Examining Clinician Differences that Influence Ratings of the
Probability of Harm and Rehospitalization on a Standard Dangerousness
Assessment
Christina M. Harris
University of Nebraska –
Lincoln
|
|
Discussion
The
present study examined how clinician differences interact to produce different
ratings of harm and rehospitalization on a
standardized forensic assessment of dangerousness in a large sample of Division
41 of the American Psychological Association. This study hoped to expand upon
the little research in the field therefore this analysis was more exploratory
than focused on hypothesis testing.
A
promising finding for both probabilities of harm and rehospitalization
ratings -and consistent with what we expected to find- is that there was not a
lot of variability among scores suggesting fairly good inter-rater reliability
among clinicians across facility, usefulness ratings, and frequency of giving
assessments of danger. This finding also supports Harris, Boccaccini,
and Murrie (2014) notion that standardized
assessments minimize the variability of scores. Interestingly, both
probabilities of harm and rehospitalization ratings
were equivalent among those who worked in a forensic facility, regardless of how
useful or frequently they gave assessments of danger. Results only reveal
differences in probability of harm and probability of rehospitalization
ratings among those who do not work in a forensic facility. This could be a
reflection the different training required of forensic professionals and the
effectiveness of that training when it comes to standardized assessments. This
finding could also support Reid’s (2003) claim that even competent clinical
professionals can’t gain the necessary forensic skills needed to run a proper
assessment in just one weekend class or after having read a book on the subject
and also that forensic practice is not simply an extension of clinical
practice. Further research is needed however these findings have implications
that point to keeping forensic assessments among forensic professionals as a
way to minimize clinician differences.
As
hypothesized, there were differences in probability of harm ratings between those
who work in a forensic facility and do not however the effect was only present
among those who gave assessments weekly. Interestingly, whether or not the
forensic professional gave higher or lower probability of harm ratings depended
on their usefulness rating. If the respondent found the assessment to be of
little to no use and worked in a forensic facility they gave higher probability
of harm ratings. Opposite of this, if the respondent found these assessments
somewhat to very useful then those who work in a forensic facility gave lower
probability of harm ratings.
There was
an interaction of usefulness ratings and facility on probability of harm
ratings and consistent with the above results that probability of harm ratings
were equivalent among those who worked in a forensic facility regardless of
usefulness. However among respondents who do not work in a forensic facility,
those who believe these assessments to be useful gave higher probability of
harm ratings than those who don’t find them. This findings are a expansion of the findings of Jensen-Doss & Hawley
(2010) that opinions of usefulness not only affect how often a clinician uses
an assessment but also opinions of usefulness can affect how an assessment is
scored, in this case through probability of harm ratings.
For
probability of rehospitalization ratings there were
also differences among forensic and clinical professionals but only if the
respondent gave assessments of dangerousness weekly and found these assessments
of no use such that those who worked in a forensic facility gave higher
probability of harm ratings than those who did not. Probability of rehospitalization ratings showed much less variability than
probability of harm ratings.
These
findings contribute to a small body of research that needs to be expanded on
and is very important to study as forensic assessments of dangerousness are so
prevalent in numerous clinical settings. Not only are forensic assessments
prevalent but they have a large impact on the
individual who is being assessed such that they can impact sentencing to their
release from a hospital. Since patients are typically only allowed one
evaluation, it is important that clinician differences have as little of an
impact on the scoring of these assessments/evaluations as possible. These
findings show that while there was little variability, there are still
clinician differences that affect the scoring of a dangerousness assessment.
These findings suggest that opinions of usefulness of assessments of
dangerousness assessments can cause different probability ratings of harm and rehospitalization and perhaps it is important for
clinicians to remain educated about the evidence there is which supports the
utility of standardized assessments so as to minimize differing opinions of
usefulness with the goal of getting professionals on the same page. These
findings also suggest that the frequency at which a respondent gave assessments
of danger only had an effect on ratings when the respondent was giving
assessments very frequently (weekly) and more research is needed to examine the
effect that frequency has on ratings since it was a new variable included in
this analysis.
There are
some limitations to the study. Although the study used actual cases, it was not
conducted in controlled setting and therefore may lack internal validity and
there may be confounds working. The study also did not collect demographic
information such as ethnicity or the location of the respondent so it is
difficult to assess generalizability and if regional and cultural differences
would have had an effect.
As there
is shortage of research regarding clinician differences in violence risk
assessment, this study may serve as a beginning for more specific study in what
constitutes difference ratings of violence as result of clinician differences.
Further research is needed to gain insight on implications of moving forward in
a manner that reduces clinician differences in scoring violence assessments.
Future directions for study might include further examining the role the
frequency a respondent issues assessments has on scoring violence assessments.
Future research might also add an element of accuracy as well to see if
differences indicate better or worse predictions of violence.
Index Introduction Methods Results Table 1 Table 2 Table 3 Figure 1 Figure 2 References