle. Settle is the data. The first part is the weather.

Track the confounders. Sleep last night, weather, period if you menstruate, stress, big meals, alcohol, exercise. These move symptoms around so much that without them you'll mistake a bad-sleep day for a treatment-failure day and pull a medication that was actually working. (Frontiers in Immunology, 2021)

That's it. Treatments in, symptoms out, confounders in the margin, four-week minimum window, one change at a time when life lets you.

What "pattern" actually means

This is the part most patients get wrong.

A pattern is not "I felt better the day after I started X." That's a coincidence at best, placebo at worst. Day-to-day movement in chronic Lyme is mostly noise.

A pattern is also not "I feel like X is helping." Feelings matter, but they're not patterns. They're hypotheses.

A pattern looks like this: you started low-dose naltrexone on March 4th. The four weeks before, your fatigue averaged 3.6 out of 5. In weeks two through five on LDN, it averaged 2.9. You stopped on April 15th because of insurance. Three weeks later, fatigue averaged 3.5 again. You restarted May 10th. By early June it was back down to around 3.0.

That is a pattern. A trend line that moved when the treatment started, moved back when it stopped, moved again when it restarted. That kind of signal survives noise. A single good day doesn't.

You don't always get something that clean. Sometimes a treatment correlates with a half-point drop in joint pain, no clear effect on fatigue, possibly worse sleep. Still useful. It tells you and your doctor the drug is doing one thing well, nothing for another, and maybe causing a side effect to factor in.

Patterns work in reverse too. The supplement you assumed was harmless might quietly correlate with worse headaches. The antibiotic everyone said would wreck you might correlate with the steadiest two months you've had.

Why your doctor needs patterns, not vibes

Walk into the LLMD's office and say "I think the doxy is helping but I'm not sure, and the herbs might be doing something, and I'm worse on Mondays," and there's nothing your doctor can do with that. It's directional vibes.

Walk in with "fatigue averaged 3.4 in March, 2.8 in April after we added LDN, 3.6 in May after we stopped it," and your doctor has something to act on. Keep LDN. Drop the supplement that didn't move anything. Try a different binder. The conversation gets concrete.

Group studies tell doctors what worked for the average patient. Your data tells them what's working for you. (Commonwealth Fund) The LLMD can't run an n-of-1 trial on you, but they can read one if you bring it.

How LymeTrack handles it

LymeTrack is built around this problem. The 5-step daily check-in lets you log treatments and symptoms in about 90 seconds, because anything longer and you'll skip days, and skipped days are where patterns die.

Every treatment gets tagged with the time you took it. Every symptom gets a 1-to-5 severity score on the same scale, every day, so two weeks of "fatigue 3" actually means something. You can do multiple check-ins per day if your symptoms swing morning to evening, and the app keeps them separate instead of averaging them into mush.

Factor tracking covers the confounders: sleep, weather, stress, activity, diet. A 2-out-of-5 fatigue day after a 9-hour sleep is recorded with that context. So is the day fatigue jumped to 4 when a storm rolled in. Without those, treatment signals get buried under daily life.

The Compass and Insights view is where the cross-correlation lives. It looks at treatment timing against symptom severity over weeks and months and surfaces the patterns you couldn't see by hand. It can flag that your joint pain dropped two weeks after you started the herbal protocol and rose again when you paused it. It can also flag that the supplement you've been loyal to for six months has no measurable effect on anything. The second one is uncomfortable but useful.

The doctor-shareable report pulls it together for the appointment: which treatments correlate with which symptoms, which factors keep showing up alongside flares, what the trend looks like over the last few months.

The app isn't going to tell you what to take. Its job is to make patterns visible so you and your doctor can argue about them with something real on the table.

Further reading

A few sources if you want to go deeper:

• Single-patient (n-of-1) trials in clinical decision making. The methodology paper for why a single patient is a valid unit of study and how clinicians use it.
• N-of-1 trials on Wikipedia. A readable plain-language overview of single-subject design, including the AB-BA-BA crossover pattern.
• Patient-reported outcomes in chronic disease (Frontiers, 2021). On why daily symptom tracking outperforms recall, and the role of confounders in chronic illness research.

The takeaway, if you want one: you can't out-feel this problem. Fourteen bottles is too many variables for a tired brain to track in real time. Write it down, give each change a real window, watch the trend line, and let your doctor see the chart instead of the vibes.