Main Findings of Included Experimental Studies

First Author (Year)DesignSample Size/PopulationIntervention/Study FocusMedical ComplexityPreventable HospitalizationEffect DirectionaMain Findings
Gadomski (1998)25Prepost without control128 025 children aged 0–18 with fee-for-service Medicaid in Maryland (3 y pre and 2 y post)MAC: Mandatory enrollment, fee-for-service primary care case management program including: (1) assigned PCP who authorizes ED, inpatient and specialty care, and (2) provides Early Periodic Screening, Diagnosis and Treatment Services, (3) higher reimbursement for physicians, hospital care and long-term care, (4) access to medical home 24/7, and (5) on-line eligibility verification system for providersInvestigator-defined set of ICD-9 codes, including: renal disease, sickle cell disease, cardiac anomalies, chromosomal anomalies, congenital anomalies, mental retardation, cerebral palsy, malignancies, neoplasms, metabolic disorders, cystic fibrosis, HIV, other immune deficiencies1. Investigators developed list of avoidable hospitalizations from combinations of inpatient ICD-9 codes with ambulatory and/or pharmacy claims, eg, inpatient asthma ICD-9 code without antecedent pharmacy claim for steroids. List refined by modified-Delphi appropriateness method with expert panel.MAC enrollment no difference in avoidable hospitalization (OR, 0.93; 95% CI: 0.86–1.01), or ACSC hospitalization (OR, 0.98; 95% CI: 0.94–1.03) in final model, but all hospitalizations decreased. Avoidable hospitalization had ACSC in 92%, whereas ACSC hospitalizations were avoidable in 38%, unavoidable 12%, and unclassified 50%. As proxy for illness severity, SSI group avoidable hospitalization AOR, 1.73 (95% CI: 1.54–1.95), and ACSC hospitalization AOR, 2.81 (95% CI: 2.63–3.01). Preventive care visits associated with lower avoidable and ACSC hospitalizations, but primary care visits associated with higher avoidable and ACSC hospitalizations.
Gordon (2007)38Prepost without control227 (of 230) children enrolled in special needs program at Medical College of Wisconsin, up to 3 y data pre and 3 y data postSpecial Needs Program: Unified care plan, coordination, and communication across specialty and primary care, participation and advocacy at specialty and primary care visits, school, with payers. Single point of contact, home visits, outreach to PCPs and community resources, RN case management, psychosocial support. MD available 24/7, rounding with inpatient services. Seventy percent in program have nurse case manager only; 30% have both program RN and MDProgram enrollment criteria = combinations of high levels of subspecialty use (≥5), organ system involvement (≥3), past utilization, technology assistance, and additional disease, geographic and socioeconomic factorsReduction in hospitalizations after interventionFewer hospitalizations and hospital days in the postperiod: 40% decrease in median admission rates, 17.6% decrease in median hospital days P < .003, with corresponding increase outpatient services and decreases in charges/payments. Suspected key activities were partnership with family and PCP, familiarity with child’s condition, close involvement during hospitalizations, and proactive ambulatory care.
Liptak (1998)12Quasi-experimental, nonequivalent comparison10 715 hospitalizations for selected acute and chronic conditions between 1984 and1995 for children <18 y, at Children’s Hospital in Rochester. Compared chronic hospitalizations versus acute (appendicitis, bronchiolitis, fracture) locally and versus consortium of 18 US tertiary academic medical centersExpanded funding by regional insurance company to provide ambulatory care coordination and wraparound services (RN, SW, Psychology, Occupational Therapy/Physical Therapy/Speech, Special Ed): 50% of population covered by this companyInvestigator-defined set of ICD-9 codes, including: cardiac disease, cancer, epilepsy, neurologic diseases, ventriculoperitoneal shunt, heme and sickle cell diseases, rheumatologic and musculoskeletal diseases, spina bifida, cystic fibrosisReduced hospital rates 30-d readmissionsAnnual hospitalizations for children with chronic conditions decreased from 2796 to 1622 between 1984 and 1995 (R2 = 0.82, P < .001), LOS decreased (R2 = 0.83, P < .001). No change for acute admissions (R2 = 0.08, P = .45). Compared with other academic centers, chronic condition readmits were lower (12.7%, 95% CI: 10.4–14.0 vs 15.0%, 95% CI: 14.2–15.7). No denominator to compare hospitalization rates directly, but using appendicitis as proxy, found lower chronic condition hospitalizations versus other centers (P < .01).
Palfrey (2004)15Prepost without control150 CSHCN, recruited by PCPs for complexity from 6 voluntary pediatric practices in Boston (4 private practices and 2 neighborhood health centers). Data collected pre (n = 150) and 2 y post (n = 117, 78%)PACC: Integrated system based on medical home principles: Multifaceted intervention to provide shared decision-making, partnership, and comprehensive care at community level, improve coordination/communication among PCPs, subspecialists, and families. Key activities: Nurse practitioner and home visits, consultation from local parent of CSHCN, modifications of office routines, individualized health plans, regularly scheduled continuing medical education, expedited referrals, and communication with specialists/hospitalProgram enrollment criteria = combinations of high levels of subspecialty use (≥2), organ system involvement (≥2), high severity, past utilization, technology assistance, or significant development problems or risks, care coordination difficultyReduction in hospitalizations after interventionAnnual hospitalization rates decreased from 58% to 43.2% of patients (P < .01). When stratified by age, there was no difference in 6–18 y olds. (The 0–5 y olds decreased from 62.9% to 45.2%). Almost 40% of families noted improvement in respite and transportation services. No change in reports of access for care, telephone advice, prescription refills, specialized equipment, and supplies. Findings between white and nonwhite children appeared to be equal.
  • a ↑, Main predictor or intervention associated with increased preventable hospitalizations; ↓, Main predictor or intervention associated with reduced preventable hospitalizations; ↔, No significant (or inconsistent) association between main predictor or intervention and preventable hospitalizations.