OBJECTIVES: The objective was to understand the extent of the insurance-loss problem in the United States over a four-year period (2004-2007) by (1) determining the number (and proportion) of individuals who experienced loss in insurance coverage and (2) for those who experienced gaps in coverage, determining the average length of those periods of time when they were without coverage.  METHODS: This study was a retrospective, longitudinal data analysis using monthly data from the 2004-05, 2005-06, and 2006-07 panels of the Medical Expenditure Panel Survey (MEPS). Individuals were categorized on the basis of continuity of insurance coverage:  a cohort of continuously insured individuals; a cohort of continuously uninsured people; and a cohort of people who experienced gaps (from one to 23 months) in coverage. The number and percent of individuals who were in each group for each panel were calculated.  A chi-square test was used to determine if there were differences in the distribution of health insurance patterns between the panels, and ANOVA was used to determine if there were differences between the average gap length over the years. RESULTS: Pooling across the three panels, there were 157,641,235 (17.37%) individuals who experienced gaps in insurance coverage, 82,826,100 (9.12%) who were continuously uninsured for 24 months, and 616,864,391 (67.95%) who had continuous coverage over a two-year period. The proportion of continuously uninsured rose from 8.73% in 2004-05 to 9.23% in 2006-07. The average gap length was 8.56, 8.44, and 8.54 months in 2004-05, 2005-06, and 2006-07 respectively.  There were no significant differences across the panels in either the distribution of insurance coverage patterns or between the average gap lengths. CONCLUSIONS: Patterns in health insurance coverage have remained stable over the past several years. Examination of changing patterns of insurance coverage will be important as health reform is introduced between now and 2014.