Implementation of a stroke/TIA e-consult increases outpatient stroke Teleneurology consults placed in VHA compared to...

E-consults for outpatient stroke/TIA care can improve timeliness of neurologic care. For rural Veterans, many of whom are eligible for VA-paid community care, it is unknown whether active implementation of an e-consult within the Veterans Health Affairs (VA) heath system impacts the overall volume of outpatient stroke care provided in the VA vs. in non-VA community care.

We studied the implementation of an e-consult for outpatient stroke/TIA care in a stepped-wedge trial in 10 VA facilities in the VA National TeleNeurology Program (NTNP). Sites were randomized to one of three sequential six-month active implementation waves (after two months of usual care where the e-consult was available) that included a stroke prevention lecture for primary care (PC) providers, as well as participation in monthly PC meetings to review cases and utilization of the consult. After the six-month active implementation (sustainability period), no data were presented to PC teams. The primary outcome was whether a consult was placed in the VA vs. community care neurology (CCN) for outpatient stroke/TIA care. A generalized linear mixed model (GLMM) with binomial distribution and log link fit to the primary outcome was used to estimate the effect of active implementation with baseline serving as the reference. Other effects in the model included a priori selected variables: demographics, site data (stroke volume, median consult wait time, rurality, availability of local neurology, indicator for project time period-baseline, active, sustainability), time block (as categorical two-month blocks) and random site effect. A similar model was fit including four control sites, with an added site-level indicator for intervention vs. control site.

Of the 1,680 consults placed for stroke/TIA, 22% (365/1,680) were NTNP (within VA). The GLMM model demonstrated that the odds of a stroke/TIA consult being placed within the VA increased during active implementation: OR 1.92 (1.17-3.16, p = 0.01) but this effect did not remain significant during sustainability (OR 1.85, 95% CI 0.83-4.13). Adding data from the four NTNP control sites who had access to the e-consult but did not receive the active implementation showed no trend in consult location over time but added additional variability to the model, resulting in a decrease in the effect (OR 1.48, 95% CI 0.96-2.28, p = 0.07).

Active implementation of a VA stroke/TIA e-consult may increase the proportion of neurology consults placed to VA compared to community care. Further implementation efforts to sustain this effect and better understand factors influencing patient and provider choices about where to receive care are needed.