Not provided
Not provided
| ID | Type | Description | Link |
|---|---|---|---|
| R01AG084343-01 | U.S. NIH Grant/Contract | View source |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Class |
|---|---|
| Stanford University | OTHER |
| National Institute on Aging (NIA) | NIH |
Not provided
Not provided
Not provided
Not provided
The goal of this clinical trial is to test an AI-based screening tool that will help to identify patients at high risk of having undiagnosed peripheral artery disease. The primary outcome measure is overall rate of new PAD diagnoses. Secondary outcomes include rate of new secondary prevention measures initiated for PAD, which will include new prescriptions for antiplatelets, PAD-dosed rivaroxaban, statins, smoking cessation counseling or referrals, and/or supervised exercise therapy referrals also aggregated at a clinic and site level.
After providers consent to participate in this study, a screening tool will be deployed for their weekly clinics to identify patients at high risk of having undiagnosed PAD. These high risk alerts will be provided after a patient has checked in for their outpatient appointment. The alert will be sent to their treating provider once the visit is initiated in the electronic health record system (EHR). The primary outcome measure is overall rate of new PAD diagnoses. Secondary outcomes include rate of new secondary prevention measures initiated for PAD, which will include new prescriptions for antiplatelets, PAD-dosed rivaroxaban, statins, smoking cessation counseling or referrals, and/or supervised exercise therapy referrals also aggregated at a clinic and site level. For secondary analysis we will specifically evaluate patients who generated an alert and assess how patient demographics and/or clinical factors are associated with likelihood of ABI testing, rate of abnormal ABIs (i.e. true positive rate), and subsequent initiation of secondary prevention measures.
UC San Diego Health (UCSDH), VA San Diego Health Care (VASDHC), and Stanford Health Care (SHC) will be the sites for study enrollment. UCSDH - La Jolla campus, UCSDH - Hillcrest campus, and VASDHC will begin a pre-intervention observation period at the same time, and then each site will be randomized to begin screening tool intervention in a stepped wedge pattern at 13-week intervals for a total of 52 weeks. We will enroll 10 clinics per site based on power calculations for number of patients needed to screen each week and to minimize the number of alerts per clinic/ provider. After this 52 week period, the Stanford site will serve as a validation site and will undergo randomization of 10 clinical sites to three 13 week intervals for a total of 52 weeks.
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| Clinical Site 1 | Experimental | Randomized to start AI-based PAD screening interventionat week 13. |
|
| Clinical Site 2 | Experimental | Randomized to start AI-based PAD screening intervention at Week 26. |
|
| Clinical Site 3 | Experimental | Randomized to start AI-based PAD screening intervention at Week 39. |
|
| Name | Type | Description | Arm Group Labels | Other Names |
|---|---|---|---|---|
| AI-based PAD screening intervention | Diagnostic Test | Providers will receive alerts for a patient that is flagged by model as being "high risk" for PAD. This will allow the provider to review the alert, check the patient's previous history, develop additional questions to assess the risk of PAD, and initiate orders prior to seeing a patient. Depending on their assessment during the patient visit the provider may choose to order an ABI test (or perform one at bedside) and/or initiate other secondary prevention measures. All patients for which an alert is triggered will be included for secondary analysis. |
| Measure | Description | Time Frame |
|---|---|---|
| PAD Diagnosis Rate | The primary outcome will be counted at a clinic and site level and will include number of new abnormal ABI tests (ABI< 0.9), and new diagnosis codes, procedures or affirmative text mentions for PAD for patients without a previous diagnosis | During 13-39 weeks prior to intervention compared to 13-39 weeks during intervention depending on timing of randomization to intervention period. |
| Measure | Description | Time Frame |
|---|---|---|
| Initiation of secondary prevention measures | New prescriptions for antiplatelets, PAD-dosed rivaroxaban, statins, smoking cessation counseling or referrals, and/or supervised exercise therapy referrals also aggregated at a clinic and site level time period. | During 13-39 weeks prior to intervention compared to 13-39 weeks during intervention depending on timing of randomization to intervention period. |
Not provided
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Not provided
Not provided
| Name | Role | Phone | Extension | |
|---|---|---|---|---|
| Kathleen Groh | Contact | 8585348103 | kagroh@health.ucsd.edu |
| Name | Affiliation | Role |
|---|---|---|
| Elsie Ross, MD, MSc | UC San Diego | Principal Investigator |
Not provided
| PubMed Identifier | Type | Citation | Retractions |
|---|---|---|---|
| 35922657 | Background | Ghanzouri I, Amal S, Ho V, Safarnejad L, Cabot J, Brown-Johnson CG, Leeper N, Asch S, Shah NH, Ross EG. Performance and usability testing of an automated tool for detection of peripheral artery disease using electronic health records. Sci Rep. 2022 Aug 3;12(1):13364. doi: 10.1038/s41598-022-17180-5. |
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D058729 | Peripheral Arterial Disease |
| ID | Term |
|---|---|
| D050197 | Atherosclerosis |
| D001161 | Arteriosclerosis |
| D001157 | Arterial Occlusive Diseases |
| D014652 | Vascular Diseases |
Not provided
Not provided
A stepped wedge cluster randomization design was chosen as a pragmatic way to evaluate the "real world" impact of AI-based PAD screening. The stepped wedge design has been used to evaluate a variety of interventions, including digital health-based studies. This particular design allows for analysis within and between clusters and can reduce the total number of clusters needed to see an effect, helping increase statistical power compared to parallel cluster randomization. A stepped wedge design, like other cluster randomization designs, also helps reduce possible contamination effects. By using institutions as the basis for clustering, we minimize the possibility that physicians increase their PAD diagnosis rates based on knowledge of the screening tool from adjacent clinics rather than direct use.
Not provided
Not provided
Not provided
Not provided
|
| D002318 |
| Cardiovascular Diseases |
| D016491 | Peripheral Vascular Diseases |