Big Data

---

Race, Ethnicity, and Infections in the Hospitalized Child with Cancer: A Role for Big Data Beth Savage, PhD candidate, CPNP, CPON, Rutgers Cancer Institute of New Jersey Charlotte Thomas-Hawkins, PhD, RN, Rutgers University School of Nursing

Basis of Inquiry

•

• •

Increased survivorship in childhood cancer has been associated with belonging to the nonHispanic White race/ethnicity group. Yet, very little research has been done to determine if there are similar disparities in the experience of infectious complications of cancer treatment (Bhatia, 2011). Big data offers Pediatric Hematology Oncology nurses a practical way to explore this complex inquiry. Big data has not been widely used by pediatric hematology oncology nurses to date. In fact, a search for the terms “Big Data”, “Data Science”, “Secondary Data Analysis”, and “Publicly Available Data” in the Journal of Pediatric Oncology Nursing failed to find any research studies applying big data to the science of pediatric hematology oncology nursing.

Methods The following statistical analyses were performed:

•

Descriptive statistics were computed using frequencies and means for categorical and continuous variables, respectively.

•

Chi square and t-test were performed to analyze bivariate relationships between predictor and dependent variables.

• •

Those predictor variables found to be significant (p<.05) in bivariate analyses were included in multivariable logistic regression models. Two separate regression analyses were performed to determine the odds of a hospitalized child experiencing an infection; and the odds of a child dying during a hospitalization in which an infection is documented.

Results Learning Outcomes

• The learner will be able to identify the association between race and infectious complications in children with cancer.

•

The two adjusted logistical regression models included: age, severity of clinical condition, race/ethnicity, and household median income by zip code.

• The learner will be able to describe to the role big

Sample This research study was a secondary analysis of the 2012 HCUP Kids' Inpatient Database (KID). The KID database is a project of the Agency for Healthcare Research and Quality and represents 80% of all inpatient pediatric hospitalizations during the year 2012.

•

The sample size was 77,299 pediatric cancer hospitalizations. Of these hospitalizations, 31,732 cases documented infection in the discharge record.

www.cinj.org

The results of the adjusted models are in Figures 2 and 3. These show forest plots of the odds ratios (OR) (red dot) with 95% confidence intervals (black lines) for factors associated with the presence of infection in the sample of cancer hospitalizations (Figure 2), and death in those cases with documented infection (Figure 3). The reference category is (White race/ethnicity) represented by the blue line at 1.0. Values fully to the right of the referent line depict factors associated with greater odds of infection while those to the left show decreased odds. Those factors with 95% CI crossing the red line are non-significant.

 Over fifty percent of children hospitalized with cancer identified as White race. White children represented a greater percentage of the cases in which there were common infectious complications (Figure 1).

 Association of race with infection in children hospitalized with cancer:

•

Compared to White children, Hispanic children have an increased likelihood of infection (OR=1.09, p<.001) (Figure 2).

•

Compared to White children, African American children have decreased likelihood of infection (OR=.850, p<.001) (Figure 2).

 Association of race/ethnicity with death in children with documented infection: • Although African American children were at decreased risk of a documented infection, the risk of death was significantly higher in these children compared to White children (OR=1.57, p=.002) when infection was present (Figure 3). • Hispanic children (OR=1.42, p=.002) and Asian children (OR=1.58, p=.025) with documented infection are also more likely to die compared to White children (Figure 3).

data has in shaping future pediatric hematology oncology nursing initiatives.

•

•

Conclusions

 The use of big data in this study provides evidence of the unequal risks facing children belonging to races and ethnicities other than the non-Hispanic White group. These findings justify future studies to determine why this is occurring and how nurses can intervene on behalf of these children.

References Figure 1 provides the distribution, in percentages (x-axis) of common infectious complications (y-axis) by race (see legend).

Agency for Healthcare Research and Quality. (2018). Kids' Inpatient Database (KID). Retrieved from https://www.hcupus.ahrq.gov/ db/nation/kid/kiddbdocumentation.jsp Bhatia, S. (2011). Disparities in Cancer Outcomes: Lessons Learned From Children With Cancer. Pediatric Blood and Cancer, 56(6), 994-1002.