Design, setting, and participants
This study is a secondary post hoc analysis. As such these analyses are exploratory and should be considered hypothesis-generating rather than confirming. Data were derived from a case–control study of a software application for detecting attention deficits in delirium, and the recruitment and assessment methods are described in detail in the original study . Briefly, patients were recruited from geriatrics and orthopaedics wards of the Royal Infirmary of Edinburgh and Glasgow Royal Infirmary, Scotland. Potential participants were first identified through consultation with the clinical care team. Patients were included if they were aged 65 years and above.
Cases and controls were frequency-matched by age within 10-year age bands and sex. Three groups of patients were recruited: patients with delirium (with or without dementia), patients with a diagnosis of dementia (without delirium), and patients without cognitive impairment. Exclusion criteria were visual or hearing impairments severe enough to preclude testing, non-fluent English speakers and photosensitive epilepsy. The study was approved by the Scotland A Research Ethics Committee.
Five researchers, all psychology graduates, carried out the recruitment and assessment of participants (ND, CC, LS, EN, and ZT). The researchers were fully trained in the use of all cognitive assessments and applying the Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5) criteria . Written informed consent was obtained from patients with sufficient capacity to understand their involvement in the study. In the case of insufficient capacity, an appropriate legal proxy was contacted to provide informed consent.
Assessments and participant groupings
The reference standard assessment for delirium based on DSM-5 criteria was informed by the Delirium Rating Scale—Revised 98 [DRS-R98 ]. As part of the assessment several scales and cognitive tests were administered: The Observational Scale of Level of Arousal [OSLA, ], the Richmond Agitation–Sedation Scale [RASS ], the Short Orientation, Memory and Concentration Test [OMCT ] and the Abbreviated Mental Test [AMT10 ], MOTYB, Days of the Week Backwards, Counting Backwards from 20 to 1, and Digit Span .
The following process for MOTYB was used. Patients were asked to recite the months of the year in backwards order starting with December. Task instructions were repeated once if the participant did not understand the task or if they started reciting the months of the year forward. No further prompts were allowed. Responses to MOTYB were recorded and transcribed (see below). Patients who were unable to engage meaningfully with MOTYB due to severe cognitive impairment and/or disturbances in arousal (as judged by the researchers) were retained for the analysis.
Patients were classified as having dementia through either a prior formal clinical diagnosis of dementia, or if they met DSM-IV criteria for dementia (using information from case notes and informants) as determined by a consultant geriatrician .
Patients for whom an OMCT score > 20 was obtained and who did not have an acute change from baseline or a diagnosis of dementia were grouped as having no cognitive impairment.
Using the above assessments, including MOTYB test results, participants were categorized into three groups: delirium according to DSM-5 diagnostic criteria (with or without dementia), dementia (without delirium), or no cognitive impairment. When grouping was unclear, a decision was sought in discussion with experienced geriatricians (AMJM and DS). When participants could not be classified into one of the predefined clinical groups, a patient was declared as indeterminate and excluded from the analysis.
MOTYB transcript analysis
The five researchers who carried out the recruitment and assessment of participants (ND, CC, LS, EN, and ZT) audio-recorded the MOTYB task. To reduce bias during the transcription process recordings were transcribed blindly by a second researcher who had not carried out the reference standard assessment battery and was not aware of the categorization of the participant. Researcher ND then assessed all transcripts blind to the categorizations and recorded the type of errors made.
Based on the framework used by Meagher et al. , the following responses were considered errors: omissions, commissions in the wrong place, non-relevant commissions and repetitions. Following inspection of the transcripts, we added the following response patterns: self-corrections, reciting the months forward, stopping part way through the task and not being able to meaningfully engage with the task (see online Table 1 for examples).
Descriptive and inferential statistical analyses were conducted using IBM SPSS Statistics version 22 (IBM, Inc., Chicago, IL). Fisher’s exact tests and Mann–Whitney U tests were used to analyse group differences. Holm–Bonferroni corrections were applied to account for multiple comparisons .