| Anupriya Lakshmipathy |
The Columbia-Suicide Severity Rating Scale (C-SSRS) questionnaire is widely used in Neuroscience studies to determine the presence of suicidal ideation or behavior. The first administration of the CSSRS within a clinical trial is conducted with respect to one or more pre-dose reference periods (PRP) (e.g., lifetime, last 6 months), with subsequent administrations conducted relative to the last visit.
Information is collected on three broad categories: suicidal ideation, suicidal behavior, and non-suicidal self-injurious behavior, with the former two categories typically analyzed as ordinal measures, each containing 5 levels. When treatment-emergent events are of interest, relative to the given PRP, it is inappropriate to include subjects for those ordinal levels for which that subject experienced that given event. For example, if the PRP is ‘lifetime' and the subject experienced Suicidal Ideation at some point during their lifetime, then that subject is not able to experience a treatment-emergent event of Suicidal Ideation, and thus must be excluded from the observed Suicidal Ideation counts and percents. This accounting results in the potential for different denominators for the different ordinal levels within a given category. However, when considering the ordinal levels in a "cumulative" way, i.e., to ascertain how may subjects had an event "as bad or worse" than the given ordinal category, all denominators should be the same across all ordinal categories. This combined accounting presents for unique programming.
This paper emphasis on deriving ADCSSRS dataset and event summary display to support the unique analysis of the ordinal levels in assessing patients suicide risk. |