Postacute Care after Pediatric Hospitalizations for a Primary Mental Health Condition.

OBJECTIVES: To determine the proportion of US children hospitalized for a primary mental health condition who are discharged to postacute care (PAC); whether PAC discharge is associated with demographic, clinical, and hospital characteristics; and whether PAC use varies by state. STUDY DESIGN: Retrospective cohort study of a nationally representative sample of US acute care hospitalizations for children ages 2-20 years with a primary mental health diagnosis, using the 2009 and 2012 Kids' Inpatient Databases. Discharge to PAC was used as a proxy for transfer to an inpatient mental health facility. We derived adjusted logistic regression models to assess the association of patient and hospital characteristics with discharge to PAC. RESULTS: In 2012, 14.7% of hospitalized children (n = 248 359) had a primary mental health diagnosis. Among these, 72% (n = 178 214) had bipolar disorder, depression, or psychosis, of whom 4.9% (n = 8696) were discharged to PAC. The strongest predictors of PAC discharge were homicidal ideation (aOR, 24.9; 96% CI, 4.1-150.4), suicide and self-injury (aOR, 15.1; 95% CI, 11.7-19.4), and substance abuse-related medical illness (aOR, 5.0; 95% CI, 4.5-5.6). PAC use varied widely by state, ranging from 2.2% to 36.3%. CONCLUSIONS: The majority of children hospitalized primarily for a mood disorder or psychosis were not discharged to PAC, and safety-related conditions were the primary drivers of the relatively few PAC discharges. There was substantial state-to-state variation. Target areas for quality improvement include improving access to PAC for children hospitalized for mood disorders or psychosis and equitable allocation of appropriate PAC resources across states.

Readmissions after Pediatric Hospitalization for Suicide Ideation and Suicide Attempt.

OBJECTIVE: To inform resource allocation toward a continuum of care for youth at risk of suicide, we examined unplanned 30-day readmissions after pediatric hospitalization for either suicide ideation (SI) or suicide attempt (SA). METHODS: We conducted a retrospective cohort study of a nationally representative sample of 133,516 hospitalizations for SI or SA among 6- to 17-year-olds to determine prevalence, risk factors, and characteristics of 30-day readmissions using the 2013 and 2014 Nationwide Readmissions Dataset (NRD). Risk factors for readmission were modeled using logistic regression. RESULTS: We identified 95,354 hospitalizations for SI and 38,162 hospitalizations for SA. Readmission rates within 30 days were 8.5% for SI and SA hospitalizations. Among 30-day readmissions, more than one-third (34.1%) occurred within 7 days. Among patients with any 30-day readmission, 11% had more than one readmission within 30 days. The strongest risk factors for readmission were SI or SA hospitalization in the 30 days preceding the index SI/SA hospitalization (adjusted odds ratio [AOR]: 3.14, 95% CI: 2.73-3.61) and hospitalization for other indications in the previous 30 days (AOR: 3.18, 95% CI: 2.67-3.78). Among readmissions, 94.5% were for a psychiatric condition and 63.4% had a diagnosis of SI or SA. CONCLUSIONS: Quality improvement interventions to reduce unplanned 30-day readmissions among children hospitalized for SI or SA should focus on children with a recent prior hospitalization and should be targeted to the first week following hospital discharge. FUNDING: Dr. Zima received funding from the Behavioral Health Centers of Excellence for California (SB852).

Mental Health Conditions and Unplanned Hospital Readmissions in Children.

OBJECTIVE: Mental health conditions (MHCs) are prevalent among hospitalized children and could influence the success of hospital discharge. We assessed the relationship between MHCs and 30-day readmissions. METHODS: This retrospective, cross-sectional study of the 2013 Nationwide Readmissions Database included 512,997 hospitalizations of patients ages 3 to 21 years for the 10 medical and 10 procedure conditions with the highest number of 30-day readmissions. MHCs were identified by using the International Classification of Diseases, 9th Revision-Clinical Modification codes. We derived logistic regression models to measure the associations between MHC and 30-day, all-cause, unplanned readmissions, adjusting for demographic, clinical, and hospital characteristics. RESULTS: An MHC was present in 17.5% of medical and 13.1% of procedure index hospitalizations. Readmission rates were 17.0% and 6.2% for medical and procedure hospitalizations, respectively. In the multivariable analysis, compared with hospitalizations with no MHC, hospitalizations with MHCs had higher odds of readmission for medical admissions (adjusted odds ratio [AOR], 1.23; 95% confidence interval [CI], 1.19-1.26] and procedure admissions (AOR, 1.24; 95% CI, 1.15-1.33). Three types of MHCs were associated with higher odds of readmission for both medical and procedure hospitalizations: depression (medical AOR, 1.57; 95% CI, 1.49-1.66; procedure AOR, 1.39; 95% CI, 1.17-1.65), substance abuse (medical AOR, 1.24; 95% CI, 1.18-1.30; procedure AOR, 1.26; 95% CI, 1.11-1.43), and multiple MHCs (medical AOR, 1.43; 95% CI, 1.37-1.50; procedure AOR, 1.26; 95% CI, 1.11-1.44). CONCLUSIONS: MHCs are associated with a higher likelihood of hospital readmission in children admitted for medical conditions and procedures. Understanding the influence of MHCs on readmissions could guide strategic planning to reduce unplanned readmissions for children with cooccurring physical and mental health conditions.

Issues Identified by Postdischarge Contact after Pediatric Hospitalization: A Multisite Study.

BACKGROUND: Many hospitals are considering contacting hospitalized patients soon after discharge to help with issues that arise. OBJECTIVE: To (1) describe the prevalence of contactidentified postdischarge issues (PDI) and (2) assess characteristics of children with the highest likelihood of having a PDI. DESIGN, SETTING, PATIENTS: A retrospective analysis of hospital-initiated follow-up contact for 12,986 children discharged from January 2012 to July 2015 from 4 US children's hospitals. Contact was made within 14 days of discharge by hospital staff via telephone call, text message, or e-mail. Standardized questions were asked about issues with medications, appointments, and other PDIs. For each hospital, patient characteristics were compared with the likelihood of PDI by using logistic regression. RESULTS: Median (interquartile range) age of children at admission was 4.0 years (0-11); 59.9% were nonHispanic white, and 51.0% used Medicaid. The most common reasons for admission were bronchiolitis (6.3%), pneumonia (6.2%), asthma (5.1%), and seizure (4.9%). Twenty-five percent of hospitalized children (n=3263) reported a PDI at contact (hospital range: 16.0%-62.8%). Most (76.3%) PDIs were related to follow-up appointments (eg, difficulty getting one); 20.8% of PDIs were related to medications (eg, problems filling a prescription). Patient characteristics associated with the likelihood of PDI varied across hospitals. Older age (age 10-18 years vs <1 year) was significantly (P<.001) associated with an increased likelihood of PDI in 3 of 4 hospitals. CONCLUSIONS: PDIs were identified often through hospital-initiated follow-up contact. Most PDIs were related to appointments. Hospitals caring for children may find this information useful as they strive to optimize their processes for follow-up contact after discharge.

Community childhood injury surveillance: an emergency department-based model.

OBJECTIVE: To describe the use of an emergency department (ED)-based injury surveillance model to determine the incidence and mechanisms of nonfatal injuries among children living in Cabrini Green, a poor urban community. METHODS: Using ED records and census data, population-based injury rates were determined for a retrospective cohort of children, 0 to 14 years old, (N = 3908) with nonfatal injuries resulting in ED treatment between January 1994 and December 1998. RESULTS: There were 1950 nonfatal injuries during the 5-year study period (annual injury incidence of 998/10,000). Age-specific rates (per 10,000 per year) were 899 among 0- to 4-year olds, 616 among 5- to 9-year olds, and 435 among 10- to 14-year olds. Sixty-three percent were male. The most common injury mechanisms were falls (339/10,000 per year), being struck by/against an object (201/10,000 per year), and being cut/pierced by an object (87/10,000 per year). Falls from a building window (2/10,000 per year) were infrequent. The incidence of housefire-related burns was 1.5/10,000 per year. Intentional injuries included alleged child abuse, 43/10,000 per year, and assaults, 30/10,000 per year. The assault rate among 10- to 14-year-old males was 100/10,000 per year. One hundred thirty-four children were admitted to the hospital (average annual rate of 69/10,000). The most frequent admission diagnoses were falls (22/10,000) among 0- to 9-year olds and assaults (13/10,000) among 10- to 14-year olds. CONCLUSION: An ED-based injury surveillance system can provide an efficient and useful way to determine injury incidence in a defined urban community. The data suggest that rates of violence-related injuries were high, while rates of window falls and housefires were low. These data have allowed targeted injury prevention efforts in Cabrini Green, and future surveillance will allow the evaluation of injury prevention activities.