Tele-ID at Community Hospital: Outcomes Unchanged, Imaging Higher

Key Takeaways

• Among patients receiving ID consultations at the community hospital that switched to telehealth, 30-day readmission was 17.4% before versus 17.7% during tele-ID, and mortality was 9.2% versus 8.7%; neither difference was statistically significant.
• Adjusted analyses found no statistically significant association of tele-ID with 30-day readmission, death, or days dead or hospitalized within 90 days of the initial consultation.
• Mean CT/MRI use during the index admission was significantly higher in the tele-ID period at the switching hospital.

Community hospitals may provide inpatient infectious diseases (ID) consultations at the bedside or through telehealth infectious diseases (tele-ID) consultations. For patients receiving these consultations, whether a shift to tele-ID coincides with differences in clinical outcomes or diagnostic imaging use remains uncertain.

Investigators assembled a retrospective cohort of patients receiving inpatient ID consultations at two community hospitals before and after one switched from in-person visits to tele-ID. The switching hospital included 3,223 patients: 1,682 before the transition and 1,541 during the tele-ID period. Patients at the second hospital were included to help account for secular trends. Investigators assessed computed tomography (CT) and magnetic resonance imaging (MRI) use during the index admission, clinical outcomes at 30 days, and days dead or hospitalized within 90 days of the initial ID consultation. Multivariable analyses adjusted for patient characteristics and hospital.

At the switching hospital, median length of stay, combined death or readmission within 30 days, readmission, and mortality did not differ significantly between periods. These were unadjusted period comparisons. In multivariable analyses that included both hospitals, tele-ID was not significantly associated with 30-day readmission, death, or days dead or hospitalized within 90 days after the initial consultation.

Imaging use differed at the switching hospital. Patients received a mean of 1.69 CT or MRI studies during the index admission before the switch, compared with 2.28 during the tele-ID period (P<.001). This was a between-period comparison, not an adjusted estimate of tele-ID's effect on imaging.

The retrospective, nonrandomized comparison cannot establish that tele-ID is equivalent or noninferior to in-person ID consultation. Likewise, the imaging difference shows an association across periods, not that tele-ID caused the increase. The observations concern patients receiving consultations at two community hospitals; whether they extend to other settings remains unresolved.

The investigators concluded that tele-ID was not associated with increased measured patient-relevant adverse outcomes, although imaging utilization was greater during the tele-ID period at the switching hospital.

Clinician Questions

Which patient characteristics were included in adjusted analyses of telehealth infectious diseases consultations?

For patients receiving inpatient ID consultations at the two community hospitals, investigators adjusted tele-ID outcome models for age, sex, race, modified Charlson score, and hospital. These were model covariates, not reported covariate-specific outcome findings.

Were adjusted effect estimates reported for 30-day readmission and death after inpatient tele-ID consultations?

For patients receiving inpatient tele-ID consultations at the community hospitals, adjusted associations with 30-day readmission and death were not statistically significant. The magnitude of any difference therefore remains unclear.