Validity of an illness severity measure for headache in a population sample of migraine sufferers

The headache impact questionnaire (HImQ) is used to measure pain and activity limitations from headache over a 3-month recall period. In a prior study, the test-retest reliability of the eight-item HImQ score was found to be relatively high (0.86). In the current study, we examined the validity of the eight-item HImQ by comparing the overall score and individual items to equivalent measures from a 90-day diary. Pain and activity limitations due to headache were assessed in a population-based sample of 132 migraine headache sufferers enrolled in a 90-day daily diary study who completed the HImQ at the end of the study. The HImQ score was derived from four frequency-based questions (i.e. number of headaches, missed days of work, missed days of chores, or missed days of non-work activity) and four summary measures of average experience across headaches (i.e. average pain intensity, and average reduced effectiveness when having a headache at work, during household chores, and in non-work activity). Diary based measures were used as the gold standard in evaluating the HImQ score. Mean and median values of frequency-based HImQ items (e.g. number of headaches) were similar to equivalent diary measures, indicating no systematic bias. In contrast, HImQ measures of average experience across attacks (e.g. average pain intensity) overestimated equivalent diary measures and, in general, better approximated diary measures for migraine headaches, rather than all headaches. The highest correlations between HImQ and diary items were observed for headache frequency and average pain intensity, the two general headache measures, followed by measures of reduced effectiveness. Among frequency-based measures, the strength of the correlation was directly related to the magnitude of the mean. The higher the mean value, the higher the correlation. The correlation between the HImQ score and diary based score was 0.49. The HImQ score is moderately valid. Frequency-based items (e.g. number of missed work days) were found to be unbiased and the highest correlation coefficients were observed for frequency-based items with relatively high mean counts (number of headaches, number of missed non-work days). These findings have implications for measuring severity of chronic episodic conditions like headache, asthma, back pain, arthritis, epilepsy, and panic disorder, which can cause limitations to activities. The validity of illness severity measures may be improved by using frequency-based questions to assess both missed activity days and days with significantly reduced effectiveness or productivity (e.g. by 50% or more). By combining the count for both missed days and days where productivity is substantially reduced, the mean of the frequency-based measure will be increased, a factor which may improve the overall validity of the item. A severity measure can be derived from such items by simple addition and provides a scale with intuitively meaningful units.