Best Practice Advisories to Reduce Inequities in Technology Use for People With Type 1 Diabetes

For the proposed study we intend to implement an automatically delivered center decision support system provided at the time and location of the decision making along with actionable recommendations and computer based actionable recommendations. The provider will be required to opt out if not prescribing ADT and to provide a reason for not prescribing in order to advance on the EMR screen (Kawamoto et al, 2015).

Project Objectives Aim 1: To develop and implement an EMR-based BPA using stakeholder feedback to standardize the approach for prescribing and documentation of advanced diabetes technologies (ADT) (CGM, insulin pump, AID) among adult and pediatric PwT1D.

Aim 2: To determine the effectiveness of an EMR-based BPA in reducing racial inequities in ADT.

Primary Objective: To compare the proportion of non-white (non-Hispanic Black and Hispanic) PwT1D with progression in ADT use (CGM, insulin pump, AID) at 6 centers in the T1DX-QI (3 pediatric and 3 adult centers) where the BPA intervention is implemented with matched control non-white PwT1D at T1DX-QI centers not receiving the intervention over a 12-month period, adjusting for baseline level of ADT use and other confounders.

Secondary Objective: To assess the difference in ADT use between White and non-white (non-Hispanic Black and Hispanic) PwT1D receiving care at the intervention centers compared with the racial difference in ADT use in matched control PwT1D.

Aim 3: To explore the reasons identified for providers decision to not prescribe ADT and whether they were patient or provider led, and the association between the reason provided and the patient's race/ethnicity.

Methodology Overview Focus group discussions and structured interviews will be conducted among adult and pediatric endocrine providers based on the Agency for Healthcare Research and Quality (AHRQ) "Five Rights" (Flores et al, 2019) which engenders effective clinical decision support. Also other guiding frameworks such as the T1d Exchange Health Equity Framework (Ebekozien et al, 2019) will guide this process. The BPA as well as the approach to be followed for the documenting discussions between the provider and patients will be refined by the results of the qualitative study.

Using a non-randomized matched-pair intervention design, we will compare ADT use following a BPA intervention among non-Hispanic Black and Hispanic PwT1D receiving care at 6 T1DX-QI centers with matched control non-Hispanic Black and Hispanic PwT1D receiving care at a non-intervention center over a 12-month period.

Six T1DX-QI centers that are data mapped to the T1DX-QI database by study initiation will be selected for the intervention centers. There will be four additional centers as possible back-up if one or more of the 6 anticipated centers are not able to participate. Matched "control" PwT1D will also be selected from centers with mapped data.

Primary Outcome Definition: Primary outcome: Progression in ADT use (as documented in EMR) during the 12-month study period. The primary outcome would be defined as positive for an individual if any of the following occurs:

• No CGM à Any CGM
• MDI à Insulin pump
• No AID à AID

Based on this definition, a PwT1D in any of the following technology transition states would be considered as meeting the primary endpoint:

• MDI/smartpen + no CGM begins using CGM
• MDI/smartpen + CGM begins using insulin pump
• Insulin pump + no CGM begins using CGM
• Insulin pump + CGM (without AID) begins using AID

Data Analysis for Aim2 Primary: Progression in ADT use in the intervention group compared with matched-pair controls adjusting for number of encounters, correlation of matched pairs, and random center effects.

Secondary: Analysis will mimic the primary analysis for adjustment of confounders.

• Progression in AID use in the intervention group compared with matched pair controls
• Progression in CGM use in the intervention group compared with matched pair controls
• Progression in Insulin Pump use in the intervention group compared with matched pair controls
• Change in the proportional difference in CGM use between white and non-white individuals (non-Hispanic black and/or Hispanic) in the intervention group compared with matched pair controls.
• Change in the proportional difference in insulin Pump use between white and non-white individuals (non-Hispanic black and/or Hispanic) in the intervention group compared with matched pair controls
• Change in HbA1c (lab result closest to end date of period) from baseline in the intervention group compared with matched pair controls Exploratory: Analysis will mimic the primary analysis for adjustment of confounders.
• Change in proportion of sustained ADT Use in in the intervention group compared with matched pair controls. Sustained use defined as at least two consecutive EMR records with ADT use documented during the post period.

We will explore if the reasons technologies are not recommended/accepted differ by race and other factors. This aim is exploratory, and the extent of the analysis will ultimately depend on information collected as part of the BPA developed during Aim 1. If patient race is associated with differences in rates of technology prescribing/use, these findings will inform further interventions to address the contributing causes.

Age ≥ 2 years with an EMR diagnosis of T1D for at least 6 months at baseline and receiving care at one of six intervention centers or a matched control at another T1DX-QI center.

Inclusion criteria

• non-Hispanic Black and Hispanic individuals

Exclusion criteria

• PwT1D with evidence of use of AID at baseline