Objectives: (1) To examine the relationship between access to care and the number of missed workdays, and (2) to determine how this relationship is confounded by the presence of having health insurance and health care use among migraineurs.
Design/methods: This retrospective, pooled, cross-sectional study used 1996 to 1999 Medical Expenditure Panel Survey data. Employed migraineurs who were between 18 and 65 years of age were included. Individuals reporting a neoplastic or an acute pain condition were excluded. An access to care index was developed using Rasch Partial Credit Analysis. A 2-part model was used to estimate the annual number of missed workdays.
Results: Of the 703 migraineurs, 538 (77%) reported missing work time. Of those who missed work, the mean (SE) annual number of missed workdays was 4.4 (.39). A higher level of access to care (P= .025) and presence of depression (P= .033) were significantly associated with missing a greater number of workdays. We created a proxy for migraine severity based on migraine-related prophylactic medication use and hospitalization(s). Severe migraines were significantly (OR = 2.01, SE = .51, P= .006) associated with an increased likelihood to miss workdays. When health insurance was included in the model, a higher level of access to care was significantly associated with the increased likelihood to miss workdays (OR = 1.04, SE = .021, P= .05). From the original model, the odds ratio (1.035 to 1.040) and the SE (.020 to .021) increased slightly. When health care use was included in the model and health insurance was removed, (1) emergency department visits were significantly (P= .006) associated with missing a greater number of workdays, and (2) access to care was significantly associated with missing a greater number of workdays (P= .028). When having health insurance and health care use were simultaneously included in the model, a higher level of access to care was significantly associated with greater likelihood to miss work (OR = 1.040, SE = .0212, P= .05) and missing a greater number of workdays (P= .005). However, a change of 1 standard deviation in the score would be associated with a 12% change in the odds to miss work and only 8 percentage points change in the number of missed workdays.
Conclusions: Contrary to expectations, a higher level of access to care is significantly associated with an increased likelihood to miss work and with missing a greater number of workdays. Depression, migraine severity, and health care use are important explanatory variables. Having health insurance may be a confounder between access to care and workplace absenteeism.