BEHAVIORAL
There are various ways to calculate accuracy of prediction. In the
analysis shown here, patients who were rated as low or low-medium risk on any risk measure were categorized as low risk. Patients who were rated as medium, medium-high, high, or very
high risk were categorized as high risk. This dichotomous categorization allowed different risk-measure ratings to be compared.
Patients rated as high risk who later were discharged from opioid
Because of the increasing pressure bearing on the issue of opioid
prescribing, certain clinical practices are indicated when using written risk assessments. Clinicians should give risk assessment questionnaires or surveys to patients with the explanation, “this is not
a process to determine whether opioids are to be offered.” Instead,
risk assessments should be used to determine what sorts of opioids
should be used and what degree of treatment monitoring is indicat-
Table Summary of Risk Assessment Comparisons: Correct Prediction of Discharge from Opioid Treatment
Risk Assessment Tool
Moore 2009
N=48
Jones 2011
N=51
Jones 2012—Study 1
N=132
Jones 2012—Study 2
N=263
Jones 2013
N=196
Clinical interview,
experienced
77%
—
70%
71%
76%
Clinical interview,
inexperienced
—
—
—
43%
—
SOAPP
73%
—
—
—
—
SOAPP-R
—
—
32%
39%
60%
ORT
45%
30%
29%
35%
48%
ORT by interview
—
57%
—
—
—
PMQ
—
—
44%
51%
—
DIRE
17%
—
—
—
—
DIRE = Diagnosis, Intractability, Risk, Efficacy ORT = Opioid Risk Tool PMQ = Pain Medication Questionnaire
SOAPP = Screener and Opioid Assessment for Patients with Pain SOAPP-R = SOAPP-revised
treatment for medication aberrant behavior were counted as a correct prediction. Likewise, patients who were rated as low risk but who
later were discharged from opioid treatment for medication aberrant
behavior were counted as an incorrect prediction. The rates of correct
prediction for each measure in each study are shown in the Table.
These data indicate that, in our population, most written risk measures
were generally poor at identifying persons who later were discharged
for medication aberrant behavior or violation of the treatment agreement. In our studies, a clinical interview by an experienced clinician
was the best method to identify future discharge from treatment.
ed for each patient in the early stages of treatment. In other words,
risk assessment should not be a pass-fail test for the use of opioids.
Opioids may be able to be used safely with individua