# Outcome codebook: which adverse-event terms count toward each outcome

Trials on ClinicalTrials.gov report adverse events as terms, almost all MedDRA preferred terms, each under an organ system (MedDRA's system organ class), in two tables: serious events (every one) and other, non-serious events (those above a frequency threshold). The analysis counts, for each outcome below, the participants in each arm with an event whose term belongs to that outcome. Your task is to say which outcomes each term belongs to. You see each term (lower-cased, spaces collapsed), the organ systems it was reported under, and which tables it appeared in, with no counts.

## How to code a term

- Give each term **every** outcome whose definition it meets. Almost all terms meet none.
- Code by what the term means clinically, the way a pharmacovigilance physician building a term list would. A term meets an outcome when an event reported under it would, in the large majority of cases, be an instance of that outcome. Don't code a term to an outcome it only sometimes accompanies: "depression" or "major depression" is not suicidal ideation, and "fall" or "laceration" is not self-harm.
- The definitions name examples, not complete lists: include other terms of the same meaning, including spelling variants, older MedDRA terms, and terms qualified by method, severity, or setting.
- For each term you code to at least one outcome, give `canonical`: the term as a MedDRA preferred term, in MedDRA's spelling (British) and capitalization (first letter capital), merging terms that are the same preferred term spelled or cased differently. For terms you code to no outcome, give the term itself with its first letter capitalized.

## Outcomes

| ID | Outcome | Includes | Excludes |
|---|---|---|---|
| S1 | Suicidal behaviour | completed suicide, suicide, suicide attempt, attempted suicide, suicidal behaviour, preparatory acts toward suicide, intentional self-injury, intentional self-harm, deliberate self-harm, self-injurious behaviour (unless stated to be without suicidal intent), intentional overdose, intentional drug misuse to harm oneself, self-poisoning, poisoning deliberate | accidental overdose, accidental injury, self-injury explicitly stated to be non-suicidal, suicidal ideation without an act (S2) |
| S2 | Suicidal ideation | suicidal ideation, suicidal thoughts, depression suicidal, suicidal plan without an act, suicide threat, passive death wish, thoughts of self-harm | depressive symptoms without suicidal content, completed or attempted suicide (S1) |
| S4 | Completed suicide | completed suicide, suicide, death by suicide, suicide (fatal) | suicide attempt, intentional overdose not stated as fatal |
| S5 | Overdose of unstated intent | overdose, drug overdose, multiple drug overdose, poisoning, toxicity to various agents, drug toxicity reported as an overdose, intentional or accidental not stated | intentional overdose (S1), accidental overdose, medication errors, overdose of the study drug stated as a dosing error |
| C1 | Nausea | nausea, nausea and vomiting, morning-sickness-like nausea outside pregnancy, retching with nausea | vomiting alone, dyspepsia, motion sickness, nausea of pregnancy |

A term can meet both S1 and S4 ("completed suicide"), or S1 and S5 only if it states intent ("intentional overdose" is S1, not S5).

## Output

A JSON array, one object per input term, in the order given:

```json
{"term": "intentional overdose", "canonical": "Intentional overdose", "outcomes": ["S1"]}
```

`outcomes` is an empty array when the term meets none.