Optimizing veteran outreach for performance on diabetes lab monitoring
Diabetes laboratory monitoring among Veterans with diabetes overdue for lab monitoring and/or retinal screening
This 2-stage SMART trial is a prospective, adaptive, individually randomized trial consisting of two randomization rounds. Randomization is computer-generated using the R package blockrand. Stage 1: Randomization: *Population: Veterans empaneled in primary care at site 663 (Puget Sound), identified on a rolling monthly basis as meeting criteria for diabetes lab monitoring or retinal eye health outreach. Specifically, eligible patients are those who appear in the denominator of at least one component eQM (dmg23h_ec subgroup, dmg31h_ec, or ked01_ec) and simultaneously fail to meet the numerator criterion for that measure (that is, they are actively due for the associated monitoring activity) at the time of monthly eligibility assessment. *Randomization: Patients are randomized 1:1 to receive either: Vet-Text: Automated SMS health system alert Letter: Automated printed letter *Stratification: By status of lab 'not done' for eQM dmg23h_ec in month prior to randomization (binary: done/not done), applied only among patients eligible for the dmg23h_ec subgroup at enrollment. Patients enrolled exclusively via ked01_ec or dmg31h_ec eligibility (and therefore not in the dmg23h_ec subgroup denominator) are not stratified on this variable. *Outcome: The primary outcome is a patient-level binary: whether the patient met the numerator criterion for at least one component eQM for which they were eligible at enrollment, by the conclusion of the trial observation window. The Weighted and Replicated Estimator (WRE) will be used for embedded dynamic treatment regimen (EDTR) means. Stage 2: Adaptive Re-Randomization *Eligibility: Patients who do not respond or complete any eligible metric numerator after a 12-week period from entry (remain as 'metric not met' in all components for which they were enrolled, and remain in the denominator of those eQMs), post-Round 1 randomization. *Randomization: Patients are randomized 1:1 to low-effort or high-effort arms: Arm 1: Vet-Text
1. Veterans appearing in the denominator of at least one component electronic quality measure (eQM) (dmg23h_ec subgroup, dmg31h_ec, or ked01_ec) who are not yet meeting the numerator criterion for that component at the time of monthly eligibility assessment (i.e., they are actively due for the associated monitoring activity) 2. Valid contact information 3. Assigned to participating VA site (663) 4. Ages 18-85 (note: dmg23h_ec and dmg31h_ec eligibility is restricted to ages 18-75 per eQM specification; ked01_ec eligibility extends to age 85; patients ages 76-85 may be enrolled via ked01_ec eligibility alone) 5. Have assigned primary care provider in PCMM 6. Have had at least one inpatient or outpatient encounter anywhere in VHA within the three years prior to the eQM reporting period end date
1. Assigned to H-PACT, GERI, SCI, or HBPC 2. Receiving hospice services in prior year (not included in eQM denominator) 3. 65 years with frailty and advanced illness (dementia medication, advanced illness definition set, or 2 instances of frailty; not included in eQM denominator) 4. Invalid/bad mailing address and invalid/non-working phone number (no valid outreach contact)