Measuring Hot Flashes via Skin Conductance
Skin conductance, the tonic level of electrical conductivity of skin and an indirect
measure of sweating, was found to be the most sensitive and specific indicator of
hotflashes in controlled laboratory studies (Carpenter et al. 2005 ; de Bakker and
Everaerd 1996 ; Freedman 1989 ; Miller and Li 2004 ). As opposed to palmar and
plantar measures of skin conductance that are responsive to psychological stimuli,
SCL is fairly inactive on the sternum under a normal variety of conditions (Rickles
and Day 1968 ). Thus, sternal measurement of SCL provides a more sensitive
measure of hotflashes than palmar measurement (de Bakker and Everaerd 1996 ).
Hotflashes show a sharp and rapid rise on a SCL trace followed by a sloping return
to baseline, or“swishy tail,”that can be distinguished from the“sawtooth”pattern
characteristic of activity or other sweating-related artifact (Carpenter et al. 1999 ).
Initial investigation into an ambulatory measure of skin conductance for the
measurement of hotflashes was performed using an analog system worn as women
went about their daily activities (Freedman 1989 ). This measure was further
developed and digitized in a commercially available device, the Biolog™
(Carpenter et al. 1999 ), a monitor worn around the waist with integrated hotflash
reporting where hotflashes are reported by pressing two buttons simultaneously on
the device which then date and time-stamps the self-report of the hotflash. This
system has been used in menopause studies over the past two decades. In initial
analyses, the monitor was deemed an acceptable and feasible method for assessing
hotflashes over 24 h (Carpenter et al. 1999 ). In an effort to improve the wear and
comfort of the hotflash monitors, another device, the Bahr Monitor™, was
designed (Webster et al. 2007 ) as a smaller device (6×6 cm) attached to the
sternum by a single electrode patch. Data on the performance of this device are still
emerging (Stefanopoulou and Hunter 2013 ; Carpenter et al. 2012 ; Mann and Hunter
2011 ; Webster et al. 2007 ).
Initial investigations into the physiologic measurement of hotflashes showed
high agreement between physiologic measurement and self-reported hotflashes
(Freedman 1989 ). However, one of the major issues facing accurate detection of hot
flashes is the algorithm used to classify hotflash events from physiologic signals.
The mostly widely used algorithm for sternal skin conductance measures of hot
flashes is a≥ 2 μmho rise in 30 s, which in both early studies (Freedman 1989 ) and
later laboratory investigations (de Bakker and Everaerd 1996 ) showed high sensi-
tivity and specificity relative to participant self-report. However, subsequent labo-
ratory and ambulatory studies have shown somewhat lower sensitivity using this
criterion (de Bakker and Everaerd 1996 ; Hanisch et al. 2007 ; Sievert 2007 ; Sievert
et al. 2002 ). Alternate thresholds for detecting hotflashes have been proposed,
including a threshold of 1.78μmho (among prostate cancer patients; Hanisch et al.
2007 ) and a 1.2μmho SCL rise (among breast cancer patients; Savard et al. 2014 ).
Others have used the 2μmho criterion with additional subjective coding to classify
events that do not meet the criterion (Thurston et al.2009a, 2011 ). As a single
threshold cannot accommodate between-subject variability in hotflash-associated
11 Hot Flashes: Phenomenology and Measurement 239