8. Comparing the DSQ with other health measures A New Survey Measure of Disability: the Disability Screening Questions (DSQ)
8. Comparing the DSQ with other health measures

Section 2.2 described two other health measures: the Health Utilities Index Mark 3 (HUI3) and the Washington Group questions (WG). Since these three measures were all included on the Canadian Community Health Survey (CCHS) either in 2010 or 2011, a study was done to compare them.^{Note 1} It was known that these three constructs measured different concepts, but still, it was of interest to see to what extent they compared. The study showed that the DSQ and HUI3 yielded a similar overall prevalence rate (27% to 29%), but prevalence rates by disability type were higher for the DSQ. The overall WG prevalence rate was much lower^{Note 2} (about 8%), and prevalence rates for some disability types were closer to the HUI3 than the DSQ.

A follow-up study^{Note 3} was conducted to better understand the differences between the DSQ and the two other measures, the results of which are summarized below.

Comparing DSQ with WG

In terms of concepts and question wording, the most comparable DSQ and WG disability types were Seeing, Hearing, Mobility, and to some extent, Dexterity; the less comparable were Learning and Memory. For each type, the question and answer categories from the DSQ module that were the most comparable to those of the WG were used to compute a “WG-modified” DSQ rate. Chart 7 shows the WG rate, the WG-modified DSQ rate, and the original DSQ rate (RACXR definition). A global rate was also computed. For the DSQ, it was computed based on only the disability types shown on the graph. For the WG, the global rate was taken directly from Bernier’s study. The Learning and Memory types are less comparable, owing to differences in concepts and question wording.

The WG rate and WG-modified DSQ rates are quite close, except for Memory, which is one of the less comparable types in terms of question wording.

The conclusion of the follow-up study is that if, from the DSQ module, only the questions on level of difficulty  (questions on activity limitations are ignored) and the most severe levels of difficulty (“A lot” or “Cannot do”) are considered, DSQ disability rates by type and overall are much closer to WG rates.

Comparing DSQ with HUI3

The DSQ and HUI3 questionnaires differ substantially. To mimic the HUI3 for some disability types, answers to several questions from the DSQ module were grouped. For example, in the HUI3, use of a Hearing or Mobility aid is considered when identifying PWD, but not in the DSQ. Thus, for the DSQ to conform to the HUI3 definition, respondents who reported using an aid had to be included among PWD, even if, according to the DSQ definition, this did not necessarily qualify them as PWD.

The most comparable disability types between the DSQ and HUI3, in terms of concepts and question wording, were Seeing, Hearing, Walking (although different from the DSQ Mobility, which also considers using stairs), and Dexterity; the less comparable were Emotional (HUI3 uses a Happy scale) and Cognition (including Learning, Memory and Developmental).

Chart 8 presents disability rates based on the HUI3, the HUI3-modified DSQ rates, and the original DSQ rates (RACXR definition). The DSQ rate for “Walking” shown in this chart is based on those people who reported difficulty walking in the DSQ (ignoring those who reported only difficulty using stairs), and whose daily activities are limited. For the DSQ, the global rate is based on the disability types in the graph plus pain. For the HUI3, the global rate was taken directly from Bernier’s study. The Emotional and Cognition types are less comparable, owing to differences in concepts and question wording.

Using the questions and answer categories from the DSQ module that are most similar to the HUI3 yields similar disability rates from both instruments. Hence, if only questions on level of difficulty are used from the DSQ module, ignoring the ones on activity limitations (as is the case in the HUI3), and if only the most severe levels of difficulty are considered (“A lot” or “Cannot do”), prevalence rates by type are close in the two instruments, and so is the global rate. To be more comparable with the HUI3 for Hearing and Walking, the DSQ must take into account the use^{Note 4} of aids. For Mobility, the HUI3 does not consider difficulty using stairs, which is taken into account in the DSQ; if only difficulty Walking is considered in the DSQ, the prevalence rate is closer to that of the HUI3. For Cognition, the DSQ definition that considers only the presence of a condition gives a closer estimate to the HUI3.

Different measures, different purposes

This follow-up study showed that the DSQ can yield results similar to those of the WG or HUI3 when comparable question and answer categories are used to define disability. Generally, a definition based on the two most severe levels of difficulty with the DSQ will produce rates similar to the WG. For the HUI, the comparison is less direct because of the different nature of questions. Nonetheless, using a definition based on the two most severe levels of difficulty and considering the use of aids for Hearing and Walking as an indicator of disability with the DSQ will yield similar results. However, for the cognitive types, only the presence of a condition should be taken into account in the DSQ to obtain similar disability rates. (The DSQ use this definition for Developmental disability.) 

The DSQ, the HUI3 and the WG each has its own purpose. The DSQ were not created to replace any of the other measures. This study only served to better understand the difference in the results from each measure. Thus, when estimates from the three measures are compared, it is expected that rates of disability will differ.

• Previous
• Next