Clinical Summaries

DELPHOS generates two types of patient summaries from accumulated clinical data — a technical summary for clinical decision-making and a rapport summary for patient relationship context. Both draw from consultation history, lab results, medications, conditions, and lifestyle data.

Patient Context

Before generating summaries, you can retrieve the structured clinical context that DELPHOS has assembled for a patient:

curl
Python

curl -X GET "https://your-instance.delphos.app/v1/patients/{patient_id}/context" \
  -H "x-api-key: $DELPHOS_API_KEY"

response = httpx.get(
    f"https://your-instance.delphos.app/v1/patients/{patient_id}/context",
    headers={"x-api-key": DELPHOS_API_KEY},
)

context = response.json()
print(f"Allergies: {len(context['allergies'])}")
print(f"Active medications: {len(context['medications'])}")
print(f"Recent consultations: {len(context['recent_consultations'])}")

The context response aggregates everything DELPHOS knows about the patient:

Section	Contents
`demographics`	Age, gender, birth date (never patient name — privacy by design)
`allergies`	Known allergies with severity and reaction type
`conditions`	Active and historical conditions with onset dates
`medications`	Current and past medications with dosage and frequency
`lifestyle`	Smoking, alcohol, exercise, and other lifestyle factors
`social`	Employment, family context, and social determinants
`recent_consultations`	Latest consultations with SOAP excerpts

Technical Summary

The technical summary is a structured clinical overview designed for physician use during consultations. It synthesizes the patient’s history into a concise narrative.

curl
Python

curl -X POST "https://your-instance.delphos.app/v1/patients/{patient_id}/summary" \
  -H "x-api-key: $DELPHOS_API_KEY"

response = httpx.post(
    f"https://your-instance.delphos.app/v1/patients/{patient_id}/summary",
    headers={"x-api-key": DELPHOS_API_KEY},
)

summary = response.json()
print(f"Summary: {summary['summary']}")
print(f"Cached: {summary['cached']}")
print(f"Generated in: {summary['generation_time_ms']}ms")

Response:

{
  "patient_id": "550e8400-e29b-41d4-a716-446655440000",
  "summary": "Paciente feminina, 41 anos, com diabetes tipo 2 diagnosticada em 2015, em uso de Metformina 500mg 2x/dia. Histórico de cefaleia tensional episódica com boa resposta a analgésicos. Última consulta em 14/04/2026: PA 120/80, sem queixas ativas. Hemograma normal em 01/04/2026. Alergia a penicilina (anafilaxia).",
  "language": "pt-BR",
  "generated_at": "2026-04-14T12:00:00Z",
  "cached": false,
  "generation_time_ms": 2500
}

Caching

DELPHOS uses content-hash-based caching for summaries. When you request a summary, the platform computes a hash of the patient’s current clinical data. If the data has not changed since the last summary was generated, the cached version is returned instantly.

To force a fresh summary (e.g., after you know new data has been added):

curl -X POST "https://your-instance.delphos.app/v1/patients/{patient_id}/summary?force_refresh=true" \
  -H "x-api-key: $DELPHOS_API_KEY"

Rapport Summary

The rapport summary provides relationship context — personal details, communication preferences, and lifestyle observations that help physicians build better patient relationships:

curl
Python

curl -X GET "https://your-instance.delphos.app/v1/patients/{patient_id}/rapport-summary" \
  -H "x-api-key: $DELPHOS_API_KEY"

response = httpx.get(
    f"https://your-instance.delphos.app/v1/patients/{patient_id}/rapport-summary",
    headers={"x-api-key": DELPHOS_API_KEY},
)

rapport = response.json()
print(f"Visit count: {rapport['consultation_count']}")
print(f"First visit: {rapport['first_visit_date']}")
print(rapport["rapport_summary"])

Response:

{
  "patient_id": "550e8400-e29b-41d4-a716-446655440000",
  "rapport_summary": "Paciente acompanhada desde janeiro de 2024. Trabalha como gerente de escritório...",
  "language": "pt-BR",
  "generated_at": "2026-04-14T12:00:00Z",
  "cached": false,
  "consultation_count": 15,
  "first_visit_date": "2024-01-10"
}

Lab Result Trends

Track laboratory values over time with trend analysis:

curl
Python

curl -X GET "https://your-instance.delphos.app/v1/patients/{patient_id}/lab-trends?\
markers=glucose,creatinine&period=12" \
  -H "x-api-key: $DELPHOS_API_KEY"

response = httpx.get(
    f"https://your-instance.delphos.app/v1/patients/{patient_id}/lab-trends",
    headers={"x-api-key": DELPHOS_API_KEY},
    params={"markers": "glucose,creatinine", "period": 12},
)

trends = response.json()
for marker in trends["markers"]:
    print(f"{marker['marker_name']}: {marker['trend']} ({len(marker['data_points'])} readings)")

Response:

{
  "patient_id": "550e8400-e29b-41d4-a716-446655440000",
  "period_months": 12,
  "total_results": 24,
  "markers": [
    {
      "marker_name": "glucose",
      "unit": "mg/dL",
      "trend": "stable",
      "reference_min": 70,
      "reference_max": 100,
      "data_points": [
        {
          "date": "2026-01-15",
          "numeric_value": 95,
          "status": "normal"
        },
        {
          "date": "2026-04-01",
          "numeric_value": 98,
          "status": "normal"
        }
      ]
    }
  ]
}

Query Parameters

Parameter	Type	Default	Description
`markers`	`string`	All markers	Comma-separated marker names
`period`	`integer`	`12`	Lookback period in months (1-120)

Trend Values

Trend	Meaning
`increasing`	Values trending upward over the period
`decreasing`	Values trending downward
`stable`	No significant change

Next Steps

Working with Records — Edit and version SOAP notes
Real-time Transcription — Audio streaming
Core Concepts — Platform fundamentals