Hospital Consultation From Outpatient Clinicians for Medically Complex Children: A Randomized Clinical Trial

Abstract

Importance: Children with medical complexity (CMC) frequently experience fragmented care. We have demonstrated that outpatient comprehensive care (CC) reduces serious illnesses, hospitalizations, and costs for high-risk CMC. Yet continuity of care for CMC is often disrupted with emergency department (ED) visits and hospitalizations.

Objective: To evaluate a hospital consultation (HC) service for CMC from their outpatient CC clinicians.

Design, setting, and participants: Randomized quality improvement trial at the University of Texas Health Science Center at Houston with an outpatient CC clinic and tertiary pediatric hospital (Children's Memorial Hermann Hospital). Participants included high-risk CMC (≥2 hospitalizations or ≥1 pediatric intensive care unit [PICU] admission in the year before enrolling in our clinic) receiving CC. Data were analyzed between January 11, 2018, and December 20, 2019.

Interventions: The HC included serial discussions between CC clinicians, ED physicians, and hospitalists addressing need for admission, inpatient treatment, and transition back to outpatient care. Usual hospital care (UHC) involved routine pediatric hospitalist care.

Main outcomes and measures: Total hospital days (primary outcome), PICU days, hospitalizations, and health system costs in skeptical bayesian analyses (using a prior probability assuming no benefit).

Results: From October 3, 2016, through October 2, 2017, 342 CMC were randomized to either HC (n = 167) or UHC (n = 175) before meeting the predefined bayesian stopping guideline (>80% probability of reduced hospital days). In intention-to-treat analyses, the probability that HC reduced total hospital days was 91% (2.72 vs 6.01 per child-year; bayesian rate ratio [RR], 0.61; 95% credible interval [CrI], 0.30-1.26). The probability of a reduction with HC vs UHC was 98% for hospitalizations (0.60 vs 0.93 per child-year; RR, 0.68; 95% CrI, 0.48-0.97), 89% for PICU days (0.77 vs 1.89 per child-year; RR, 0.59; 95% CrI, 0.26-1.38), and 94% for mean total health system costs ($24 928 vs $42 276 per child-year; cost ratio, 0.67; 95% CrI, 0.41-1.10). In secondary analysis using a bayesian prior centered at RR of 0.78, reflecting the opinion of 7 experts knowledgeable about CMC, the probability that HC reduced hospital days was 96%.

Conclusions and relevance: Among CMC receiving comprehensive outpatient care, an HC service from outpatient clinicians likely reduced total hospital days, hospitalizations, PICU days, other outcomes, and health system costs. Additional trials of an HC service from outpatient CC clinicians are needed for CMC in other centers.

Trial registration: ClinicalTrials.gov Identifier: NCT02870387.

Figure 1.. CONSORT Diagram

Patient recruitment, randomization, and follow-up.

Figure 2.. Posterior Probability of Decreased Total Hospital Days With Hospital Consultation (HC) vs Usual Hospital Care (UHC) With a Skeptical Prior (A) and an Expert Prior (B)

A, The curves represent probability densities and are scaled so that the total area under the curve is 1, and the area between any 2 values on the x-axis equals the probability of observing a value in that range. The blue line plots a skeptical prior distribution centered at a rate ratio of 1.0 (95% credible interval, 0.3-3.3) and indicates an equal number of children would be expected to benefit from either HC or UHC. The posterior probability of intervention effect is derived by combining the prior distribution with the trial results. The posterior distribution (orange line) is shifted to the left of a rate ratio of 1.0 with a median of 0.61. The area under the curve that is less than a rate ratio of 1.0 (light gray) represents the posterior probability of any reduction in total hospital days in HC group (91% for this trial). The area under the curve that is greater than a rate ratio of 1.0 (blue) represents the posterior probability of an increase in total hospital days in HC group (9% for this trial). B, Under an expert prior that is more optimistic a priori of intervention benefit (blue line centered at rate ratio of 0.78), the posterior distribution is more tightly distributed around the point estimate (median) of 0.69 indicating greater benefit of the intervention reflected in a 96% probability of reduced total hospital days in HC group.