| HF design principles | Implementation in PE-Dx CDS |
| Automation of information acquisition35 | Auto-population of some Wells’ and PERC criteria using EHR data |
| Automation of information analysis35 | Computation of Wells’ score by CDS |
| Support of decision selection35 | Provision of recommendation for next step in diagnostic pathway, for example, doing nothing, ordering D-dimer, ordering CTA scan |
| Explicit control/flexibility35 36 | Ability to change values for Wells’ criteria—for example, possible to change heart rate/pulse—to support physician clinical judgement and unique patient situations |
| Minimisation of workload36 | Minimisation of data entry; for example, data for Wells’ automatically populated in PERC PERC appears only if Wells’ score is low No need to enter data for all PERC criteria once any PERC criterion is positive Automatic generation of text for documentation of medical decision-making |
| Consistency36 37 | Consistency with how information is presented in other parts of the EHR; for example, use of Yes/No toggle Consistency with how Wells’ and PERC criteria are listed on MDCalc website that is routinely used by physicians |
| Chunking/grouping36 74 | Wells’ criteria and PERC rule presented separately Placement of CDS in ED navigator of EHR |
| Visibility37 | Indication of points/weights assigned to each Wells’ criterion to make it clear/transparent how Wells’ score is computed |
| Error prevention36 37 | To avoid documenting wrong Wells’ score, all Wells’ criteria must be addressed |
HF, human factors; PERC, pulmonary embolism rule-out criteria.