How can we measure disability in research related to the COVID-19 response?

Challenge #1: Various approaches to measuring disability exist, reflecting different models and definitions of disability

• Both the International Classification of Functioning, Disability and Health (ICF) and the UN Convention on the Rights of People with Disabilities (UNCRPD) describe disability as an interaction between impairments and various barriers, that causes or risks participation restriction.
• Not all available tools that are commonly used for measuring disability are compatible with this definition and amongst those that are, many measure only one component of the interaction, e.g. self-reported functioning, participation restrictions, environmental barriers or impairments separately, leading to difficulties comparing estimates
• In countries such as the United Kingdom and elsewhere in Europe, where the definition of disability in legislation includes participation restrictions, different tools are used that incorporate both self-reported functioning and participation restrictions together
• Question variations of “do you have a disability” will only identify people with the most substantial functioning limitations and can be stigmatising, yet are still commonly used

Challenge #2: Use of the Washington Group Questions is common, but the pros and cons of the different modules is not always clear

• The most commonly used tool is the Washington Group Short Set (WGSS), which aims to identify the majority of people at risk of participation restrictions because of a health condition or impairment. It focuses on six basic functional domains: seeing, hearing, walking, cognition, communicating and self-care.
• The WGSS are widely used and recommended, including for SDG indicator disaggregation. They are simple, quick, reproducible and easy to translate into different languages, allowing results to be compared over time and between countries. However, the WGSS provides limited data on other components of disability, such as participation restrictions or impairments. It also excludes psychosocial functioning and mental health.
• The Extended Set of Washington Group questions (WGExS) contains up to thirty-five questions, and can be used in surveys where more time is available. They capture a more complete picture of disability but take longer to complete. When using the Extended set, prevalence tends to be much higher, due to high rates of pain and fatigue. The Washington Group have also collaborated with UNICEF to develop an extended set of questions on functioning for children aged 2 to 17, and the International Labour Organisation to develop a livelihoods module.
• The Washington Group “Enhanced Short Set” (WGES) is a hybrid tool which includes the 6 questions of the Short Set, plus six additional questions related to upper body strength and mental health.

Challenge #3: Tools are translated, delivered, analysed, interpreted and presented in different ways, making comparison challenging

• Rigour in translating and testing tools to measure disability is mixed, with anecdotal evidence that cognitive testing (testing whether key concepts are understood) is not regularly completed
• The way that questions are delivered – for example whether to the individual or a proxy via household head or caregiver report, and whether or not response options are read aloud to participants, affect the outputs
• The most common tool, the WGSS, can be analysed in different ways including considering anyone who reports “a lot of difficulty” or higher in any domain (or “daily” and “a lot” for either of the psychosocial functioning questions) as having a disability; or considering anyone who reports “some” or greater in any domain as having a disability
• Different analytical and theoretical approaches are applied to data to assess the impact on participation or wellbeing of people with disabilities, that can be complex or poorly explained.
• Disaggregation of data is important, as age group, gender and type of limitation can add additional layers of barriers or exclusion amongst people with disabilities, but often sample sizes are too low to disaggregate data meaningfully, and data is often presented for “all people with disabilities” only.
• Some tools, such as the Model Disability Survey, require more complex statistical analyses to categorise people as having a disability or not

Challenge #4: Stakeholders may refuse to include tools with multiple questions on account of the number of questions included

• Tools such as the UK Equality Act Disability Definition (EADD) and Global Activity Limitation Instrument (GALI) provide holistic information on health conditions that cause participation restriction and are more commonly used in High Income settings
• Because they do not ask about specific functions, they are shorter than the WGES. However, this means they also do not collect data on the type of limitation the participant has, or whether or not they have considered the use of assistive devices (e.g. glasses) in their response

Challenge #5: Research questions related to health and rehabilitation would benefit from information on impairments and health conditions

• Impairments or health conditions are components of disability that can be measured directly using objective testing. These data are important for planning appropriate health and rehabilitative services amongst those who would benefit from these (e.g. provision of cataract surgery, hearing aids, mobility devices). This may be particularly needed in low resource settings where inadequate access to health care is closely related to disability
• However, impairment and health condition testing in isolation does not consider the effect on the individual’s level of functioning or participation. This type of assessment may also be more resource intensive to undertake than gathering self-reported information.