To determine the feasibility of a full-scale randomised controlled trial evaluating the clinical effectiveness of the RISKINDEX based on experience gained from conducting this pilot RCT. ML mortality predictions will be compared with clinical…
ID
Source
Brief title
Condition
- Other condition
Synonym
Health condition
Acute aandoeningen bij patiënten op de SEH
Research involving
Sponsors and support
Intervention
Outcome measures
Primary outcome
- Calculated ML risk scores and observed mortality, to evaluate discriminatory
performance of ML risk score to predict 31-day mortality.
- Physicians self-reported policy changes to evaluate whether presentation of
the ML risk score causes changes in clinical decision making. Policy changes
include treatment policy, requesting ancillary investigations, treatment
restrictions (i.e., no intubation or resuscitation).
Secondary outcome
- Clinical endpoints such as 31-day mortality, ICU and MC admission and
readmission will be compared between the control an intervention group to
evaluate differences.
- Diagnostic performance of other clinical risk scores and physicians will be
compared to the ML score.
Background summary
Identifying emergency department (ED) patients at high and low risk shortly
after admission could help decision-making regarding patient care. Several
clinical risk scores and triage systems for stratification of patients have
been developed, but often underperform in clinical practice. Moreover, most of
these risk scores only have been diagnostically validated in an observational
cohort, but never have been evaluated for their actual clinical impact. In a
recent retrospective study that was conducted in the Maastricht University
Medical Center (MUMC+), a novel machine learning (ML) model was introduced that
predicted 31-day mortality of sepsis patients presenting to an ED. Follow-up
studies underlined the potential of the model also in a prospective set-up.
However, it remains unknown to what extent these models have any beneficial
value when it is actually implemented in clinical practice.
Study objective
To determine the feasibility of a full-scale randomised controlled trial
evaluating the clinical effectiveness of the RISKINDEX based on experience
gained from conducting this pilot RCT. ML mortality predictions will be
compared with clinical impression of physicians. Furthermore, the impact of the
ML model on clinical decision making will be monitored.
Study design
The MARS-ED study is designed as a multi-center, randomized, open-label,
non-inferiority pilot clinical trial.
Intervention
Physicians will be presented with the ML risk score of the patients they are
actively treating, directly after assessment of regular diagnostics has taken
place.
Study burden and risks
The only intervention in this study is presentation of a ML risk score to the
treating physician. However, this risk estimation might influence clinical
decision making and may result in ordering ancillary testing or extra
consultations. These extra investigations will probably not harm the patient
and may even improve diagnostics. In addition, the higher mortality prediction
made by the ML model will stimulate the physician to check once more on his/her
judgment of the patient and will probably help to improve this judgement. We
conclude that the benefits outweigh the risks.
P. Debyelaan 25
Maastricht 6229HX
NL
P. Debyelaan 25
Maastricht 6229HX
NL
Listed location countries
Age
Inclusion criteria
- Adult, defined as >= 18 years of age
- Assessed and treated by an internal medicine specialist in the ED
- Willing to give written consent, either directly or after deferred consent
procedure
Exclusion criteria
- <4 different laboratory results available (hematology or clinical chemistry)
within the first two hours of the ED visit (calculation ML prediction score
otherwise not possible)
- Unwilling to provide written consent, either directly or after deferred
consent procedure
Design
Recruitment
Followed up by the following (possibly more current) registration
No registrations found.
Other (possibly less up-to-date) registrations in this register
No registrations found.
In other registers
| Register | ID |
|---|---|
| Other | n.n.t.b. |
| CCMO | NL78478.068.21 |