Reliability and validity of a global question on self-reported chronic morbidity

Aim

A global question on chronic morbidity is included in many national health interview surveys. According to a recent EU Commission regulation, information on this item should be collected in all EU member states. However, little is known about the reliability and validity of such a question.

Subject and methods

The reliability of a global question on chronic morbidity was investigated among persons who participated in 2001 in both the Belgian health interview survey (HIS) and the national population census (n?=?2,871) by using kappa statistics and logistic regression. In addition, data from the HIS 2001 and 2004 (n?=?21,376) were used to study estimates and determinants of the sensitivity of this global chronic morbidity measure among people with specific chronic diseases.

Results

In terms of reliability, the kappa statistic showed only moderate agreement (0.559; 95 % CI 0.523–0.594). Additionally, the sensitivity of the global question on chronic morbidity ranged from 49.9 to 87.2 %, depending on the type of disease. A much higher sensitivity was observed among people who rated their health status to be moderate to bad (adjusted OR 3.85; 95 % CI 3.17–4.69).

Conclusion

Self-reported chronic morbidity, measured by a single and global question, is a reasonably reliable instrument to measure ill health. The global instrument provides useful information on the burden of disease, because it takes into account the relevance of the diseases for the people themselves.