Predictors of Outcomes in Cerebellar Stroke: A Retrospective Cohort Study From the National Inpatient Sample Data.

ABSTRACT

Cerebellar strokes have high morbidity and mortality due to bleeding or edema, leading to increased pressure in the posterior fossa. This retrospective cohort study analyzed three outcomes following a cerebellar stroke in-hospital mortality, length of hospital stay, and total hospitalization costs. It uses data from the National Inpatient Sample (NIS) and aims to identify the predictors of outcomes in cerebellar stroke patients, including 464,324 patients, 18 years of age and older, hospitalized between 2010 and 2015 in US hospitals with cerebellar strokes. In our study, for every decade age increased beyond 59 years, there was a significant increase in mortality; those aged 80+ years had 5.65 odds of mortality (95% CI 5.32-6.00; P < 0.0001). Significant differences in patient characteristics were observed between patients who survived to discharge and those who did not, including older age (77.4 vs. 70.3 years; P < 0.0001), female sex (58% vs. 52%; P < 0.0001), and being transferred from another healthcare facility (17% vs. 10%; P < 0.0001). Patients admitted directly rather than through the emergency department were more likely to die (29% vs. 16%; P < 0.0001). The mortality rate was lower for blacks (OR 0.75; P < 0.0001), Hispanics (OR 0.91; P = 0.005), and Asians (OR 0.89; P = 0.03), as compared to the white population, for females in comparison to males, and geographically, in all other areas (Midwest, South, and West) in contrast to the Northeast. Cerebellar stroke incidence and high mortality were seen in the traditional stroke belt. Mortality is also affected by the severity of the disease and increases with the Charlson Comorbidity Index (CCI), All Patient Refined Diagnosis Related Groups (APR-DRG) scores, and indirectly by place of receiving care, length of stay (LOS), cost of stay, type of insurance, and emergency department admissions. LOS increased with age, in males in the Northeast, and was less in whites compared to other races. Trend analysis showed a decrease in LOS and costs from 2010 to 2015. Increased costs were seen in non-whites, males, higher household income based on zip code, being covered under Medicaid, transfers, CCI ≥ 5, and discharges in the western US. Median household income based on the patient's zip code was well-balanced between those who lived and those who died (P = 0.091). However, payers were not evenly distributed between the two groups (P < 0.0001 for the overall comparison). A higher proportion of discharges associated with in-hospital mortality were covered under Medicare (70% vs. 65% in the died vs. lived groups, respectively). Fewer discharges were associated with death if they were covered by commercial insurance or paid for out-of-pocket (15% vs. 19% for commercial insurance and 3% vs. 5% for out-of-pocket). In-hospital mortality was associated with a longer length of hospital stay (5.6 days vs. 4.5 days; P < 0.0001) and higher costs ($16,815 vs. $11,859; P < 0.0001). Variables that were significantly associated with lower total costs were older age, having commercial insurance, paying out-of-pocket or other payers, not being admitted through the emergency department, having a lower comorbidity index (CCI = 1-2), and being discharged from a hospital that was small- or medium-sized, located in the Midwest or South, and/or was non-teaching (rural or urban).