Table 4.

Health Care Utilization Outcomes for Negative Binomial Models

CharacteristicUrgent-Care VisitsEmergency Department VisitsHospita Visits
Unadjusted,
IRR (95% CI)
Adjusted,
IRR (95% CI)
Unadjusted,
IRR (95% CI)
Adjusted,
IRR (95% CI)
Unadjusted,
IRR (95% CI)
Adjusted,
IRR (95% CI)
Intercept0.68 (0.55-0.83)0.45 (0.23-0.88)0.38 (0.29-0.50)0.31 (0.14-0.70)0.15 (0.10-0.22)0.07 (0.02-0.25)
Stabilization1.22 (0.93-1.61)1.11 (0.86-1.44)2.32 (1.64-3.28)a1.74 (1.27-2.39)a2.10 (1.29-3.42)a2.01 (1.25-3.23)a
Outreach1.10 (0.81-1.48)1.14 (0.86-1.52)1.60 (1.10-2.38)a1.31 (1.01-1.79)a1.21 (0.69-2.11)1.06 (0.60-1.85)
Self-care generativityReferenceReferenceReferenceReferenceReferenceReference
Emergency visits prior1.49 (1.38-1.61)a
Hospital visits prior2.07 (1.64-2.61)a
Urgent-care visits prior1.38 (1.31-1.47)a
Age1.00 (0.99-1.01)0.99 (0.98-1.01)1.01 (0.99-1.03)
Sex0.79 (0.63-1.00)1.05 (0.80-1.36)1.32 (0.89-1.98)
Insured1.26 (1.00-1.59)1.37 (1.04-1.80)a1.39 (0.91-2.1)
Immigrant status1.08 (0.70-1.65)0.84 (0.51-1.39)0.53 (0.25-1.11)
Language preference1.09 (0.83-1.43)1.14 (0.82-1.58)1.23 (0.74-2.04)
  • IRR = incidence rate ratio.

  • a Statistically significant result.

  • Note: Health care utilization outcomes—IRRs and 95% CIs for health care events over cohort follow-up, adjusted for age, sex, health insurance status, immigrant status, language preference, and prior year health care utilization. The model yields parameter estimates representing a log-count value (not shown) which are then used as an exponential of e. yielding an interpretable result (shown in table), representing the percentage of the original count units.