Linguistic and Psychometric Validation of the Cognition Bolt-On Version of the Japanese EQ-5D-5L for the Elderly

Forward translation

After obtaining the permission from the EuroQol Group, the English cognition dimension was obtained from the original developer of the cognition dimension (M. F. Janssen), and it was cross-culturally adapted for the Japanese language in accordance with previously published guidelines [12, 13]. Two independent translators translated the cognitive dimension to Japanese and reconciled the two translated versions into one. One of the authors (MI), medical expert, reviewed the draft, and it was revised where needed.

Back translation

Another Japanese-English bilingual translator who was blinded to the original English version translated the Japanese draft back into English. The original developer of the cognition dimension reviewed the back translation whether it was conceptually equivalent to the original cognition dimension. Following his confirmation, a provisional Japanese cognition dimension was finalized.

Pilot test

The pilot test was conducted in July 2021. Five Japanese-speaking individuals aged 20 years or older who worked as a caregiver at a nursing home or who were the main caregiver of his/her family member with dementia at home were recruited and completed the provisional Japanese cognition dimension added to the Japanese EQ-5D-5L proxy version after their consent to participate in the pilot test. Following its completion, the participants were debriefed online regarding cognition dimension and the overall impression of the EQ-5D-5L+C, to ensure that the cognitive dimension was understood and answerable. The summarized debriefing results were reviewed. The recruitment and debriefing following an interview guide were conducted by INTAGE Healthcare Inc. (Tokyo, Japan).

Finalization of the Japanese version

After the pilot test, the provisional Japanese cognition dimension was to be revised where needed and was finalized after the authors’ agreements. The Japanese cognition dimension embedded in the Japanese EQ-5D-5L (EQ-5D-5L+C) proxy version was finalized.

Participants

The finalized Japanese EQ-5D-5L+C proxy version was administered in all the nursing home residents of six nursing homes associated with Life Group, Inc. (Tokyo, Japan). The collected anonymous processed data were used for psychometric validation of the Japanese EQ-5D-5L+C. Consent was obtained from the nursing home residents or his/her potential surrogate (e.g., family member and adult guardian) using the informed consent form approved by the ethics committee. Residents with data of the EQ-5D-5L+C, EQ-5D-5L [2, 14– 17], and Mini-Mental State Examination (MMSE) [18] available at least two time points of month 0 and month 3 during the study period were included in the analysis. No exclusion criteria were applied.

The target sample size was a total of 250 individuals which was considered plausible to collect data at those participating six nursing homes.

Data source

Data originally collected every three months between October 2021 and April 2022 were extracted. Data were originally collected via questionnaires that were responded by proxy, such as facility managers, care managers, nurses, or other caregiving staff at nursing homes at 0 month, 3 months, and 6 months from October 2021(baseline).

The data source included demographics and clinical characteristics of participants, including age, sex, independence in activity of daily living owing to dementia assessed using the “Criteria for determination of the daily life independence level of the elderly with dementia” [19], prescribed anti-dementia drugs (i.e., donepezil, galantamine, rivastigmine, memantine, and yokukansan/YiganSan), activities of daily living (ADL) assessed using Barthel Index (BI) [20], occupation/position of the proxies who responded to the EQ-5D-5L+C and EQ-5D-5L [2, 14– 17], EQ-5D-5L+C (proxy version responded by proxy), EQ-5D-5L (proxy version responded by proxy), and MMSE [18].

Criteria for determination of the daily life independence level of the elderly with dementia

The independence in activities of daily living in relation to the presence and symptoms of dementia was assessed using the criteria for determination of the daily life independence level of the elderly with dementia. The criteria are commonly used to assist in identifying individuals eligible for long-term care under the insurance system of Japan. This tool was originally developed to enable healthcare professionals to objectively and quickly determine the level of independence in daily living (i.e., level of care required) for elderly people diagnosed with dementia by a physician [19]. Ranks are categorized as I, II, II a, II b, III, III a, III b, IV, and M. The higher rank indicates the presence of more severe symptoms that interfere with daily life and the more frequent difficulties in behavior and communication.

BI

Assessment of ADL was performed by BI, which measures functional independence in the domains of personal care and mobility such as feeding, moving from wheelchair to bed and return, grooming, transferring to and from a toilet, bathing, walking on a level surface, going up and down stairs, dressing, and continence of bowels and bladder on a 10-item scale [20]. The score ranges from 0 to 100, the lower score indicating higher degree of one’s dependence on assistance from others.

EQ-5D-5L (proxy version)

The EQ-5D-5L was used to evaluate HRQOL. The EQ-5D-5L has a descriptive profile of the five dimensions and the EQ Visual Analogue scale (EQ-VAS). Each dimension has five-level responses, i.e., no, slight, moderate, severe, and unable to/extreme problems [2, 15]. The EQ-5D index score ranges from 0 (death) to 1 (full health), calculated based on value sets developed for the Japanese population [16, 17]. The EQ-VAS, a person’s self-perceived health on a vertical VAS, provides a single score ranging from 0 (worst imaginable health) to 100 (best imaginable health).

MMSE

To assess cognitive function and severity of dementia, the MMSE was used. The MMSE is an 11-item questionnaire which covers five areas of cognitive function (i.e., orientation, registration, attention and calculation, recall, and language) [18]. The maximum score is 30, and lower score indicates higher severity of dementia. The MMSE score is categorized as follows: 30, no cognitive impairment; 26– 29, mild cognitive impairment; 21– 25, mild dementia; 11– 20, moderate dementia; and 0– 10, severe dementia with reference to a previous report [21].

Translation

Regarding the description of the bolt-on dimension “COGNITION (memory, comprehension, concentration, thinking)”, the discussion focused on the translation of “cognition” and “memory”. The “cognition” was translated as “cognitive function” initially at the forward translation stage for easy understanding although neither “cognition” nor “cognitive function” can be familiar in laypersons. Through the discussion with the medical expert, as these Japanese terms, which were not necessarily easy to understand, are interchangeably used even among clinical professionals, this translation was then revised to “cognition” to simply retain the original expression in English. Instead, based on the medical expert’s opinion, the translation of “memory” was additionally supplemented with the expression “forgetfulness”, which Japanese speakers tend to be more familiar with as memory-related symptoms entailed by cognitive impairment and which is even used for as a part of outpatient department name in the hospitals. Thus, the translation was revised as “memory [forgetfulness]”.

In the back translation draft, “cognition” was translated as “cognitive function”. Although the original developer confirmed that there was no problem with this translation as well as the supplemental description of “forgetfulness”, the authors and translators also agreed that “cognitive function” is more appropriate in this context and corresponds to the expression of the status. A provisional Japanese cognition dimension was finalized after “cognition” was revised to “cognitive function” in Japanese.

Pilot test

Five formal or informal caregivers responded to the provisional version of the EQ-5D-5L+C without difficulties in understanding and answering. With this, we confirmed that no change was needed for the provisional version of the EQ-5D-5L+C. The final version of the cognition dimension is shown in the Supplementary Figure 1.

Participant’s characteristics

Baseline characteristics of the participants are shown in Table 1. The mean (standard deviation [SD]) age of the participants was 87.14 (7.29) years, and they were predominantly women (68.9%). The mean MMSE score was 15.76 (8.46), with over 85% classified into mild to severe dementia categories. Donepezil was the most frequently prescribed anti-dementia drug in less than 8% of the participants. Proxies who responded for the participants were primarily registered nurses and other nursing home staff other than facility managers or care managers (39.8% and 46.1%, respectively).

Item analysis

Table 2 shows the distributions of the EQ-5D-5L+C response at three time points. No ceiling or floor effect (>70%) was observed in any of the six dimensions of the EQ-5D-5L+C. Missing data was as small as 1.6% at 3 months at most among three time points.

Validity

For construct validity, the strongest correlation (95% confidence interval, CI) with the MMSE score was found in the cognition dimension (r_s = – 0.640 [CI: – 0.730 to – 0.550]), followed by a strong and moderate correlation in the self-care and usual activities dimensions, respectively (r_s = – 0.530 [– 0.632 to – 0.429] and – 0.497 [– 0.598 to – 0.395], respectively) (Table 3).

Reliability

In a total of 138 and 119 participants, the MMSE category was unchanged at 3 months and 6 months from baseline, respectively. Of these participants, a good level of agreement was observed both between baseline and 3 months and between baseline and 6 months (k = 0.644, 95% CI, 0.541 to 0.746; and k = 0.656, 0.549 to 0.763, respectively) (Table 4).

Sensitivity to change

The mean MMSE score changed by – 1.207 (SD: 5.346) at 3 months and by – 1.149 (5.899) at 6 months from baseline among 237 and 221 participants, respectively (Table 5). A similar but weak correlation between changes in the cognition dimension and those in the MMSE score was observed for both two sets of time points (3 months: r_s = – 0.191, 95% CI, – 0.325 to – 0.058; 6 months: r_s = – 0.267, – 0.396 to – 0.138). Although weak, a correlation with changes in the MMSE score at 3 months was higher when the cognition dimension was added to the EQ-5D-5L compared to the EQ-5D alone (r_s = – 0.142, – 0.275 to – 0.009; versus r_s = – 0.074, – 0.207 to 0.059) (Table 6).