Before implementing the R&M in routine ward rounds, the head of the hospital's geriatric unit (RO) gave the staff (nurses and residents) comprehensive instructions and training in the administration of their respective tests. An R&M field was added to the electronic medical records. At the start of RADAR implementation, a geriatric nurse conducted spot scans of the nurses' records to verify they were filled out as required. Six months after R&M had been implemented in the department's routine practice, the 4AT was administered by a resident in geriatrics with previous experience with this test (MS), as part of the implementation process, to ratify the reliability of R&M. The results of the 4AT were also entered into the patients' medical records. The current manuscript reports the results of a retrospective analysis of the data extracted from the patients' records, performed after a sufficient number of patients underwent both tests.
Subjects: Patients aged over 70 years, hospitalized in the internal medicine wards in the Bnai-Zion Medical Center, a municipal hospital affiliated with the faculty of medicine, were included in this study. Patients with advanced dementia and disability referred from nursing homes, terminally ill patients and patients unable to cooperate due to language barriers were excluded. The study was approved by the Human Investigations Review Board of Bnai Zion Medical Center, Haifa, Israel.
Delirium tests: The R&M test was administered to each patient by ward staff during the morning hours, followed by the 4AT within 4 hours, at noon. Unlike the R&M, which was administered routinely every day, the 4AT was performed randomly 1–3 times a week on weekdays. The 4AT was only administered once in a given patient, sometimes on the day of admission, but usually a few days into their stay.
The R&M test was performed in 2 stages: first, the RADAR test was administered by the nurse based on her observation of the patient while dispensing the morning drugs, as previously described (22, 23). In the electronic medical records, nurses answered "yes" or "no" to 3 questions: 1. Was the patient unusually sleepy? 2. Did the patient have difficulty following instructions? 3. Were the patient's movements unusually slow? If the answer was "yes" to 1 of the questions, the RADAR test was considered positive and a red flag appeared in the resident’s field of the patient's electronic medical records, indicating possible delirium.
Every patient whose records indicated a positive RADAR test was evaluated by their treating physicians using the MOYB during their routine morning rounds to confirm the presence of cognitive impairment. An abnormal MOYB confirmed the presence of an attention/concentration deficit, and was used, together with the nurse's assessment, to diagnose the presence of cognitive disturbance.
Finally, based on information provided by the patient’s relatives, friends, or caregivers (for new admissions), or on assessments from previous days, the cognitive decline was marked as "chronic" (usually dementia) or "new onset" – indicating delirium.
Performing 4AT was started after a sufficient test-run of R&M, and used to validate the reliability of the new test. The 4AT (Table 1) was performed as previously described (14). Only records that included all required R&M data obtained within the predetermined 4 hour time-span were used. The ward staff’s R&M scores were not accessible to the resident who performed the 4AT. However, given the brevity of his visit, he was often unable to determine whether mental decline was a new development or was pre-existing, so the principal investigator clarified this after the 4AT in the case of patients diagnosed with cognitive impairment.
Table 1
4AT parameters and scoring
(adapted from https://www.the4at.com).
4AT parameters | evaluation | scoring |
---|
Alertness | drowsiness/sleepiness | normal or mild brief sleepiness 0 clearly abnormal 4 |
AMT4 | age, date of birth, place, current year | no mistake 0 1 mistake 1 2 or more mistakes 2 |
Attention (MOYB) | ability to say the months of the year in backward order, starting with December | > 6 months 0 < 7 months 1 Untestable 2 |
Acute change in cognition | acute or fluctuating change in mental function | no 0 yes 4 |
AMT4 – Abbreviated Mental Test; MOYB – Months Of the Year Backwards Test. |
Scoring key: acute (including fluctuating) change in cognition was considered mandatory for the diagnosis of delirium. Accordingly, a score > 4 in the presence of a new change in cognition was defined as delirium. |
Statistical analysis. Results were presented in binary form - positive or negative for delirium. In addition to new onset or fluctuating course, both positive RADAR and MOYB were required for a positive R&M test. For 4AT, the "gold-standard" in this study, a score of at least 1 in addition to the 4 points scored for the acute event (i.e. a total of 5 points or more) was required for the diagnosis of delirium. From a clinical point of view, a high sensitivity (> 90%), and therefore a low false negative rate (low β error) was considered to be most important to avoid overlooking patients with delirium. The sample size required for a sensitivity of 95% and a 95% confidence interval was calculated according to Buderer et al. (25), and was n = 365, assuming delirium prevalence of > 20% in elderly patients hospitalized in internal medicine wards (3). In addition to R&M sensitivity, specificity, and positive and negative predictive values, the Pearson correlation, Altman-Bland plot, non-inferiority test and ROC analysis were used to compare R&M and 4AT.