In their systematic review of scales assessing quality of life (QoL) in people living with dementia, Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019) did not identify any existing instrument applicable for long-term care with sufficient psychometric properties, feasibility, and availability.

The most important quality of any instrument is its content validity, which is defined as “the degree to which the content of an instrument is an adequate reflection of the construct to be measured” (p 1160; Terwee etal., Reference Terwee2018). As Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019) concluded, content validity was only formally evaluated in one of the nine instruments reviewed (QUALIDEM; Ettema etal., Reference Ettema, Dröes, de Lange, Mellenbergh and Ribbe2007).

The findings of Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019) raise two central issues of concern. First, implementing QoL assessments requires certainty that the measurement is tapping what we are interested in. We need to know if the conceptualizations and items in the instrument reflect the important aspects under study from the view of people living with dementia. Next, lack of proper standardized measurement of QoL is a potential source of error in both clinical settings and research reports. Working in evidence-based practice (Melnyk etal., Reference Melnyk, Fineout-Overholt, Stillwell and Williamson2010), this has significant implications for reaching correct conclusions about effective and helpful approaches in any assessment of health interventions.

However, as several recent initiatives have focused on developing clinically relevant core outcome sets for people living with dementia, we are optimistic about implementing systematic QoL assessment in long-term care. These initiatives target both issues by recommending standardized assessment packages, as well as ensuring experts living with dementia inform the content validity of these sets of instruments. The COMET initiative (COMET Initiative | Home [comet-initiative.org]) is currently working on and publishing core outcomes important for home-dwelling people living with dementia (Harding etal., Reference Harding2019; Reilly etal., Reference Reilly2020). ICHOM have developed a standard core outcome set for people living with dementia throughout the disease process (Dementia – ICHOM Connect).

While Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019) used a systematic approach from the Consensus-based Standards for Selection of Health Measurement Instruments (COSMIN) initiative, more refined and comprehensive COSMIN ratings for reviewing patient rated outcome instrument (PROMS) have been published recently (L. B. Mokkink etal., Reference Mokkink2017; Prinsen etal., Reference Prinsen2018; Terwee etal., Reference Terwee2018). In particular, Terwee etal. (Reference Terwee2018) describes a systematic evaluation of content validity, allowing reviewers to evaluate the content validity based on available knowledge about the instruments. We strongly advice researchers and practitioners to attempt this systematic approach, to gain information about the appropriateness of instruments.

For PROMS, securing content validity includes the relevance, comprehensiveness, and comprehensibility of the instrument, as evaluated by members of the target group. If evidence of content validity is absent, investigating this by conducting one’s own qualitative interviews or focus groups (Brod, Tesler, and Christensen, Reference Brod, Tesler and Christensen2009) is a better option than developing a new instrument. As Santana-Berlanga etal. (Reference Santana-Berlanga, Porcel-Gálvez, Botello-Hermosa and Barrientos-Trigo2020) identified 24 instruments assessing QoL applicable for persons with dementia living as “institutionalized older adults”, adding new QoL-instruments to this pool will increase the problem of a lack of systematic measurement and make comparisons in meta-analyses complicated. Investigating the content validity of existing PROMS, where people living with dementia are included as experts, is not only the gold standard (Terwee etal., Reference Terwee2018) but is also something we believe people living with dementia want to contribute to. Participating in research may well provide side effects of feeling useful, empowered, and valued (Øksnebjerg etal., Reference Øksnebjerg2018).

One part of content validity is “relevance”, and here a central criterion is an appropriate recall period (Terwee etal., Reference Terwee2018). In general, recall bias is a problem for all aggregated retrospect ratings evaluating long-time intervals (Shiffman etal., Reference Shiffman, Stone and Hufford2008). For residents in care homes living with moderate to severe dementia, their current emotional state may interfere with the scoring of their past experiences (Kolanowski etal., Reference Kolanowski, Hoffman and Hofer2007). Thus, a proper recall period is critical to ensure valid QoL ratings, and several inquiries advocate for momentary assessment to increase validity and decrease the recall bias of self-reports (Clarke etal., Reference Clarke2020; Øksnebjerg etal., Reference Øksnebjerg2018; Shiffman etal., Reference Shiffman, Stone and Hufford2008). For implementing QoL assessments in long-term care, applying a momentary focus, or conducting assessment through interviews, will probably facilitate self-report for people living with dementia longer throughout the time of illness (Ferring and Boll, Reference Ferring and Boll2010; Logsdon etal., Reference Logsdon, Gibbons, McCurry and Teri2002). Meta-analyses suggest dementia severity is not correlated strongly with self-reported QoL (Martyr etal., Reference Martyr2018), and implementation of momentary assessments would help clinicians detect drift from individual baselines and intervene when needed.

The feasibility assessment by Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019) is valuable for practitioners and researchers seeking to use any of the reviewed instruments. In the conclusion, it is stated that the instruments “function poorly” based on their psychometric properties. But do they function poorly, or are they just not investigated thoroughly enough? COSMIN (Mokkink etal., Reference Mokkink2010; Mokkink etal., Reference Mokkink2017) offers a strict assessment of instruments. In our experience, many instruments will easily get a minimal score as the “lowest score counts”.

As opposed to Hughes etal. (Reference Hughes, Farina, Page, Tabet and Banerjee2019), we strongly discourage using dementia-specific instruments for the whole population in long-term care to compare aggregated scores or individual development. The content validity of dementia-specific instruments is developed to capture the specific issues for this condition and using a dementia-specific instrument for all care home residents would in our view suggest dementia is inevitable for this multifaceted population. We fear implementation of dementia-specific scales in general would lead to implicit attitudes of expecting cognitive impairment in people in care homes living there for other reasons. If the aim is to compare individual scores or aggregated scores over time, using generic instruments for all residents may be a better option.

Content validity is always context-specific (Terwee etal., Reference Terwee2018), and instruments with low content validity in one context may have high content validity in other contexts. Nevertheless, the most important assessment is to evaluate what one wants to measure and find the best instrument fit for this purpose. Other instruments have been recommended for this population based on their total evidence and feasibility, including assessment from the view of people living with dementia. ICHOM recommends QoL-AD (Logsdon etal., Reference Logsdon, Gibbons, McCurry and Teri1999) and Quality of Wellbeing Scale-Self-administered (Sieber etal., Reference Sieber, Groessl, David, Ganiats and Kaplan2008) to assess overall QoL and well-being. These are freely available for nonprofit use when registering on the ICHOM website. Another interesting approach is to look at, or even develop, item banks based on item response theory (Edelen and Reeve, Reference Edelen and Reeve2007), where generic questions are applicable for a variety of populations (see, for example, PROMIS [healthmeasures.net]). These item banks are created to ensure optimal scalability of the unidimensional items, where each item corresponds to a level of QoL. In practice, the item bank is adjusting to the resident and terminates the questionnaire when the QoL level is detected. We hope establishing content validity in the most promising instruments is a target for future research, and we encourage feasible and available instruments to be refined.