Strength and Limitations
This study has several strengths and limitations. Strengths of this study include a sample size, study design and time of data collection. The HCUP NIS is designed to produce regional and national estimates of hospitalization utilization and outcomes. The analyses are conducted using weighted estimates that represent more than 35 million annual hospitalizations nationally. The large sample size allowed for robust analysis of rare exposures like PDM and outcomes like stillbirth in specific racial/ethnic groups at a national level using a standardized dataset. Use of this data reduces selection bias and increases precision of our analyses. Further, our findings are based on data from 2002 to 2017, more recent and comprehensive than previous national studies.7,10,47
Limitations of this study include the data source. HCUP NIS is limited to an inpatient population and does not account for women who delivered outside of the hospital. Additionally, results of our studies are limited to the analysis of only three racial/ethnic groups despite differences in prevalence of pregnancy-associated diabetes in other race/ethnic groups including Native Americans and NH-Asians.7,31,33
Another limitation of this study is variation in data reporting for PDM and GDM. Despite efforts to standardized documentation, there are differences in reporting across data sources likely due to changes in diagnosis criteria over time, coding, and health care provider experience.48 The extent to which variation in reporting is likely to influence our results is not clear. However, a study that compared birth certificate data to other government collected data found the HCUP NIS survey to be a more acceptable source.48
Furthermore, our study is limited by the inability to account for sociodemographic characteristics such as education level, employment, and marital status. Additionally, we could not account for factors such as physical activity, diet, hemoglobin a1c level and diabetes type which are associated with pregnancy-associated diabetes and/or stillbirth because of NIS HCUP data limitations.49,50 Lastly, given the design of this study, a causal relationship between PDM and stillbirth could not be determined.