Table 2.
Implementation barriers
| Barriers (stage) | Component or element | A | B | C | D | E | F | G | H | Total | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| B1 | Lack of clinician trust (IV+)18 | ICA | 2 | 1 | 1 | 1 | 5/4 | ||||
| B2 | MLA retraining concerns: Feedback loops arise when alerts lead to timely treatment. (IV+) | AI | 1 | 1/1 | |||||||
| B3 | Alerts dismissed for wrong reasons, for example, patients with no sepsis symptoms or with higher acute complexity (IV+) | CW | 2 | 2 | 1 | 5/3 | |||||
| B4 | Alert fatigue (II+)18 | AI; CW; ICA | 1 | 1 | 1 | 3/3 | |||||
| B5 | Differential nurse/Doctor role, perceptions of role and value (IV+)18 | CW; ICA | 1 | 1 | 2/2 | ||||||
| B6 | Inherent limitations of EHR data, which can be plagued by missingness, inaccuracies, and changes in practice patterns over time (II+) | DP | 1 | 1 | 2/2 | ||||||
| B7 | Data entry delays, leading to delayed predictions (III+) | DP | 1 | 1/1 | |||||||
| B8 | Inventors/company equity owners may have COI and inadvertently act in bias ways towards the evaluation of their system (III+) | ET | 1 | 1/1 | |||||||
| B9 | Surveillance bias: important to monitor impact of Alerts on non-septic patients for over-prescription of antibiotics (IV+) | ET; EM | 1 | 1/1 | |||||||
| B10 | Substantial cost involved for infrastructure, implementation personnel time and ongoing maintenance (All)18 | ICA; GOV | 1 | 1/1 | |||||||
| B11 | Lack of individual proficiency of health professionals in the use of hardware and software (IV+) | CW; ICA | 1 | 1/1 | |||||||
| B12 | Clinicians perceive they are better at diagnosing sepsis than the AI and the alert occurs after they already suspect (IV+)18 | CW; AI; ICA | 1 | 1 | 2/2 | ||||||
| B13 | Lack of machine learning foundational knowledge and firsthand experience (II+)63 | CW; ICA | 1 | 1/1 | |||||||
| B14 | Clinician concern over reliance on system (IV+)63 | GOV; QS; ICA | 1 | 1/1 | |||||||
| Total papers | 8 | 0 | 3 | 9 | 5 | 1 | 0 | 1 | 27 | ||
| Count of barriers identified by the group | 6 | 0 | 3 | 9 | 4 | 1 | 0 | 1 | 24 |
For reach barrier, the number of papers that identify the barrier within each group are noted in columns A to H. The totals column is in the format of: total number of papers/total number of groups. The associated element or component in the derived framework is also identified where ICA: Implementation, change management & adoption; AI: AI model; CW: clinical workflow; DP: data pipeline; GOV: governance; QS: Quality & safety; EM: Evaluation and monitoring. Beside each barrier is listed the stage in parentheses, that is associated with that barrier.