4.2.1 The role of technologies for behavioral observation in OUD
In addition to simple observation of pain-related behaviors, more
objective tools have been employed to quantify several behaviors
commonly associated with pain in individuals with OUD. Of note, an
important limitation is that the following studies did not specifically
assess pain, and therefore, these results need to be extrapolated with
caution.
In a human laboratory study, Teeters and colleagues randomized 39
individuals with OUD to either a 15-minute laboratory stress or
no-stress condition followed by exposure to opioid
cues130. Opioid craving was measured, using a craving
VAS before and after exposure to opioid cues, and sleep duration using
the Pittsburgh Sleep Quality Index131 and actigraphy.
The study found that participants in the no-stress control group who
reported shorter average nightly sleep duration had higher levels of
opioid craving following opioid cue exposure. These suggest that poor
sleep increases vulnerability to opioid craving, which could increase
the perceived need for opioids to manage pain. As a final example,
Salgado GarcĂa and colleagues analyzed biosensor data from 46 patients
who underwent dental surgery and received opioids after extraction.
Based on metrics such as skin conductance and accelerometry, machine
learning models could identify periods in which patients were using
opioids with an accuracy of up to 83.7%132.
Lambert and colleagues used sternal accelerometers to monitor
involuntary movements in 23 patients undergoing opioid withdrawal, which
has substantial clinical overlap with the pain
experience133. The study revealed that patients
exhibiting sinusoidal wave patterns in their accelerometry data,
indicative of periodic leg bouncing and foot tapping, had worsening
withdrawal symptoms, measured using the COWS. This finding suggests that
accelerometry can identify those at risk of worsening opioid withdrawal.
Bertz and colleagues also utilized actigraphy watches and electronic
diaries to assess the sleep of 37 patients with OUD undergoing methadone
or buprenorphine treatment134. Their findings
suggested that patients experienced shorter sleep periods and delayed
sleep timing during periods when they used non-prescribed opioids and
cocaine based on urine drug screens. This suggests the potential for
actigraphy to be useful in detecting return to non-medical substance
use. Therefore, the ability to monitor behaviors quantitatively over
time may also provide insights about pain trajectories and volatility,
which have also been found to predict non-medical opioid use among
persons with OUD135.