Factors Influencing Variations in Hospitalization for Diabetes with Hypoglycemia

Abstract

Many studies have explored risk factors associated with Hypoglycemia (HG) and examined the variation in healthcare utilization among HG patients. However, most of these studies failed to integrate a comprehensive list of personal risk factors in their investigations. This empirical study employed the Behavioral Model (BM) of health care utilization as a framework to investigate diabetes' hospitalizations with HG. The national inpatient sample with all non-pregnant adult patients admitted to hospitals' emergency departments and diagnosed with HG from 2012 to 2014 was used. Personal factors were grouped as predictors of the length of stay and the total charges incurred for hospitalization. High-risk profiles of hospitalized HG patients were identified. The analysis shows the need for care factors are the most influential predictors for lengthy hospitalization. The predisposing factors have a limited influence, while enabling factors influence the variation in hospital total charges. The presence of renal disease and diabetes mellitus (DM) complications played a key role in predicting hospital utilization. Furthermore, age, socio-economic status (SES), and the geographical location of the patients were also found to be vital factors in determining the variability in utilization among HG patients. Findings provide practical applications for targeting the high-risk HG patients for interventions.

Keywords: costs of care; high-risk profile of hospitalized diabetes; hospital utilization; hypoglycemia; predictors of hospital length of stay.

Figure 1

DTREG tree of the predictors for LOS.

Figure 2

Relative importance of statistically significant variables in explaining the variance in the overall LOS.

Figure 3

DTREG tree of the predictors for the total charge.

Figure 4

Relative importance of statistically. significant predictor variables in explaining the variation in total charge.