Ten participants (3 males, 7 females) with a mean age of 90 (SD = 5.80; Table 1) were recruited from a specialized dementia long-term care home. Age, dementia diagnosis, mental health diagnoses (if any), and current medications were extracted from records. Participants’ cognitive functioning was assessed using the Montreal Cognitive Assessment (MoCA) [11], which confirmed that all participants had cognitive impairments (indicated by a MoCA score of < 26/30). The Hospital Anxiety and Depression Scale anxiety subset (HADS-A) [12] was used to assess participants’ anxiety, with a score of greater than 10/21 considered abnormal. This measure was chosen as it is a short assessment used previously to evaluate self-reported anxiety in PWD [13]. HRV data were collected using the Zoom HRV wrist-worn monitor. A wrist monitor was used as it is completely non-invasive and easily portable [7]. The data were saved to the Elite HRV iPhone application. Kubios HRV Standard software was used for the analysis of the RR-interval data.
Insert Table 1 about here
Data were collected at the care home, in either a small library or lounge. The wrist-worn monitor was placed on participants’ non-dominant wrist to reduce the residual effects of daily activities frequently performed using the dominant hand [14] on the data. Participants were asked to sit in an upright position, keep their eyes open, and remain still and quiet during this portion of data collection. Despite these instructions, most participants moved occasionally or engaged in conversation with a staff member or volunteer. RR-interval recordings were taken after participants had been seated for approximately one minute. Two recordings of RR-interval data were collected from each participant.
Following HRV data collection, the Hospital Anxiety and Depression Scale (HADS) statements and response options were read aloud to the participants as many of them had poor eyesight which interfered with their ability to read the statements. Participants were asked to articulate their response to each statement aloud. Since the HADS assessment contained statements regarding sensitive subject matter, special attention was paid to the participants’ well-being during the administration of this assessment. Subsequently, a second researcher administered the MoCA as per the administration guidelines.
Data analysis
Two shorter epochs, rather than one longer one, were collected due to the limitation of the Zoom HRV device and to minimize non-stationarities in the data [15]. These epochs ranged from 65 seconds to 156 seconds (M = 126.1). Prior to analysis, a few of the epochs were shortened due to signal loss. The range of epoch times remained the same, but the mean differed slightly (M = 125.5).
Artefact correction was performed using threshold-based artefact correction. Each epoch was corrected using a different threshold strength, as there was variation in the quality of the RR-interval data between different participants and epochs. The lowest strength threshold that allowed for the identification and correction of the most artefacts was chosen. In a majority of the epochs, the percentage of beats corrected did not exceed 5% [16]. Some epochs required more rigorous corrections. Just over 5% of beats were corrected in two of the epochs [17]. The artefact corrected data were detrended using the smoothn priors method [18] with a lambda value of 500 [16]. The data were then analysed in the time domain to produce RMSSD values. The two RMSSD values per participant, in units of milliseconds (ms), were prorated to equal time intervals of 60 seconds to correct for differing epoch durations [19] and averaged so each participant had one RMSSD value.
The Shapiro-Wilk test was run for the averaged, prorated RMSSD values to ensure normal distribution. The data were not normally distributed (p < 0.05). To correct for this, the data were transformed using the natural log function (ln) [6]. The log transformed RMSSD (lnRMSSD) values were normally distributed, W(10) = 0.927, p > 0.05.
To our knowledge, there are no data reporting average HRV for the dementia care home population to use as a direct means of comparison. Table 2 summates data from past studies investigating time-domain HRV in older adults of a variety of ages and physical capabilities.
Insert Table 2 about here