Table 4 Examples of clinical questions and a corresponding framing.

From: The Framing of machine learning risk prediction models illustrated by evaluation of sepsis in general wards

Question (*Will this patient…)

When to predict

Framing structure

Lead window

Prediction window

Window Shift

*Develop sepsis within the next 12 h?

Each hour

SEQ/SW

NA

12 h

1 h

*Develop sepsis within the next 12 h?

Preoperatively

At event

NA

12 h

NA

*Experience a preventable hospitalization within 90 days?

Each week

SEQ/SW

NA

90 days

1 week

*Experience a preventable hospitalization within 90 days?

GP visits

At event

NA

90 days

NA

*Develop acute kidney injury during this hospitalization?

Admission

At event

NA

Entire admission

NA

*Need medical attention during the day tomorrow?

During round

At event/SW/SEQ

12 h

12 h

NA

How long does this patient have to live?

At diagnosis

Survival analysis

NA

Until study ends

NA

Expected P-sodium levels this time tomorrow

At home visit

OCD

18 h

12 h

NA

  1. SEQ sequential, SW sliding window, OCD on clinical demand, GP general practitioner, NA not available.