483 Observations Search

Semantic search across FDA 483 observations using AI-powered natural language queries.

Tier Access

Required Tier: Starter+ (Starter, Pro, Enterprise)

Rate limits apply per tier — see Rate Limits for details.

POST /v1/483/observations/search

Search for FDA 483 observations using natural language. The API uses vector embeddings to find semantically similar observations, enabling you to discover relevant compliance issues based on meaning rather than just keyword matching.

Request

POST https://api.ctwise.ai/v1/483/observations/search
X-Api-Key: YOUR_API_KEY
Content-Type: application/json

Request Body

{
  "query": "sterility assurance failure in aseptic processing",
  "filters": {
    "classification": ["OAI"],
    "product_type": "drug",
    "date_range": {
      "start": "2023-01-01",
      "end": "2024-12-31"
    },
    "program_area": "Drugs",
    "cfr_reference": "211.113",
    "category": "Aseptic Processing"
  },
  "top_k": 10,
  "include_facility_context": true,
  "include_regulatory_mapping": true,
  "enable_query_expansion": true
}

Parameters

Parameter	Type	Required	Description
query	string	Yes	Natural language search query
filters	object	No	Filter criteria for search results
filters.classification	string[]	No	Filter by inspection classification: NAI, VAI, or OAI
filters.product_type	string	No	Product type: drug, biologics, medical_device, food, etc.
filters.date_range	object	No	Inspection date range
filters.date_range.start	string	No	Start date (ISO 8601 format)
filters.date_range.end	string	No	End date (ISO 8601 format)
filters.program_area	string	No	FDA program area: Drugs, Biologics, Devices, Foods, etc.
filters.cfr_reference	string	No	CFR citation (e.g., "211.113")
filters.category	string	No	Observation category (see Categories)
top_k	integer	No	Maximum results to return (default: 10, max: 100)
include_facility_context	boolean	No	Include facility details in results (default: false)
include_regulatory_mapping	boolean	No	Include CFR and category mappings (default: true)
enable_query_expansion	boolean	No	Enable 483-specific query expansion (default: true, see Enhanced Search)

Response

{
  "query": "sterility assurance failure in aseptic processing",
  "results": [
    {
      "observation_id": "483-2024-001234-obs-003",
      "inspection_id": "483-2024-001234",
      "fei_number": "3005012345",
      "facility_name": "Acme Pharmaceutical Manufacturing",
      "city": "Boston",
      "state": "MA",
      "country": "USA",
      "inspection_end_date": "2024-03-15",
      "classification": "OAI",
      "product_type": "drug",
      "program_area": "Drugs",
      "act_cfr_number": "21 CFR 211.113",
      "short_description": "Sterile manufacturing area lacks adequate environmental monitoring",
      "observation_text": "Your firm failed to establish an adequate system for monitoring environmental conditions in aseptic processing areas. Specifically, during the inspection, it was observed that: 1) Viable air sampling was not conducted during aseptic filling operations in Room 203. 2) Surface monitoring in the ISO 5 environment was performed weekly rather than each production day. 3) Trending and investigation of environmental monitoring excursions was inadequate...",
      "similarity_score": 0.94,
      "relevance": "high",
      "confidence_level": 0.91,
      "categories": [
        "Aseptic Processing",
        "Environmental Monitoring",
        "Sterility Assurance",
        "Process Controls"
      ],
      "severity_score": 8.5,
      "facility_context": {
        "total_inspections": 8,
        "recent_classifications": ["OAI", "VAI", "NAI"],
        "warning_letters": 1,
        "last_inspection_date": "2024-03-15",
        "parent_company": "Acme Holdings Inc."
      },
      "data_attribution": {
        "api_version": "v1.0",
        "data_source_name": "FDA ORA Inspection Database",
        "data_source_url": "https://datadashboard.fda.gov/ora/cd/inspections.htm",
        "last_updated": "2024-09-15T00:00:00Z"
      }
    },
    {
      "observation_id": "483-2023-008765-obs-001",
      "inspection_id": "483-2023-008765",
      "fei_number": "3005067890",
      "facility_name": "BioTech Sterile Solutions",
      "city": "San Francisco",
      "state": "CA",
      "country": "USA",
      "inspection_end_date": "2023-11-20",
      "classification": "OAI",
      "product_type": "biologics",
      "program_area": "Biologics",
      "act_cfr_number": "21 CFR 211.113",
      "short_description": "Inadequate aseptic process validation",
      "observation_text": "Your firm failed to validate the aseptic filling process for Product X. The media fill validation performed in 2022 did not adequately simulate the commercial filling process, including line speed, fill volume, and container closure system...",
      "similarity_score": 0.88,
      "relevance": "high",
      "confidence_level": 0.85,
      "categories": [
        "Aseptic Processing",
        "Process Validation",
        "Sterility Assurance"
      ],
      "severity_score": 8.2,
      "facility_context": {
        "total_inspections": 5,
        "recent_classifications": ["OAI", "VAI"],
        "warning_letters": 0,
        "last_inspection_date": "2023-11-20",
        "parent_company": "BioTech Global Corp."
      }
    }
  ],
  "query_metadata": {
    "execution_time_ms": 1247,
    "total_results": 2,
    "indexes_searched": ["483-observations-2024", "483-observations-2023"],
    "enhanced_pipeline": true,
    "enhanced_search": {
      "query_expansion_enabled": true,
      "expanded_query": "sterility assurance failure in aseptic processing sterile manufacturing contamination",
      "abbreviations_found": [],
      "synonyms_added": ["sterile manufacturing", "contamination"],
      "domain_terms_added": [],
      "intent": {
        "program_area": "drugs",
        "query_type": "analytical",
        "cfr_section": null
      }
    }
  }
}

Response Fields

Field	Type	Description
query	string	Original search query
results	array	Array of matching observations
results[].observation_id	string	Unique observation identifier
results[].inspection_id	string	Parent inspection identifier
results[].fei_number	string	FDA Establishment Identifier
results[].facility_name	string	Facility legal name
results[].city	string	Facility city
results[].state	string	Facility state/province
results[].country	string	Facility country
results[].inspection_end_date	string	Inspection end date (ISO 8601)
results[].classification	string	Inspection classification (NAI/VAI/OAI)
results[].product_type	string	Product type
results[].program_area	string	FDA program area
results[].act_cfr_number	string	Primary CFR regulation cited (21 CFR format)
results[].short_description	string	Brief description of observation
results[].observation_text	string	Full observation text
results[].similarity_score	number	Semantic similarity score (0.0-1.0)
results[].relevance	string	Relevance level (low/medium/high)
results[].confidence_level	number	AI confidence score (0.0-1.0)
results[].categories	string[]	Categorized themes
results[].severity_score	number	Severity score (0.0-10.0)
results[].facility_context	object	Facility inspection history (if requested)
results[].data_attribution	object	Data provenance and source information
results[].data_attribution.api_version	string	API version used to generate this data
results[].data_attribution.data_source_name	string	Name of the source database
results[].data_attribution.data_source_url	string	URL to the source database
results[].data_attribution.last_updated	string	Last data refresh timestamp (ISO 8601)
query_metadata	object	Query execution metadata
query_metadata.execution_time_ms	integer	Processing time in milliseconds
query_metadata.total_results	integer	Number of results returned
query_metadata.indexes_searched	string[]	Indexes searched
query_metadata.enhanced_pipeline	boolean	Whether enhanced AI pipeline was used
query_metadata.enhanced_search	object	Query expansion and intent classification details
query_metadata.enhanced_search.query_expansion_enabled	boolean	Whether query expansion was active
query_metadata.enhanced_search.expanded_query	string|null	Expanded query used for embedding (null if unchanged)
query_metadata.enhanced_search.abbreviations_found	string[]	Abbreviations expanded (e.g., "CGMP" -> "Current Good Manufacturing Practice")
query_metadata.enhanced_search.synonyms_added	string[]	Synonyms injected from 483-specific synonym groups
query_metadata.enhanced_search.domain_terms_added	string[]	Domain-specific terms added via pattern matching
query_metadata.enhanced_search.intent	object	Classified query intent
query_metadata.enhanced_search.intent.program_area	string|null	Auto-detected FDA program area (drugs, devices, biologics, food)
query_metadata.enhanced_search.intent.query_type	string	Query type (general, analytical, definition, requirements, examples)
query_metadata.enhanced_search.intent.cfr_section	string|null	CFR section detected in query

Examples

Basic Search

curl -X POST "https://api.ctwise.ai/v1/483/observations/search" \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "data integrity issues with laboratory records",
    "top_k": 5
  }'

Search with Classification Filter

curl -X POST "https://api.ctwise.ai/v1/483/observations/search" \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "cleaning validation deficiencies",
    "filters": {
      "classification": ["OAI"],
      "product_type": "drug"
    },
    "top_k": 10,
    "include_facility_context": true
  }'

Search with Date Range and Program Area

curl -X POST "https://api.ctwise.ai/v1/483/observations/search" \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "medical device design control failures",
    "filters": {
      "date_range": {
        "start": "2024-01-01",
        "end": "2024-12-31"
      },
      "program_area": "Devices",
      "product_type": "medical_device"
    },
    "top_k": 20,
    "include_regulatory_mapping": true
  }'

Python Example

import requests

API_KEY = "your_api_key"
BASE_URL = "https://api.ctwise.ai/v1"

response = requests.post(
    f"{BASE_URL}/483/observations/search",
    headers={
        "X-Api-Key": API_KEY,
        "Content-Type": "application/json"
    },
    json={
        "query": "CAPA effectiveness not verified",
        "filters": {
            "classification": ["OAI", "VAI"],
            "date_range": {
                "start": "2023-01-01",
                "end": "2024-12-31"
            }
        },
        "top_k": 15,
        "include_facility_context": True
    }
)

data = response.json()
print(f"Found {data['query_metadata']['total_results']} results\n")

for result in data['results']:
    print(f"[{result['classification']}] {result['facility_name']}")
    print(f"  Date: {result['inspection_end_date']}")
    print(f"  Similarity: {result['similarity_score']:.2f}")
    print(f"  Description: {result['short_description']}")
    print(f"  CFR: {result['act_cfr_number']}")
    print()

JavaScript Example

const API_KEY = 'your_api_key';
const BASE_URL = 'https://api.ctwise.ai/v1';

const response = await fetch(`${BASE_URL}/483/observations/search`, {
  method: 'POST',
  headers: {
    'X-Api-Key': API_KEY,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    query: 'supplier qualification and oversight deficiencies',
    filters: {
      classification: ['OAI'],
      program_area: 'CDER'
    },
    top_k: 10,
    include_facility_context: true,
    include_regulatory_mapping: true
  })
});

const data = await response.json();
console.log(`Found ${data.query_metadata.total_results} results`);

data.results.forEach(result => {
  console.log(`${result.facility_name} (${result.state})`);
  console.log(`  Severity: ${result.severity_score}/10`);
  console.log(`  Categories: ${result.categories.join(', ')}`);
});

Error Responses

Missing Query Parameter

{
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Field 'query' is required"
  }
}

Status: 400 Bad Request

Invalid Classification Filter

{
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Invalid classification value. Allowed values: NAI, VAI, OAI"
  }
}

Status: 400 Bad Request

Invalid API Key

{
  "message": "Unauthorized"
}

Status: 401 Unauthorized

Rate Limit Exceeded

{
  "error": {
    "code": "RATE_LIMIT_EXCEEDED",
    "message": "Rate limit exceeded. Please try again later.",
    "retry_after": 60
  }
}

Status: 429 Too Many Requests

Performance

Metric	Target
Response time	< 3 seconds
Max results	100 per request
Timeout	30 seconds

Tips for Better Results

1. Use natural language - Describe what you're looking for as you would to a colleague
2. Be specific - Include context like product type, compliance area, or specific deficiency
3. Combine filters - Use classification + date range + program area for focused results
4. Request context - Set include_facility_context: true to understand repeat offenders
5. Adjust top_k - Start with 10-20 results, increase if needed for broader coverage

---

Enhanced Search Pipeline

The 483 observations search endpoint includes a domain-specific query expansion pipeline that automatically improves search quality. When enable_query_expansion is true (the default), queries go through three enhancement stages before embedding:

Stage 1: Abbreviation Expansion

Common FDA/GMP abbreviations are expanded inline. For example:

Query	Expanded
CGMP violations	CGMP (Current Good Manufacturing Practice) violations
OOS investigation	OOS (Out of Specification) investigation
CAPA effectiveness	CAPA (Corrective and Preventive Action) effectiveness
API contamination	API (Active Pharmaceutical Ingredient) contamination

Over 30 FDA/GMP abbreviations are recognized, including: OAI, VAI, NAI, GMP, CGMP, QSR, HACCP, CDER, CBER, CDRH, SOP, CAPA, OOS, OOT, DMF, BMR, BPR, HVAC, WFI, CIP, SIP, and more.

Stage 2: Synonym Injection

Domain-specific synonyms are added to broaden semantic coverage:

Term Found	Synonyms Added
data integrity	audit trail, electronic records
contamination	cross-contamination, microbial contamination
cleaning validation	cleaning verification, residue testing
sterility	sterile, aseptic, sterility assurance
deviation	nonconformance, discrepancy
complaint	adverse event, product complaint

Stage 3: Domain Pattern Enrichment

Regex patterns detect domain-specific references (CFR citations, GMP parts, 483-specific language) and add contextual terms.

Intent Classification

The search pipeline also classifies query intent to auto-apply filters:

• Program area detection: Queries mentioning "drug", "tablet", "API" auto-filter to drugs; "device", "implant" to devices; "biologics", "vaccine" to biologics; "food", "dietary" to food
• CFR section detection: Queries containing CFR references (e.g., "211.68") are tagged for cross-reference enrichment
• Query type: Classified as general, analytical, definition, requirements, or examples

Disabling Query Expansion

To use the raw query without expansion, set enable_query_expansion to false:

{
  "query": "CGMP violations in sterile manufacturing",
  "enable_query_expansion": false,
  "top_k": 10
}

---

Observation Categories

Observations are automatically categorized using NLP-based classification. Use the category filter to narrow search results to a specific category.

Available Categories

Category	Description	Example Observations
Data Integrity	Audit trails, electronic records, ALCOA+	"Failed to maintain complete data..."
Process Controls	SOPs, batch records, process parameters	"Procedures not followed for..."
Quality System	CAPA, complaints, deviations, change control	"Corrective action not effective..."
Laboratory Controls	OOS, OOT, calibration, stability	"Laboratory equipment not calibrated..."
Facility & Equipment	HVAC, water systems, cleaning, maintenance	"Facility not maintained in clean..."
Production Controls	Manufacturing, packaging, labeling	"Production records incomplete..."
Sterility Assurance	Aseptic processing, environmental monitoring	"Sterile area environmental monitoring..."
Material Controls	Raw materials, suppliers, components	"Supplier qualification inadequate..."
Documentation	Records, signatures, retention	"Batch records missing signatures..."
Personnel & Training	Training, qualifications, hygiene	"Personnel not adequately trained..."
Regulatory Compliance	GMP, 21 CFR Part 11, general compliance	"Failed to comply with CGMP..."

Example: Filter by Category

curl -X POST "https://api.ctwise.ai/v1/483/observations/search" \
  -H "X-Api-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "equipment not calibrated",
    "filters": {
      "category": "Laboratory Controls"
    },
    "top_k": 10
  }'

---

Related Endpoints

• GET /v1/483/facilities/{fei_number} - Facility inspection history
• GET /v1/483/citations/trending - Trending citations
• GET /v1/483/regulatory-mapping - CFR-to-ICH E6(R3) mappings
• POST /v1/483/watchlist - Monitor facilities

KG Intelligence (Cross-Domain)

• POST /v1/intelligence/classify - Classify an observation into event types and CFR sections
• POST /v1/intelligence/investigate - Deep investigation with evidence chain for any observation
• GET /v1/intelligence/regulations/{cfr} - Look up full CFR regulation text for cited sections
• Platform Intelligence Overview - How 483 data connects to the broader intelligence platform