# Frozen Ukrainian PHQ-9 and GAD-7 package

Status: research source package only. It does not change application code.

## Selected editions

| Instrument | Frozen primary edition | Locator | Implementation gate |
|---|---|---|---|
| PHQ-9 | [MOH/DEC Unified Depression Protocol, Order No. 1003, 25 December 2014](https://www.dec.gov.ua?ZG93bmxvYWQ=d3AtY29udGVudC91cGxvYWRzLzIwMTkvMTEvMjAxNF8xMDAzX3lrcG1kX2RlcHJlc2l5YS5wZGY%3D) ([Rada mirror](https://zakon.rada.gov.ua/go/v1003282-14)) | Section VIII, Appendix 1, appendix sheets 1–2 | Approved only with non-diagnostic display labels, the item-9 safety contract, and screening-only language. |
| GAD-7 | [MOH Appendix 5, Order No. 288, 20 February 2024](https://moz.gov.ua/uploads/10/54593-dn_288_20022024_poriadok_dod_5.pdf) ([Rada registry](https://zakon.rada.gov.ua/go/z0332-24)) | Complete standalone Appendix 5 | **NO-GO for app implementation.** Frozen as evidence only until the exact validated 2024 Ukrainian item corpus and provenance are obtained. |

The exact items, response values, scoring bands, source metadata, validation evidence, and fixtures are in [`phq9-gad7-uk.json`](./phq9-gad7-uk.json). File and canonical-payload hashes are in [`phq9-gad7-uk.sha256`](./phq9-gad7-uk.sha256).

## Implementation rules

- Use the stored item and response text as one indivisible edition. Do not substitute wording from UIKPT, PHQ Screeners, another Ukrainian adaptation, or the English original while retaining these source citations.
- Accept only integer answers 0–3. Require all scored items; never prorate or impute a missing response.
- PHQ-9 total is the sum of items 1–9, range 0–27. Its functional-impairment response is never added to the total.
- Preserve the MOH diagnostic-style PHQ band text only as source metadata. User-facing results must use `displayLabelUk`: “Мінімальні або відсутні симптоми депресії,” “Легкі симптоми депресії,” “Помірні симптоми депресії,” “Помірно виражені симптоми депресії,” and “Виражені симптоми депресії.”
- Show the PHQ functional question when any scored item is above zero. It is required for questionnaire completion when shown, although the numeric score itself can be computed before it is answered.
- GAD-7 total is the sum of items 1–7, range 0–21. The selected MOH appendix has no functional-impairment follow-up, so none is imported.
- Severity bands are screening descriptions, not diagnoses or Ukrainian population norms. No result may provide treatment advice.

## Safety contract

Any PHQ-9 item 9 response above zero independently and immediately opens a prominent help route—even when the remaining questionnaire is incomplete or the total is low. The route must contain release-time-verified Ukrainian emergency/crisis resources, an option to contact a trusted person, and prompt professional-assessment guidance. PHQ-9 alone must not assign a suicide-risk category or infer imminent intent.

Official resource references shown by the app are the [MVS 112 information page](https://mvs.gov.ua/news/sluzba-112-mvs-ekstrenu-dopomogu-mozna-viklikati-navit-bez-mobilnogo-zviazku), the [MOH suicide-help guidance](https://moz.gov.ua/uk/yak-dopomogti-lyudini-yaka-maye-suyicidalni-dumki) (serves HTTP 403 to non-browser clients as of 2026-07-11; real-browser rendering pending verification), and the dynamic [“Ти як?” hotline directory](https://howareu.com/hot-lines). On 2026-07-11, the [Lifeline Ukraine](https://lifelineukraine.com/en) home page stated that its hotline was paused and its short number and chats were temporarily not accepting contacts. The 7333 contact is therefore deliberately excluded from the actionable resource list, despite stale 24/7 wording remaining elsewhere on that page. Availability and contact details must be verified at release/display time.

Every result must state:

> Цей опитувальник є скринінговим інструментом і не встановлює діагноз. Результат не є рекомендацією щодо лікування або зміни лікування.

## Evidence limits and edition differences

PHQ-9 has Ukrainian reliability/construct evidence, including a [2025 study](https://e-medjournal.com/index.php/psp/article/view/597) and a [Ukrainian civilian wartime study](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1553555/full). Neither provides gold-standard Ukrainian diagnostic-cutoff calibration for this exact MOH wording. The latter cites the UIKPT 2012 form, whose item/anchor wording differs from the frozen MOH edition.

GAD-7 has a substantial [2024 Ukrainian adaptation study](https://insight.journal.kspu.edu/index.php/insight/en/article/view/180) and a [2026 student study](https://mpp-journal.com/index.php/journal/article/view/101). The 2024 study warns that wartime conditions can produce high scores without establishing GAD; it did not establish item-level identity with this MOH appendix. The MOH form itself has distinctive wording: item 1 contains “дратівливість,” item 5 uses “Місця собі не знаходжу,” and item 6 again contains “неспокою чи дратівливості.” A final regular-search attempt could not execute because the configured Tavily plan quota was exhausted, and earlier searches did not expose the validated seven-item corpus or supplement. Therefore the MOH form is evidence only and GAD-7 app implementation is **NO-GO pending exact validated wording provenance**.

The PHQ source has a missing space in the functional question (`обов’язків,домашніх`). JSON preserves that source text and separately provides a display version with only the space restored.

Remote PDF-byte hashes are intentionally absent: Tavily returned extracted text/search snippets rather than the source bytes. The manifest therefore hashes the frozen canonical payload and local research artifacts without pretending to authenticate the remote PDF bytes.

## Validation

Run:

```powershell
node research/validate-phq9-gad7-uk.mjs
```

The validator checks structure, item counts, response domains, contiguous bands, all min/max/boundary fixtures, missing-response behavior, the PHQ item-9 independent trigger, functional-question completion rules, canonical payload hashes, and the artifact manifest.