Bastiaan Sallevelt

132 CHAPTER 2.3 SUPPLEMENTARY INFORMATION SI1 Reading Instruction & Implementation Guideline The 34 START and 76 STOPP algorithms described in the paper are supplied as a single Excel spreadsheet (.xlsx) of the original publication: Huibers CJA, Sallevelt BTGM, de Groot DA et al. Conversion of STOPP/START version 2 into coded algorithms for software implementation: A multidisciplinary consensus procedure. Int J Med Inform. 2019 May;125:110-117. doi: 10.1016/j. ijmedinf.2018.12.010. In order to implement these algorithms, please follow the steps described in this document. For additional file formats (such as XML or JSON), please contact the authors ( Each criterion is encoded as a separate decision table, which leads to one or more inference rules per criterion. These rules are meant to run on a dataset composed of a single patient’s health record, including his or her episodes, medicines, and measurements. These values are expected to be complete and accurate. Each criterion has a number of columns containing metadata, conditions, and actions. These columns span five rows each and are formatted as such: Table SI1.1. Sample rule with Metadata, Condition, and Action columns. Name Sample Metadata Sample Condition Sample Action Type METADATA [ADDITIONAL] CONDITION ACTION Object ID Episode exists Medicine Attribute value icd10 atc Operator equals (=) equals (=) start if not present Description Criterion ID Atrial fibrillation Vitamin K antagonist Value START A1 I48 B01AA01