Table 3 Relevancy of an AI-enhanced care pathway planning and scheduling system functionalities

Relevancy of unit-level recommendations for operation.

1. It is important that AI is able to make individual patient profiles based on previous data

0.82

2

2. It is important that AI can suggest the best possible timing for a treatment or visit based on patient risks and predicted patient flow

0.9

3

3. It is important that AI recognizes if the patients are in risk for adverse events during the care

0.96

1

4. It is important that AI takes into account uncontrollable/external factors (weather, major events, holidays) while scheduling for the care

0.42

9

5. It is essential that AI identifies the most relevant examinations and tests when planning a treatment event

0.86

5

6. It is important that AI is able to recognize unnecessary laboratory tests for individual patients for the treatment/care

0.66

13

7. It is important for AI to identify patient level patterns that cause last minute cancellation of treatment/care

0.72

11

8. AI must estimate the duration of all phases of treatment events

0.54

10

9. AI idenfies and takes into account adverse or atypical events when predicting the care pathway and duration of its phases.

0.71

12

10. It is important that AI recommends the best time for the planned treatment for the patient, taking into account other scheduled patients, patient flow and available resources.

0.84

7

11. It is important for AI to identify patterns from the live data that predicts deterioration in the patient`s chronic diseases.

0.82

6

12. It is important that AI is able to use data from home sensors, wearables and robots when predicting individual patient’s symptom progression

0.76

8

13. It is important that AI recognizes patterns from patient data that are associated with deterioration during the following 24 h

0.88

4

Relevancy of unit-level recommendations for patients predicted perioperative process/patient flows

1. It is important that AI is able to predict available resources for certain time points based on data of internal and external factors

0.86

2

2. It is important that AI is able to estimate the needed time resources for the scheduled day.

0.84

4

3. It is important that AI is able to recognize the possible factors and patterns causing adverse events after care or prolonged need for care

0.82

1

4. It is important that AI is able to recognize the days of increased need for care and increased need for resources (“high flow days”) and the factors causing those

0.88

3

5. It is important that AI is able to predict the days of decreased personnel resources in care

0.76

5

6. It is important that AI is able to predict the average duration of provided care

0.76

6

Importance of patient-level functionalities in the UI

1. The user interface has a visualization of predicted patient flow and a reasoning behind it for a particular patient (predicted duration of each phase of his/her care path)

0.78

2

2. The user interface has functionalities for finding an appropriate and right timing for a particular patient’s treatment

0.88

3

3. The user interface updates the visualization of the patient flow for a particular patient during care based on existing data enriched with live data

0.73

4

4. The user interface has a report/list of the patients whose treatment is anticipated to be at risk for cancellation

0.66

5

5. The user interface updates the visualization of predicted evolution of patient’s condition based on the historical and live patient data

0.84

1

Importance of unit-level functionalities in the UI

1. The user interface has a visualization of units/hospitals general patient flow (all predicted patients flows)

0.72

3

2. The user interface updates the visualization of the patient flow for a particular patient during care based

0.76

2

3. The user interface has a view of the recommended order of patients’ treatment

0.82

1

4. The user interface has a report/list of similar patients to replace cancelled ones

0.70

4

5. The user interface has a view of predicted changes in the patient flow based on uncontrollable factors like weather, major events and holiday seasons

0.54

5

Importance of functionalities in the UI (unit resources)

1. The user interface has a functionality to check if staff availability is anticipated to be limited during planned treatment time

0.78

1

2. The user interface has a functionality to check if hospital capacity is anticipated to be limited during plan planned treatment time

0.86

2

3. The user interface has a visualization of predicted hospital capacity as a replicate of hospital environment (i.e. digital twin)

0.56

3

4. The user interface has a report/list of predicted financial impacts of the clinical care process

0.44

4