Format

Send to

Choose Destination
J Pediatr Endocrinol Metab. 2019 Sep 6. pii: /j/jpem.ahead-of-print/jpem-2019-0156/jpem-2019-0156.xml. doi: 10.1515/jpem-2019-0156. [Epub ahead of print]

Validation of a risk screening tool for pediatric type 1 diabetes patients: a predictor of increased acute health care utilization.

Author information

1
Nationwide Children's Hospital, Pediatric Endocrinology, 700 Children's Drive, Columbus, OH 43205, USA, Phone: +614-722-8836, Fax: +614-722-4440.
2
Nationwide Children's Hospital, Pediatric Endocrinology, Columbus, OH, USA.
3
Nationwide Children's Hospital, Service Line Quality Improvement, Columbus, OH, USA.
4
Nationwide Children's Hospital, Biostatistics Core and Critical Care Medicine, Columbus, OH, USA.
5
Nationwide Children's Hospital, Ohio State University, Pediatric Endocrinology, Columbus, OH, USA.

Abstract

Background Both psychosocial and socioeconomic risk factors contribute to poor glycemic control (GC). Previous research has identified that diabetes care behaviors are generally 'set' by late childhood, further highlighting the importance of psychosocial screening and intervention in the early course of disease management. The purpose of the current study was to determine whether this brief risk assessment tool is associated with GC and acute health care (HC) utilization, and to evaluate the discriminatory utility of the tool for predicting poor outcomes. Methods This was a retrospective cohort design in which we compared risk assessment scores with health outcomes at 6, 12, and 18 months after new-onset type 1 diabetes diagnosis for 158 patients between 2015 and 2017. The two primary outcome variables were GC and acute HC utilization. Results Our data demonstrate that the greatest utility of the tool is for predicting increased acute HC utilization. It was most useful in differentiating between patients with vs. without any acute HC utilization, with excellent discriminatory ability (area under the receiver operator characteristic curve [AUC] = 0.93), sensitivity (90%), and specificity (97%). Conclusions Knowledge of the risk category in addition to identification of individual risk factors within each domain allows for not only clear treatment pathways but also individualized interventions. The risk assessment tool was less effective at differentiating patients with poor GC; however, the tool did have high specificity (83%) for predicting poor GC at 18 months which suggests that the tool may also be useful for predicting patients at risk for poor GC.

KEYWORDS:

acute health care utilization; hemoglobin A1c; psychosocial risk; type 1 diabetes

PMID:
31490774
DOI:
10.1515/jpem-2019-0156

Supplemental Content

Full text links

Icon for Sheridan PubFactory
Loading ...
Support Center