Adults and healthcare professionals attach much value to social participation, as it contributes towards a person’s general health and post-illness recovery. However, past research has shown that social participation is difficult to measure and contains different subdomains that do not apply to everyone. By way of illustration, social participation to some people means undertaking paid work combined with informal care, while to others it is about looking after children in combination with voluntary work. In addition, the importance of subdomains can change over time, such as when a person retires or becomes seriously ill.
For these reasons, we have to be able to measure social participation according to a flexible but standardized method that can be used for all adults in different situations. The American National Institutes of Health (NIH) has funded the development of Patient-Reported Outcomes Measurement Information System (PROMIS). PROMIS consists of item banks that can be applied universally to adults or children. These item banks are large sets of questions (items) that all measure the same construct, such as social participation. PROMIS item banks are developed based on Item Response Theory (IRT). An advantage of IRT is that it is possible to use a selected number of automatically generated items, with the help of computer adaptive testing, that are relevant to each specific individual.
Because of the smart measuring method, people have to fill in only 3 to 7 items for a reliable score, while they can still be compared with each other. It is also possible to compare scores of people who have filled in the whole item bank with those who have completed items from a particular subdomain from the item bank, given that the scores are subject to the same metric. Social participation is measured with the ‘Ability to Participate in Social Roles and Activities’ item bank. The item bank consists of 35 items and assesses the perceived capacity to carry out normal social roles and activities (for example, ‘I feel limited in the amount or time I have for my family,).
First study
In the first MEES study, we interviewed 46 adults with and without health problems and asked them what social participation meant to them. The participation subdomains that were identified were compared to the 35 items from the current item bank. It appeared that not every subdomain of social participation that is meaningful to them is included in the current item bank, such as informal care and voluntary work.
Second study
In the second MEES study, and in collaboration with social participation experts, new items that measure/cover these subdomains have been generated. The generation and testing of the new items was guided by PROMIS scientific standards. Interviews with adults with and without (chronic) health problems were held to test whether these new items were relevant and comprehensible , and whether social participation subdomains were still missing. In total, 17 new items that are relevant and comprehensible were proposed to be added to the original item bank. A field-test study is currently underway to see whether the item bank with the 17 new items is indeed better able measure social participation.
Third study
Social participation can be measured with the PROMIS item bank ‘ability to participate in social roles and activities’ (APS). To date, there is insufficient knowledge on the psychometric properties of the PROMIS APS item bank. Therefore, the aim of this third study is to investigate the responsiveness and test-retest reliability of the PROMIS APS item bank.
To measure patient reported outcomes within Amsterdam UMC several PROMIS item banks, including the PROMIS APS item bank, have been integrated in our hospital information system EPIC as computerized adaptive tests (CATs). In collaboration with Strategy and Innovation and the departments Rehabilitation Medicine and Neurology of Amsterdam UMC, we have implemented the PROMIS CAT APS in routine care. Before the PROMIS CAT APS can be optimally used for care purposes it is important to gain more insight in the psychometric properties, which can be used to study whether change scores can be used to measure treatment success.
Contact: Sietske Tamminga: s.tamminga1@amsterdamumc.nl and Astrid de Wind: a.dewind@amsterdamumc.nl
