Pediatric Emergency Department Recidivism: Demographic Characteristics and Diagnostic Predictors
from Caring For Our Future, Fall 2005
Introduction
Interest in recidivism was generated by The Children’s Hospital’s participation with the Emergency Nurses Association LUNAR II project (Learning and Using New Approaches to Research) in 1998. The purpose of this national multi-site study was to describe the characteristics of individuals who make repeated visits (recidivists) to the Emergency Department. The LUNAR II study identified demographic data, illness patterns, health service utilization patterns, socio-cultural and economic characteristics and the relationship among these variables to describe the factors associated with recidivism. ED overcrowding continues to be a national problem. The volume of ED visits has steadily increased over the last several decades because EDs have assumed many of the responsibilities of primary-care facilities in addition to emergency care. Studies have estimated that two-thirds of all ED visits may be inappropriate.
Approximately 60,000 children seek emergency care each day in the United States . (Sharma, et al, 2000). This population represents a unique group of patients whose medical needs are quite different from their adult ED patient counterparts. For example, a fever in this age group may be more ominous, and a respiratory illness may give rise to greater morbidity. Common themes identified as contributing to ED misuse and overcrowding include parents’ failure to understand the requirements for referral as well as misinterpretation of severity of illnesses. Health disparities for pediatric patients have been identified as barriers to care that impact emergency service utilization. (Howard, 2005; Blank, et al, 2005; Doobinin, et al, 2003) Efforts to determine why patients with certain diagnoses are at high risk for return can help further define the relationship between recidivism and the nature of pediatric emergency health-care delivery.
Purpose
The research questions included:
- What are the demographic characteristics of frequent users of pediatric emergency departments?
- Are certain demographic and/or diagnostic categories predictive of pediatric recidivism?
- How does the rate of recidivism at our hospital compare with other similar pediatric facilities?
Methods/Design
The hospital decision support system (DSS) was used to first identify a base population of patients who were treated in the Emergency Department during a one-week, non-holiday study period. Two readmission indicators were developed for the purpose of establishing the study populations. The first readmission indicator flagged each ED patient from the base population who had an ED visit during the previous three months as previously done with the LUNAR II study. The second readmission indicator flagged a subset of patients from the base population who had an ED visit during the previous 48 hours. Forty-eight hour return visits were selected by a national benchmarking initiative (ORYX initiative of the Joint Commission on Accreditation of Healthcare Organizations) to facilitate comparisons among other similar pediatric institutions. Table 2 summarizes these findings.
Analysis
Descriptive statistics were used to compare the demographic and diagnostic profile of the study populations. Logistic regression techniques were utilized to assess the association of risk factors with the outcome variable of return visits. Univariate logistic analysis was used to screen the risk factors associated with the outcome (return visit). The risk factors with p values less than or equal to 0.1 were entered into the multivariate logistic regression to examine their association with return visits while adjusting for the effects of the other risk factors in the model. Risk factors were considered significantly associated with the outcome if the p values were less than or equal to 0.05. The associations are presented as odds ratios (OR), and the confidence interval was set at 95 percent.
In an attempt to limit the number of comparisons made during the regression analysis, it was often necessary to take the raw demographic and diagnostic data and create groupings. For example, the American Academy of Pediatrics age groupings were used to conduct the univariate and multivariate logistic analysis on patient age. International Classification of Diseases – 9th Revision (ICD-9) diagnosis groupings were used for the analysis on patient principle diagnosis. The three most frequently occurring principle diagnoses in each ICD-9 diagnosis grouping are displayed in Table 2. In cases where there were numerous groups, an “Other” group was created. This “Other” grouping represented the most infrequent demographic characteristics or diagnostic categories found in the sample as well as unreported /unknown data.
Findings
Demographic Characteristics
The data set consisted of all children who came to the ED during the one-week period (N =932). Approximately 25 percent of these patients had at least one visit during the preceding three months (N= 237). Four percent of the study-week patients had 48-hour return visits (N=38). Demographic characteristics are summarized in Table 1 and include comparison data from the LUNAR II study.
Study findings related to the subset population of 48-hour returns did not confirm the disparities frequently described in the literature that demonstrate socioeconomic distress. For example, chart reviews revealed that 70 percent of this subset had a primary-care provider and insurance coverage. Data from this study suggested that the traditional determinants of health-care service utilization such as age, health insurance, and the seriousness of symptoms are not reliable predictors of return visits. The most frequent reason for ED utilization was lack of availability of the PCP at the time of visit. PCP referrals to the emergency department were made for only 20 percent of the sample, while family member insistence accounted for 60 percent of the 48-hour return visit population.
Predictability of Return Visits
Diagnostic categories and demographic variables were examined to determine predictability of return visits. Regression techniques were only possible with the three-month study population due to the small sample size of the 48-hour return subset. The risk factors included age, gender, race, insurance coverage, diagnosis, primary-care provider and day/time of ED visits, which are the variables representative of health disparities frequently described in the literature.
Univariate logistic analysis was used to screen the risk factors associated with return visits. The risk factors of age, race, principle diagnosis, visit day and gender were significantly associated with the outcome at a significance level of 0.1; while the variables of insurance status, admit times and PCP status were not significant. Those factors with p-values less than or equal to 0.1 were entered into the multivariate logistic regression. The variables of age, race and principle diagnosis group remained significantly associated with return visits to the ED (p<0.05), while the variables of visit day and gender were no longer significant. Thus, visit day and gender were dropped out of the multivariate model. The results from the multivariate logistic regression are summarized in Table 2.
For the age variable, patients aged 1 year or older were 39 percent to 47 percent as likely to return to the ED within three months as those patients less than 1 year old (reference group). In other words, patients younger than 1 year old were more than twice as likely to return to the ED within t hree months compared to other age groups when the other risk factors were adjusted in the model.
The odds of ED return visits for Caucasian patients were 39 percent less likely to have an ED return visit when compared to the Hispanic population (reference group). African-American patients had no statistically significant difference from the Hispanic patients in their probability of ED return visits (p=0.71). The other race categories of Asian, Native American, and unknown were nearly 50 percent less likely to return to the ED compared to Hispanic patients.
Statistically significant findings also were obtained in the analysis of principle diagnosis. Patients with nervous-system and sense-organ diseases (see Table 3 for examples) were 1.86 times more likely to return to the ED within three months than patients with respiratory-system diseases (reference group). Patients with digestive-system diseases were 48 percent less likely to return to the ED than the reference group.
These findings confirm that some demographic and diagnostic predictability of return visits to the Emergency Department within three months is indeed evident for pediatric patients. Age may be considered a predictable variable of recidivism, especially for patients less than 1 year old. Health disparities in terms of race were significant, and indicated that Hispanic and African-America n populations were twice as likely to be recidivists. Finally, the principle diagnosis was a third predictor of recidivism. Patients with a principle diagnosis that fell under the broad category of nervous-system and sense-organ diseases have a much higher chance of returning to the Emergency Department than any of the other disease categories.
Comparison of Pediatric Recidivists
Our hospital reports 48-hour return visit data to the BENCHmarking Effort for Networking Children’s Hospitals (BENCH) supported by Medical Management Planning, Inc. (MMP) on a quarterly basis. BENCH is a pediatric-specific benchmarking and best-practices effort that, since 1992, has facilitated the comparison of unblinded data among its member children’s hospitals for the purposes of improving the care delivered to pediatric patients and their families. The institutions participating in the BENCH initiative define recidivism as an unscheduled return to the ED within 48 hours with the same chief complaint or reason for visit. This definition was utilized in our study to insure a more accurate comparison with similar pediatric hospitals. Data revealed that our hospital has an average rate of recidivism of 1.0 percent to 1.86 percent, which is significantly less than reported rates in the literature of 13 percent to 18 percent. (Blank, et al, 2005; Doobinin, et al, 2003; Cardin, et al, 2003). Additional comparisons to BENCH-reporting facilities indicated that the maximum rate of recidivism for the study quarter was 3.79 percent, with our institution reporting 1.49 percent.
Significance/Nursing Implications
The literature describing return pediatric ED visits is limited and largely concerns validating return visits as a medical quality assurance tool. These studies provide some insight into why some patients return, but generally do not offer methods for identifying high-risk patients before they return. Our study provided an understanding of parental illness behavior and the influence of the health-care system on a parent’s utilization of health-care services to identify high-risk pediatric populations.
Any strategy to address frequent use of the ED requires an understanding of patients with health disparities (Alessandrini, et al, 2004). Our study allowed investigators to evaluate assumptions about ED frequent users and to identify the link between demographic characteristics, diagnostic predictors, and return visits.
The description of these relationships provides emergency nurses with valuable information to create preventive strategies focused on health disparity profiles.
The predictability of return visits for certain diagnostic categories further delineated the issue of recidivism. These data illuminated specific diseases warranting focused attention such as nervous-system and sense-organ diseases, digestive-system diseases, and infectious and parasitic diseases for children less than 1 year old. These populations can be considered to be at greater risk for return visits. Patients under the age of 1 year with nervous-system/sense-organ diseases should receive focused attention to discharge instructions, follow-up arrangements including specific symptoms that would require return, and referral to primary-care providers or centers providing non-urgent care.
Future studies might explore hospital-wide interinterventions pertaining to the structure and the process of care. Screening return ED populations for high-frequency diagnoses may reveal under-recognized target groups for specific cross-disciplinary preventive interventions. Future studies also might examine the relationships associated with recidivism such as family satisfaction, compliance with provider follow-up recommendations and illness severity.
References
Alessandrini, E.A., Lavelle, J.M., Grenfell, S.M., Jacobstein, C.R., Shaw, K.N. (2004). Return visits to a pediatric emergency department. Pediatric Emergency Care, 20 (3): 166-171.
Bayley, EW, MacLean, SL, Desy, P., McMahon, M. (2004). ENA’s Delphi Study on National Research Priorities for Emergency Nurses in the United States . Journal of Emergency Nursing, 30(1): 12-21.
Blank, F.S., Blank, J., Li, H., Henneman, P., Smithline, H., Santoro, J., Provost, D., Maynard, A. (2005). A descriptive study of heavy emergency department users at an academic emergency department reveals heavy ED users have better access to care than average users. Journal of Emergency Nursing, 31(2): 139-144.
Cardin, S., Afilalo, M., Lang, E., Collet, J.P., Colacone, A., Tselios, C., Dankoff, J., Guttman, A. (2003). Interventions to decrease emergency department crowding: does it have an effect on return visits and hospital readmissions? Annals of Emergency Medicine, 41 (2): 173-185.
Doobinin, K.A., Heidt-Davis, P.E., Gross, R.K., Isaacman, D.J. (2003). Non-urgent pediatric emergency department visits: care-seeking behavior and parental knowledge of insurance. Pediatric Emergency Care, 19(1): 10-14.
Howard, P.K. (2005). Overcrowding: Not just an emergency department issue. Journal of Emergency Nursing, 31(3): 227-228.
Kelly, A.M., Chirnside, A.M., Curry, C.H. (1993, Aug 11). An analysis of unscheduled return visits to an urban emergency department. New Zealand Medical Journal, 106(961): 334-346.
Prah-Fuger, J., Richter, C., Spitznagel, E., Lewis, L. (2004). Analysis of costs, length of stay, and utilization of emergency department services by frequent users: implications for health policy. Academy of Emergency Medicine , 2(12): 1311-1317.
Sharma, V., Simon , S.D. , Bakewell, B.A., Ellerbeck, E.F., Fox, M.H., Wallace, D.D. (2000). Factors influencing infant visits to emergency department. Pediatrics, 06(5): 1031-1039.
Sun, B., Burstin, H., Brennan, T. (2003). Predictors and outcomes of frequent emergency department users. Academy of Emergency Medicine , 10(4): 320-328.
Wetta-Hall, R., Ablah, E., Dismuke, S., Molgaard, C., Fredrickson, D., Berry, M. (2005). Emergency department use by people on low income. Emergency Nurse, 13 (3): 12-18.
