This page describes indices we developed as part of our Cancer and Comorbidities project and Multimorbidity projects. The indices are intended to allow adjustment for comorbidity when considering survival outcomes for patients with cancer (the C3 and PBCI indices) and for use in more general settings not specific to cancer (the M3 and P3 indices). All indices have been validated and published in peer-reviewed journals.
Visit our indices publications in OUR Archive
This page provides links to the main validation papers for each index, along with code for scoring the various indices.
Acknowledgements:
The code provided below was created by James Stanley, using original code that with lead roles or contributions from Diana Sarfati, James Stanley, Jason Gurney, Clare Salmond, and Jane Zhang (see published papers for fuller details of involvement of individual researchers).
Both projects were funded by the NZ Health Research council (HRC ref 10/496 and 14/173 respectively)
Publications (with details on the methods used in development and results of validation)
C3 index
Sarfati, D., Gurney, J., Stanley, J., Salmond, C., Crampton, P., Dennett, E., Koea, J. & Pearce, N. (2014b). Cancer-specific administrative data-based comorbidity indices provided valid alternative to Charlson and National Cancer Institute Indices. Journal of Clinical Epidemiology, 67(5), 586-595. doi: 10.1016/j.jclinepi.2013.11.012
PBCI index
Sarfati, D., Gurney, J., Stanley, J., Lim, BT, McSherry C (2014). Development of a pharmacy-based comorbidity index for patients with cancer. Medical Care, 52, 586-593. doi: 10.1097/MLR.0000000000000149.
M3 index
Stanley J, Sarfati D. (2017) The new measuring multimorbidity index predicted mortality better than Charlson and Elixhauser indices among the general population. Journal of Clinical Epidemiology 2017;92:99-110. DOI: 10.1016/j.jclinepi.2017.08.005
Gurney J, Stanley J, Sarfati D. The M3 multimorbidity index outperforms both Charlson and Elixhauser indices when predicting adverse outcomes in people with diabetes. Journal of Clinical Epidemiology 2018; 99: 144-52.
P3 index
Stanley J, Doughty RN, Sarfati D. A Pharmaceutical Dispensing-based Index of Mortality Risk From Long-term Conditions Performed as well as Hospital Record-based Indices. Med Care 2020;58(2):e9-e16. DOI: 10.1097/MLR.0000000000001217.
Macros and functions:
SAS code is currently available for the C3 and M3 indices. The file is one SAS macro that can score C3 or M3 for a given dataset. The macro can also code for the Charlson and Elixhauser indices (see details in the M3 index paper cited above as to how these are coded, with respect to ICD codes used and weights applied).
R code for scoring the M3 index is available – note that this code does not currently allow for scoring of Charlson or Elixhauser indices.
SAS and R code for the PBCI and P3 indices is currently planned for development in early 2023. Please contact james.stanley@otago.ac.nz for more information.
C3 & M3 Code (with example data layout files)
The current files are available from the following Dropbox folder: https://www.dropbox.com/sh/q4alo89ahnzjkke/AABGOQ_74XFARbrlujSQOi5Qa?dl=0. This contains the SAS macro for scoring the C3 and M3 index (currently version 1.0); two separate function scripts for R for the C3 and M3 indices (currently both version 1.0); notes on setting up data for the macro/function (separate files for SAS and R), and format of output (Word document); and an example file illustrating potential data layouts for the macro/function (Excel file).