We thank Caban-Martinez et al for their important observation that the CIDI-SF and other diagnostic tools used in many community-based surveys may overestimate the rates of depression in individuals with arthritis. Caban-Martinez et al suggest that the response of those with arthritis to certain scale items (e.g., “Did you feel tired out or low on energy all the time?”) may reflect their physical symptoms of arthritis, rather than their state of mental health.
To our knowledge, there has not been a validation study of the CIDI-SF focused on a sample of those with arthritis. We agree with Caban-Martinez and colleagues that this is an important issue that should be addressed in future research. Blalock and colleagues examined the impact of scale items assessing physical symptoms such as sleep disturbances in the CES-D scale in a population with arthritis (1). Blalock et al concluded that inclusion of these items was “unlikely to have any effect” on studies examining the determinants of depression among individuals with arthritis. However, they noted that including these scale items may result in a “modest” bias toward overestimation of the prevalence of depression among those with arthritis.
There are 2 additional suggestions for future research that we would like to propose. Researchers comparing the mental health of those with arthritis with those without arthritis in large community surveys should consider a range of mental health outcomes. In particular, suicidal ideation is extremely important to evaluate. Our finding that those with arthritis had twice the odds of suicidal ideation in comparison with those without arthritis provides strong support for regular mental health screening of this population.
Second, Statistics Canada and other institutions producing large data sets for epidemiologic studies should provide responses to each scale item in their public use data sets. In the public use data sets of the CCHS, only the aggregated score of the CIDI-SF scale was provided. Individual scale items were not available. Therefore, we were not able to explore how responses varied between those with and without arthritis on each scale item.
As Caban-Martinez and colleagues have suggested, validation of the CIDI-SF among those with arthritis would be helpful. Individuals with arthritis who are experiencing comorbid depression are vulnerable to many negative outcomes, including poorer medical compliance (2) and higher rates of suicide (3), than those who are not experiencing depression. Yet, depression often goes undiagnosed and untreated in patients with arthritis (4). In order to improve outreach and provision of services to the most vulnerable community dwellers with arthritis, additional population-based research using abbreviated diagnostic interviews is needed.