810 - Utilization and Perceived Utility of Genetic Testing for Inpatient Neonates

Publication Number: 810.7019

Emily A. Messick, Nationwide Children's Hospital, Columbus, OH, United States; Chance Alvarado, Nationwide Children's Hospital, Columbus, OH, United States; Sara Conroy, Nationwide Children's Hospital, Columbus, OH, United States; Austin A. Antoniou, Nationwide Children's Hospital, Columbus, OH, United States; Julie Strominger, Nationwide Children's Hospital, Columbus, OH, United States; Bimal P. Chaudhari, Nationwide Children's Hospital, Columbus, OH, United States

Background: Chromosomal microarray has historically been a first line genetic test for critically ill neonates. With the recent advent of rapid genome sequencing (rGS), genetic testing strategies in this population are in flux. Identification of which infants should receive genetic testing, when testing should occur, and which test is most useful remains uncertain.

Objective: Evaluate the clinical utility of genetic testing from the clinician perspective for inpatient neonates as currently practiced and identify factors associated with the highest utility.

Design/Methods: Interim analysis of a prospective observational study of neonates admitted 7/1/24-10/22/24 to Nationwide Children's Hospital at age < 28 days with genetic testing during index hospitalization. Infants were excluded if test results were not available during the admission. The primary provider’s global impression of genetic testing utility was captured on a Likert scale via electronic survey 2-14 days after result return for each test. Rater demographics, infant characteristics and outcome data were collected. An “early-broad” testing strategy was defined as rGS results ≤14 days of admission. When multiple tests were assessed for an individual patient, aggregate maximum utility was considered when evaluating on the patient-level basis. Associations of clinical utility with strategy, indication, diagnostic yield, and type of test were estimated using univariable logistic regression. Utility differences were calculated on the survey level for result timing, rater role, and years of clinical experience stratified by type of test.

Results: Forty-one utility measures obtained from thirty-nine infants with 68 genetic tests were included (Figure 1). Most testing (85%) was sent from the neonatal intensive care unit (Table 1). Genetic testing significantly or somewhat improved care for 64% of tests, with none worsening or impeding care. An early-broad strategy significantly or somewhat improved care for 73% (n=8) vs. 62% (n=13) for other strategies. Patient-level utility was higher when at least one test was diagnostic and when indication for testing was other than “multiple congenital anomalies” (Figure 2).

Conclusion(s): In this single center study, an early-broad testing strategy was associated with higher clinical utility compared to other strategies for inpatient neonates. No testing was considered to have worsened or impeded care. Continued data accrual and evaluation into the testing strategies with the highest utility for this population is needed.

Figure 1: Study Flow Diagram


Table 1a-c: Infant Characteristics, Patient Level Testing Factors, and Test Level Factors by Testing Strategy


Figure 2: Percentage of Positive Utility Rating by Patient Level Factors and Test Level Factors
Figure 2 illustrates percent stating testing improved or enabled care at the patient level (A) and test level (B and C) stratified by type of test, for factors of interest. For comparative summaries, perceived utility was dichotomized as Yes-“testing improved or enabled care” vs. No-“care was unchanged by testing”. Yes encapsulated “significantly improved or enabled care” and “somewhat improved or enabled care”. This was performed given low counts not facilitating an ordinal-response model. Categorical features were binned given a small overall sample and sparse responses for certain feature levels (Diagnostic yield: diagnostic vs. all other results; Respondent experience: > 10 years vs. ≤ 10 years; Respondent role: attending vs. all other roles; Indication for testing: multiple anomalies vs. all other indications). No adjustment was made for within-respondent correlation. Early-broad strategy was defined by any genome sequencing that yielded results ≤14 days of admission. Odds of perceived utility were evaluated using univariable linear regression for patient-level factors. Unable to model estimated odds of perceived utility at the test-level due to small counts and numerous utility ratings at 100% within test-type stratifications. Figure 2 (A) illustrates the percent stating testing improved or enabled care at the patient level by testing indication, testing strategy, and maximum diagnostic yield. The aggregate maximum indicates greater utility with early-broad strategy and when at least one test is diagnostic [OR 1.64, 95% CI (0.35-9.20), OR 3.38, 95%CI (0.66-25.9), respectively]. There was lower utility when indication for testing was multiple anomalies [OR 0.60, 95% CI (0.13-2.73)]. Figure 2 (B) depicts percent stating testing improved or enabled care at the test level for microarray by indication, testing strategy, diagnostic yield, time to results (≤14 days of admission vs. after), rater role, and rater clinical experience. Figure 2 (C) depicts percent stating testing improved or enabled care at the test level for rapid genome by indication, testing strategy, diagnostic yield, time to results (≤14 days of admission vs. after), rater role, and rater clinical experience.
OR,Odds ratio; CI,Confidence interval

Figure 1: Study Flow Diagram


Table 1a-c: Infant Characteristics, Patient Level Testing Factors, and Test Level Factors by Testing Strategy


Figure 2: Percentage of Positive Utility Rating by Patient Level Factors and Test Level Factors
Figure 2 illustrates percent stating testing improved or enabled care at the patient level (A) and test level (B and C) stratified by type of test, for factors of interest. For comparative summaries, perceived utility was dichotomized as Yes-“testing improved or enabled care” vs. No-“care was unchanged by testing”. Yes encapsulated “significantly improved or enabled care” and “somewhat improved or enabled care”. This was performed given low counts not facilitating an ordinal-response model. Categorical features were binned given a small overall sample and sparse responses for certain feature levels (Diagnostic yield: diagnostic vs. all other results; Respondent experience: > 10 years vs. ≤ 10 years; Respondent role: attending vs. all other roles; Indication for testing: multiple anomalies vs. all other indications). No adjustment was made for within-respondent correlation. Early-broad strategy was defined by any genome sequencing that yielded results ≤14 days of admission. Odds of perceived utility were evaluated using univariable linear regression for patient-level factors. Unable to model estimated odds of perceived utility at the test-level due to small counts and numerous utility ratings at 100% within test-type stratifications. Figure 2 (A) illustrates the percent stating testing improved or enabled care at the patient level by testing indication, testing strategy, and maximum diagnostic yield. The aggregate maximum indicates greater utility with early-broad strategy and when at least one test is diagnostic [OR 1.64, 95% CI (0.35-9.20), OR 3.38, 95%CI (0.66-25.9), respectively]. There was lower utility when indication for testing was multiple anomalies [OR 0.60, 95% CI (0.13-2.73)]. Figure 2 (B) depicts percent stating testing improved or enabled care at the test level for microarray by indication, testing strategy, diagnostic yield, time to results (≤14 days of admission vs. after), rater role, and rater clinical experience. Figure 2 (C) depicts percent stating testing improved or enabled care at the test level for rapid genome by indication, testing strategy, diagnostic yield, time to results (≤14 days of admission vs. after), rater role, and rater clinical experience.
OR,Odds ratio; CI,Confidence interval