WP05 – Pilot testing and clinical validation of SENATOR in the co-ordination site

Objectives

  • test the performance of the prototype SENATOR software in a cohort of hospitalized older people with multimorbidity.
  • complete any necessary modifications to the prototype SENATOR software in order to progress to WP06 and WP08.

Workpackage Description

Before the prototype SENATOR software can be translated into languages other than English (WP06) or put to use in the planned RCT (WP08), it must be tested and validated in comparison to the application of the assessment scales intrinsic to SENATOR by trained and experienced clinical personnel. For this purpose, 20 selected cases of hospitalized older people with multimorbidity, defined as e 3 chronic medical disorders, will be assessed independently both by two trained, experienced clinical pharmacy staff and by two trained, experienced clinical gerontology staff applying the package of assessments listed in SENATOR. The data from these two pairs of assessors will then be compared directly to the automated data output obtained from application of the prototype SENATOR software. The data will be subdivided into the 10 categories corresponding to the 10 assessment procedures that are inherent in the prototype SENATOR software engine, i.e. (i) correctness of drug indications, (ii) potential for severe adverse drug-drug interaction and drug-disease interaction, (iii) presence of PIMs (STOPP criteria), (iv) presence of PPOs (START criteria), (v) estimated one year mortality risk (CIRS-G score), (vi) a new list of generic drugs, following optimization using the above optimization scales, (vii) a list of least expensive drug brand options, based on current MIMS monthly (30 day) drug ingredient costs, (viii) overall medication appropriateness (Medication Appropriateness Index score), (ix) list of indicated non-drug therapies (ONTOP data); (x) ADRROP, when validated and finalized, constitutes the tenth input component of SENATOR. Technical validation of the SENATOR software prototype is described in WP03. Clinical validation of SENATOR is the essence of WP05, and will be subdivided into 3 key tasks as follows:

Task 1

This will involve assessment of clinical validity, i.e. to show that the performance of the SENATOR software prototype matches that of trained clinical operators applying the 10 assessment components of SENATOR as described. For this task, two trained clinical gerontologists will independently assess 20 selected older multimorbid patient data files, applying the 10 assessment components of SENATOR manually (i.e. without computerized support). A third independent operator will apply the SENATOR software prototype to the 20 patient data files. From this analysis, the reliability of the SENATOR software will be assessed.

Task 2

This will involve inter-rater reliability evaluation of the SENATOR software using physician:physician comparisons. Three pairs of raters will assess 3 separate sets of 20 older multimorbid patient data files, i.e. a total of 60 older patient data files that will prepared in a standardized manner especially for the validation exercise. Inter-rater reliability will be calculated and expressed by means of the Kappa statistic method.

Task 3

This will involve full statistical analysis of the clinical validation exercises and the description of the statistical results.