Figure. Adjusted odds ratios of in-hospital death by race/ethnicity relative to non-Hispanic whites (black boxes), which is represented by the horizontal line. The vertical lines indicate the 95% CI for each odds ratio.
Ramirez M, Chang DC, Bickler SW. Pediatric Injury Outcomes in Racial/Ethnic Minorities in CaliforniaDiversity May Reduce Disparity. JAMA Surg. 2013;148(1):76–80. doi:10.1001/2013.jamasurg.3
Author Affiliations: Department of Surgery, University of California, San Diego, La Jolla.
Hypothesis Differences in health outcomes are well documented in adult racial/ethnic minorities. We hypothesize that similar differences exist in pediatric racial/ethnic minorities because their care is a function of their parents' access. We investigated this issue by examining pediatric injury outcomes in California.
Design Retrospective analysis of the California Office of Statewide Health Planning and Development hospital discharge database.
Setting Sample of all California hospitalized patients.
Patients In a sample of patients aged 18 years or younger at admission, injury patients were defined as having International Classification of Diseases, Ninth Revision, primary diagnosis codes between 800 and 959, with certain exclusions, from January 1, 1999, through December 31, 2010.
Main Outcome Measures Adjusted risk of in-hospital death, controlling for age, sex, injury severity measured by survival risk ratio, Charlson comorbidity index, insurance status, admission year, teaching hospital status, and mechanism of injury.
Results A total of 47 000 pediatric patients were identified. Bivariate analysis showed a significant difference in mortality by race/ethnicity among non-Hispanic whites (0.8%), blacks (1.2%), Hispanics (1.1%), Asians (1.2%), and American Indians/others (0.6%) (P = .01). However, with the exception of Asians (odds ratio, 0.32; 95% CI, 0.11-0.90), adjusted odds ratios of death relative to non-Hispanic whites for blacks (1.33; 0.71-2.46), Hispanics (1.06; 0.71-1.58), and American Indians/others (0.60; 0.17-2.10) showed no significant differences.
Conclusions Unlike previous studies that have shown that adult racial/ethnic minorities (age, 18-64 years) have higher mortality relative to non-Hispanic whites, our study demonstrated no significant racial/ethnic differences among pediatric patients with injuries. It may be that differential access does not exist for children. In addition, it may also be possible that the diversity in California leads to culturally competent care and such care has been reported to improve patient outcomes.
Disparities in health outcomes in adult racial/ethnic minorities have been shown extensively in the trauma literature.1- 6 Whether this exists in pediatric patients is less clear. We hypothesized that similar differences exist in pediatric racial/ethnic minorities because their care is a function of their parents' access to health care. The goal in this study was to examine this issue in a statewide database in California.
We performed a retrospective analysis of the California Office of Statewide Health Planning and Development (OSHPD) hospital discharge database from 1999 to 2010. The OSHPD is 1 of 13 departments within the California Health and Human Services Agency; it reports health care outcomes for specific procedures and medical conditions from California-licensed hospitals and health care practitioners. Our study reflects publicly available data. We chose to include data beginning in 1999 to capture a more complete data set.7
Injury patients aged 18 years or younger at admission were identified by International Classification of Diseases, Ninth Revision (ICD-9), primary diagnosis codes between 800 and 959, excluding 905-909 (late effects of injury), 930-939 (foreign body), 940-949 (burn), and 958 (early complications of trauma).
The primary outcome variable for our multivariate analysis was in-hospital death. Our primary independent variable was race/ethnicity defined as non-Hispanic whites, blacks, Hispanics, Asians, and American Indians/others. Our covariates included age, sex, admission year, insurance status, injury severity measured by survival risk ratios (SRRs), comorbidities determined using the Charlson comorbidity index, teaching hospital status, and mechanism of injury. Children were classified into 3 age groups: 0 to younger than 1, 1 to 12, and 13 to 18 years, with the youngest age category as our reference group. The effect of admission years (2000-2010) on in-hospital death was compared with in-hospital deaths in 1999. Our reference group for insurance status included patients with private insurance coverage, which was compared with patients with Medi-Cal, indigent and other government coverage, Medicare, worker's compensation, self-pay, and other payer. Injury severity was measured by SRRs per ICD-9 Injury Severity Score methods based on the diagnosis code of the patients.8 Traditional SRRs were calculated. Comorbidities were adjusted for using the Charlson comorbidity index, which was calculated per the Deyo et al adaptation for administrative data sets as described by Romano et al.9Teaching hospitals were defined as those having a surgical residency program. In this study, mechanism of injury was defined as motor vehicle crashes, handguns, falls, and other.
We performed a multivariate analysis on the data using commercially available software (Stata special edition 11.2; StataCorp). Statistical significance was defined as P < .05. This study and its publicly accessible database were deemed exempt from review by our institutional review board.
Of 47 000 pediatric patients, the overall in-hospital death rate was 1.0%. Table 1 reports the demographic information for the study population. Most patients were Hispanic (51.4%) and non-Hispanic white (37.0%); black (7.0%), Asian (2.5%), and American Indian/other (2.1%) composed the rest of our patient sample. Slightly more than two-thirds of the patients were male, and nearly all patients had private or Medi-Cal insurance coverage. Most of the population (78.5%) received care at nonteaching hospitals.
Table 2 reports a comparison of patient demographics by race/ethnicity. There was a significant difference in mortality rates, ranging from 0.6% to 1.2%. There were significantly more males in all groups, most notably in the black group. There were also more patients with private insurance coverage among non-Hispanic whites and Asians, whereas more than half of blacks and Hispanics had Medi-Cal insurance coverage. There were significant differences in mechanism of injury between races/ethnicities; blacks had higher rates of injuries due to handguns (4.3%) and Asians were more likely to have fallen (46.2%).
Our multivariate analysis showed no association between mortality and race/ethnicity for most groups compared with non-Hispanic whites (Figure). The only exception was with Asians, who demonstrated a 68% decreased risk of mortality compared with non-Hispanic whites. There was an association between age and mortality, with increased survival in the older age groups (1-12 years and 13-18 years) compared with the youngest age group (0-<1 year). In addition, there was a statistically significant increased risk of death in patients who were in the insurance category of other payer or were involved in motor vehicle crashes (odds ratios, 4.53 and 1.61, respectively) (Table 3).
For all groups with the exception of Asians, our study found no significant difference in mortality with respect to race/ethnicity. Although our unadjusted analysis showed a significant difference in mortality between racial/ethnic groups, this significance disappeared on adjusted analysis. Our work is in contrast to disparity studies1- 6 in adult racial/ethnic minorities, which showed higher mortality in nonwhites. One possible reason for the difference between children and adults is that in adults, there is differential access to insurance based on their financial, health, and immigration status; this differential access does not exist for children. Children have many avenues for insurance coverage in California via Medi-Cal,10 the Healthy Families11 program, and various other government programs.
Some studies in children have shown the disparity in trauma outcomes among racial/ethnic minorities that we see in adults.12,13 Using the National Trauma Databank, Hakmeh et al12 showed increased mortality in blacks and Hispanics compared with non-Hispanic whites. However, our study focused on the pediatric outcomes in California and, interestingly, we did not see any significant differences in outcomes based on race/ethnicity with the exception of Asians. The meaningfulness of this difference in Asians is unclear. Our study is consistent with that of Salim et al,14 which showed that race did not appear to be an independent risk factor for mortality in a large trauma center in southern California.
The explanation for our observation that there is no significant disparity in care among children of diverse races/ethnicities is likely multifactorial. One possible reason for the difference between our study vs studies conducted nationally may be our setting. California uniquely consists of a growing diversity of racial and ethnic groups. We speculate that exposure to such a population rich in diversity may allow us to provide culturally competent care for our youth, resulting in similar rates of survival for all groups. The link between culturally competent care and patient outcomes has been reported. A review15 assessing the effectiveness of cultural competence training for health professions in community-based rehabilitation showed positive outcomes for most of the programs that used cultural competency training. Beck and Gordon16(p1085) addressed the relationship between cultural competence and outcomes in the mental health field, stating, “Engaging the patient in their own path to recovery or well-being improves engagement in, and adherence to, the treatment plan, and ultimately improves outcomes.” Indeed, sites with cultural competency training had better patient satisfaction and increased adherence compared with sites with less training.17 Furthermore, cultural competency training continues to benefit the patient beyond the care of the culturally competent provider. For adolescent patients, Soriano et al18(p175) reported, “Physicians, by recognizing the problems that may occur in an adolescent attempting to gain competence in two cultures, may be able to provide appropriate interventions and referral services.”
In addition, it may be that the expansion of health insurance to low- and middle-income families has had a beneficial effect for injured children. Beginning in 2007 via the California Health Initiative, California expanded health care coverage to those who are not eligible for the Medi-Cal, Healthy Families, or Access for Infants and Mothers program.19 With insurance coverage expanding to care for more of our youth, hospitalizations associated with ambulatory care–sensitive conditions were reduced after implementation of the California Health Initiative.20
It appears that Hispanic children may be overrepresented in trauma. In the current study, 51.4% of the injured children were Hispanic. However, this percentage is a reflection of the change in demographics in California. Based on the 2010 census, the number of Hispanic children increased by 17%, which accounts for why 51% of Californians younger than 18 years are Hispanic.21
The strengths of our findings include use of the OSHPD database, a nontrauma database, which allowed us to include data from trauma and nontrauma centers during 12 years. A large sample size of 47 000 individuals provides comprehensive and generalizable data. However, our study has several limitations. Because the OSHPD database is a nontrauma registry, our ability to adjust for injury severity with the commonly used injury severity scores was limited. This anatomic scoring system yields an overall score for patients with multiple injuries. Instead, we used the SRR, which is a calculated number based on ICD-9 diagnosis codes that carry different risks of death. Because this was a large administrative database, there may be errors in data reporting; however, we believe that these errors are randomly distributed. Data pertaining to blacks and American Indians were lacking, providing less than 10% of the data. Such a limited sample size may mask significant results. In addition, the impact of immigration status is unknown in our study.
In closing, we believe that the favorable outcomes in the care of pediatric racial/ethnic minorities may reflect both culturally competent care and California's efforts to offer insurance coverage to more children. However, our work is also a call to action. We see disparities in trauma outcomes appear when these children grow into adulthood. Continued advances in the realm of health insurance coverage are needed as well as interventions during the transition to adulthood. Other states may be interested in replicating the study in their own populations. Further studies are needed to focus on issues between diversity, cultural competency, and disparity. California would be an ideal population in which to compare the magnitude of disparity within regions with varying levels of diversity to determine whether diversity reduces disparity.
Correspondence: Michelle Ramirez, BA, Department of Surgery, University of California, San Diego, 9152 Regents Rd, Unit D, La Jolla, CA (email@example.com).
Accepted for Publication: June 6, 2012.
Published Online: September 17, 2012. doi:10.1001/2013.jamasurg.3
Author Contributions:Study concept and design: Chang and Bickler. Acquisition of data: Ramirez and Chang. Analysis and interpretation of data: Ramirez, Chang, and Bickler. Drafting of the manuscript: Ramirez, Chang, and Bickler. Critical revision of the manuscript for important intellectual content: Ramirez, Chang, and Bickler. Statistical analysis: Ramirez and Chang. Administrative, technical, and material support: Ramirez and Chang. Study supervision: Ramirez, Chang, and Bickler.
Financial Disclosure: None reported.
Funding/Support: This project was supported by the Hispanic Center of Excellence at University of California, San Diego (UCSD) School of Medicine, UCSD Program in Medical Education, and the UCSD Department of Surgery.
Previous Presentation: This paper was presented at the 83rd Annual Meeting of the Pacific Coast Surgical Association; February 18, 2012; Napa Valley, California, and is published after peer review and revision.