Type

Range

Units

Status

Start Date

End Date

Financing

Study Design

Brief description/objective

Introduction: To evaluate current delivered multidimensional interventions in the context of palliative cancer care, prospective checklist work (Temel NEJM 2010), video-/audio-tape of consultation (Jenkins VA JCO 2011) or retrospective chart reviews (Mack JW ASCO 2011/#6026) can be applied. We developed a practical tool to extract symptom management performance in real practice settings of oncologist in a multicentric setting (E-Mosaic trial / SAKK 95/06). We aim to explore the inter-rater reliability of data extraction and quantify types of interventions retrieved.
Methods: Based on a comprehensive overview of key palliative care interventions (text books, contents of professional meetings, published chart review tools in palliative care) we developed a checklist with 6 categories of main symptoms: 1/pain, 2/fatigue, 3/anorexia, 4/dyspnea, 5/depression/anxiety, 6/nausea. In addition other symptoms can be checked (n=19). In each category interventions can be mentioned (pharmacological, interventional, non-pharmacological [counseling, referral to specialist], diagnostic). For medication it can be sub-specified: new start, stop, increased, reduced, dose maintained, dosis per need. In order to be assigned to a specific symptom, either the physician mentioned it explicitly in the chart or the intervention is undoubtful assigned to one symptom. For inter-rater reliability 10 charts were blindly reviewed by two oncologists/palliative care specialist in addition to the main investigator. Agreement was checked for overall interventions, for main symptom category interventions and for pharmacological, interventional, diagnostic.
Results: 140 charts have been reviewed. pharmacological interventions are well documented with exception of maintaining the dose lacking often justification. Counseling or communication interventions are found in approximately in 50% of patients, however substantial differences between centers involved. Exploratory inter-rater reliability reveals substantial agreement for the main categories, also for communication interventions. Diagnostic interventions potentially, but not explicitly associated with symptom control (e.g. hemoglobin, chest x-ray) could only be included if explicitly stated by the physician. Final analysis is in progress.
Conclusion: Consistent with others work chart review of unselected oncologist seems reliable for main symptom management performance, however with substantial variation between centers. Future work may require basic training for documentation or also checklist for intervention done prospectively in clinical trials.