The question that follows every treatment
You start a new medication. After two weeks you ask yourself: is it working yet? Am I imagining it? Are these side effects normal? That uncertainty is part of every pharmacotherapy and it's exhausting.
Structured tracking gives you an answer your gut feeling can't. This article shows how.
What you should track
Four dimensions are enough:
- Mood as a daily scale
- PHQ-9 weekly
- Side effects as a list or free text
- Adherence as a binary field (taken / missed)
InnerPulse integrates all four. You don't need separate apps.
Realistic timelines
Antidepressants don't work immediately. If you don't know that, you'll frustrate yourself.
- SSRI/SNRI: First effects often after 2 to 4 weeks, full effect after 6 to 8 weeks
- Bupropion: Similar, often slightly faster on drive and energy
- Tricyclics: 2 to 4 weeks for mood, often faster for sleep
- Mood stabilisers: Variable, often 2 to 6 weeks
- Augmentation with atypicals: 1 to 3 weeks
These timelines are guidelines. Your personal trajectory may differ, but the frame helps you wait realistically instead of stopping too early. A meta-analysis by Posternak and Zimmerman (2005) showed that the largest improvement often actually happens in the first two weeks. So if after 4 weeks you haven't had a single better day, talk to your prescriber instead of waiting longer.
What an effect looks like
When a medication works, you typically see three things in your data:
1. Variance reduction. Daily variance decreases. Bad days become less bad.
2. Trendline rises. Your weekly average mood goes up — sometimes slowly, often only visible after 4 weeks.
3. PHQ-9 falls. A reduction of 5 points or more is clinically considered a response. A reduction below 5 is considered remission.
PHQ-9 trajectory after starting treatment (example)
What no effect looks like
Just as important: what does it look like when the medication isn't working?
- PHQ-9 unchanged or higher after 6 to 8 weeks
- Daily variance unchanged
- Trendline flat or falling
- Side effects without compensating mood improvement
In that case a switch or augmentation makes sense. The data gives you and your doctor a basis to decide instead of guess.
Tracking side effects structurally
Side effects are often transient. Nausea in week 1 to 2, fatigue, headaches, sexual dysfunction. Some fade, others stay.
Track a short list per day. Example:
- Nausea: yes/no
- Dizziness: yes/no
- Sexual side effects: yes/no
- Sleep disturbance: yes/no
After 3 to 4 weeks you see which side effects faded and which didn't. That makes the conversation with your doctor concrete instead of anecdotal.
Adherence is its own variable
If you don't track intake, you don't know whether missing effect is from the medication or from missed doses.
Track every day adherence: yes or no. 90 percent adherence is good. Below 80 percent and you can't make any statement about effectiveness.
Many apps remind you. Use that reminder instead of relying on memory.
How to prepare for the doctor's appointment
You go to your next appointment. Instead of telling, you show data. Three things are ideal:
- PHQ-9 trajectory since treatment start as a curve
- Side effect frequency as a list
- Adherence rate as a percentage
That saves you 10 minutes of telling and gives your doctor an objective impression. Many patients are surprised how much more precise the conversation becomes after the first appointment with data.
InnerPulse exports all three as CSV or PDF.
What you should not track
Pharmacological micro-management. Example: measuring your mood half an hour after intake to "see the effect kick in". That's noise and often triggers hyperfocus.
Stay at the daily perspective. Antidepressants work over hours to weeks, not minutes.
Tracking is not a diagnosis
An important reminder. Mood data doesn't replace a doctor or therapy. It makes trajectories visible, but it doesn't diagnose and it doesn't prescribe.
If your data shows something is off, the answer isn't "change the dose yourself" — it's talk to your prescriber. Data gives you a stronger voice in the conversation, no more.
Read more
- PHQ-9, GAD-7 & more explains the tests you need for effectiveness measurement.
- PHQ-9: Every question explained helps you read your score trajectories more honestly.
- Keeping a Mood Journal: The Complete Guide shows the daily foundation under the weekly test.
- Recognising patterns in your mood helps separate effects from noise.
- Antidepressant onset meta-analysis: Posternak & Zimmerman (2005)
- Smartphone-based mood tracking in bipolar disorder: Faurholt-Jepsen et al. (2015) RCT