Individually significant risk factors do not provide an accurate clinical prediction rule for infant underimmunization in one disadvantaged urban area

Ambul Pediatr. 2006 May-Jun;6(3):165-72. doi: 10.1016/j.ambp.2006.01.002.


Objective: To define a clinical prediction rule for underimmunization in children of low socioeconomic status.

Methods: We assessed a cohort of 1160 infants born from July 1998 through June 1999 at an urban safety net hospital that received primary care at 4 community health centers. The main outcome measure was up-to-date status with the 3:2:2:2 infant vaccine series at 12 months of age.

Results: Latino infants (n = 959, 83% of cohort) had immunization rates of 74%, at least 18% higher than any other racial/ethnic group. Multivariate logistic regression demonstrated the following independent associations (relative risk, 95% confidence interval) for inadequate immunization: non-Latino ethnicity (1.7, 1.4-2.0), maternal smoking (1.3, 1.1-1.7), no health insurance (1.9, 1.4-2.3), late prenatal care (1.9, 1.5-2.3), no pediatric chronic condition (2.1, 1.2-3.1), and no intent to breast-feed (1.3, 1.1-1.6). However, the index of concordance (c-index) for this model was only 0.69. Neither excluding infants who left the health care system nor accounting for infants who were "late starters" for their first vaccines improved the predictive accuracy of the model.

Conclusions: In this predominantly Latino population of low socioeconomic status, Latino infants have higher immunization rates than other infants. However, we were unable to develop a model to reliably predict which infants in this population were underimmunized. Models to predict underimmunization should be tested in other settings. In this population, interventions to improve immunization rates must be targeted at all children without respect to individual risk factors.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Cohort Studies
  • Colorado
  • Humans
  • Immunization / statistics & numerical data*
  • Infant
  • Predictive Value of Tests
  • Risk Factors
  • Socioeconomic Factors
  • Urban Health*
  • Vulnerable Populations*