Application of the data quality framework to administrative data on child maltreatment

Abstract

Background

The absence of knowledge regarding the quality of child maltreatment administrative data poses significant risks to the validity of field research findings, hinders research reproducibility, and increases the potential for overgeneralized results.

Objective

To provide a scientific framework for systematically and computationally assessing the quality of child maltreatment administrative data.

Participants and setting

To test the data quality examination approach, data from a child maltreatment database subscription service utilized between 2000 and 2023, including maltreatment records from 393 child and victim advocacy programs across 45 states and Washington, D.C. in U.S.

Methods

Four core dimensions of data quality (accuracy, consistency, completeness and timeliness) were measured through quantified data quality metrics. Further statistical analyses examined the relationship between data quality and geographic locations.

Results

Moderate to good overall data quality was found, with significant variation across agencies. Some exhibited exceptionally high data quality, and geographic location was associated with variation in data quality.

Conclusions

This study is the first to comprehensively evaluate the data quality of a nationwide database on child maltreatment, offering a valuable scientific reference for future research applying this framework to assess the quality of administrative data.