fda-catalyst-devices

Methodology inspired by publicly taught medtech catalyst frameworks; all text is an original paraphrase.

Defaults

Parameter	Default Value	Rationale
forward_window_quarters	4	Standard catalyst window; extend to 8 on request
include_superseded	false	Active-only events by default; supersede chains on request
coverage_lens	true	Coverage and NTAP milestones tracked alongside decision events
recall_scan	true	Recalls and watch actions surfaced as risk signals by default
holiday_labeling	true	Holiday-shifted weeks flagged on every calendar row

Preflight

Run canonical pre-flight per contracts/preflight.md. Propagate X-Agentii-Trace per contracts/x-agentii-trace-header.md. Confirm ticker resolution via search_companies before catalyst queries.

Triggers

• "What device decisions (510(k)/PMA/De Novo) are coming up for [ticker]?"
• "When is the PMA decision date for [device/ticker]?"
• "Which medtech names have device decisions this quarter?"
• "What coverage or NTAP milestones are pending for [ticker]?"
• "Has [ticker] had any recalls or watch-list actions recently?"
• "Build a device catalyst calendar for my medtech watchlist."
• "How do De Novo classifications differ from 510(k) as catalysts?"
• "Evaluate the approval odds for [device] ahead of its panel."
• "What post-market surveillance obligations does [ticker] carry?"
• "Which device decisions overlap with [ticker]'s earnings date?"
• "How did similar devices trade around their decision dates?"
• "What is the installed-base story behind [ticker]'s next label expansion?"

Production Grounding

• Device decision tracks: 510(k) clearance (substantial equivalence), PMA approval (higher evidence bar), De Novo classification (novel low-to-moderate-risk path) — each carries a binary-event dynamic.
• Reimbursement is the second gate: CMS coverage determinations and NTAP milestones convert an approval into revenue; an approval without a coverage path is a hollow catalyst.
• Recall and watch actions invert the signal: recall classes, watch lists, and post-market surveillance findings are negative catalysts that can compound.
• Device vocabulary: installed base, placements, procedures, utilization, ASP with label-expansion annotations — the commercial context for sizing the post-decision move.
• Never fabricate decision dates, recall classes, or coverage outcomes; annotate coverage gaps (per references/knowledge-frameworks.md).

Data Source Priority

1. Structured decision data: get_device_decision (device_decision_event); asset identity via search_universe_devices / get_company_devices.
2. Filing context: search_documents / search_sec_filings for PMA submissions, recall disclosures, MDRs.
3. Coverage and NTAP milestones: search_documents keyword coverage/NTAP + read_source_outline/read_source_pages.
4. Historical grounding: search_investment_cases (device decision/recall events), search_by_analogue.

Methodology

Retrieval Scope

unstructured_document_search

Retrieval Strategy

1. Resolve ticker(s) via search_companies; collect company_id for device queries.
2. Pull active decision events (get_device_decision) and asset inventory (get_company_devices).
3. Overlay coverage/NTAP milestones and recall/watch actions from filings.
4. For panel meetings, retrieve briefing documents via read_source_outline/read_source_pages; apply scrutiny axes.
5. Ground the play: search_investment_cases / search_by_analogue with device filters; cite /v/ records.

Temporal Scope

See frontmatter temporal_scope block. Post-market surveillance windows may require looking back up to 8 quarters.

Tool Allowlist

See frontmatter allowed_tools.

Protocol

1. Ticker resolution
2. Decision enumeration (510(k) / PMA / De Novo)
3. Coverage and NTAP milestone overlay
4. Recall and watch risk scan
5. Analogue grounding + calendar synthesis

Modes

• Watchlist (default, multi-ticker): one dated calendar with per-ticker sections, sorted by date.
• Single ticker deep-dive: full decision timeline, coverage state, and recall history.
• Event type: restrict to PMA / De Novo / 510(k) / coverage / recalls on request.
• Risk scan: recall and watch actions as a forward-looking risk screen.

Tool Fallbacks

Failure	Fallback
get_device_decision empty	search_documents keyword decision/PMA + search_sec_filings; annotate coverage_gap
Coverage milestone not found	Flag reimbursement status unknown; never assume a coverage path
Recall data missing	State surveillance data unavailable; do not infer recall classes
Knowledge tools empty	Proceed with structured data only; annotate knowledge coverage_gap

Output File

{ticker}/{YYYY-MM-DD_HHMM}_fda-catalyst-devices_{affix}.md (watchlist: watchlist/{YYYY-MM-DD_HHMM}_fda-catalyst-devices-calendar_{affix}.md)

Output Structure

1. Executive Summary — highest-impact device-track catalysts in the window, 2-3 sentences
2. Catalyst Calendar — dated table: date, ticker, event type, device, application, holiday-week flag, expected-impact framing
3. Per-Event Analysis — decision-track assessment (evidence bar, scrutiny axes) + decision history
4. Coverage & Reimbursement State — coverage determinations, NTAP milestones, ASP context
5. Recall & Surveillance Scan — recall classes, watch actions, post-market obligations
6. Risk Assessment & Coverage Gaps — binary-risk sizing, signal clusters, missing data

Error Handling

Error	Fallback
No device decision rows	Fall back to earnings + pipeline events; annotate coverage_gap — never fabricate dates
510(k) vs PMA ambiguity	State both possibilities with the evidence-bar difference; do not guess the track
Ticker not in med universe	Resolve via CIK/device-universe match; surface the classification gap

Memory Load

See contracts/memory-load.md.

Snapshot

See contracts/snapshot-synthesis.md.

Final Summary (TUI)

Include ### Key Citations block with 0-10 clickable /v/ URLs.

References

• contracts/citation-and-memory.md
• contracts/output-frontmatter-schema.md
• contracts/memory-load.md
• contracts/snapshot-synthesis.md
• contracts/preflight.md
• references/knowledge-frameworks.md