Table 3.
Survey results.
| Survey question | Values | |
| What is your preferred format for the tool? (average rank, higher score=more desirable) | ||
|
|
Patient header | 4.59 |
|
|
Text message | 4.03 |
|
|
Glucose management report | 3.81 |
|
|
BPAa | 3.65 |
|
|
Patient list | 2.94 |
|
|
Epic InBasket message | 1.98 |
| What piece of information would you like to see included in the real-time alert? (select all that apply; n=102), n (%) | ||
|
|
Specific reason or reasons a patient is at risk based on the prediction model | 80 (78.4) |
|
|
Categorized risk of hypoglycemia (eg, low, medium, high) | 66 (64.7) |
|
|
Patients’ estimated probability of hypoglycemia | 64 (62.8) |
|
|
Hyperlink to actual prediction model | 24 (23.5) |
|
|
Validated accuracy of the model | 17 (16.7) |
|
|
None of the above is necessary | 3 (2.9) |
| What feature would you like to see incorporated in the real-time alert? (select all that apply; n=102), n (%) | ||
|
|
Recommended action | 90 (88.2) |
|
|
Ability to acknowledge the alert | 61 (59.8) |
|
|
Ability to ignore or override alert | 61 (59.8) |
|
|
Direct link to subcutaneous insulin order set | 54 (52.9) |
|
|
Direct link to subcutaneous insulin decision support tool | 29 (28.4) |
|
|
Ability to consult endocrinology or inpatient diabetes management service | 43 (42.2) |
|
|
None of the above | 1 (0.9) |
| Should the real-time informatics alert trigger an endocrinology consult? inpatient diabetes management service consult? (n=101), n (%) | ||
|
|
No | 85 (84.2) |
|
|
Yes | 16 (15.8) |
| What is your preferred channel for the alert tool? (select all that apply; n=102), n (%) | ||
|
|
CORUS text message | 63 (61.8) |
|
|
BPA tool | 52 (50.9) |
|
|
Other channels | 15 (14.7) |
|
|
Epic inBasket message | 5 (4.9) |
| When would you like to receive the alert in your workflow? (single choice; n=102), n (%) | ||
|
|
As soon as hypoglycemia risk is detected | 47 (46.1) |
|
|
At the same time everyday | 27 (26.5) |
|
|
When opening the EMRb of a patient predicted to be high risk | 24 (23.5) |
|
|
Other | 4 (3.9) |
| Who should receive the alert? (select all that apply; n=102), n (%) | ||
|
|
Person listed as first call on the EMR | 86 (84.3) |
|
|
Nurse | 47 (46.1) |
|
|
Attending physician | 15 (14.7) |
|
|
Clinical nurse specialist (nurse practitioner) | 15 (14.7) |
| What is lowest sensitivity you’d consider clinically acceptable for the proposed alert tool? (single choice; n=101), n (%) | ||
|
|
50%-59% | 3 (2.9) |
|
|
60%-69% | 10 (9.9) |
|
|
70%-79% | 37 (36.6) |
|
|
80%-89% | 37 (36.6) |
|
|
90%-100% | 14 (13.9) |
| What is lowest specificity you’d consider clinically acceptable for the proposed alert tool? (single choice; n=100), n (%) | ||
|
|
50%-59% | 3 (3) |
|
|
60%-69% | 6 (6) |
|
|
70%-79% | 30 (30) |
|
|
80%-89% | 41 (41) |
|
|
90%-100% | 20 (2) |
| Are you interested in participating in a clinical design team to build the informatics alert? (n=96), n (%) | ||
|
|
No | 73 (76) |
|
|
Yes | 23 (24) |
aBPA: best practice advisory.
bEMR: electronic medical record.