Reference

variables

Contains all the variables needed for algorithm development and implementation

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable	string
role	The different processes that the variable was part of during the algorithm development	string
variableType	The statistical type of the variable	category	Categorical: Categorical type Continuous: Continuous type
databaseStart	The databases that the variable can be harmonized from	string
variableStart	The names of the variable for each starting database	string
units	The units for the variable	string
label	A short description of the variable	string
labelLong	A long description of the variable	string
description	Should contain any issues that need to be taken into account before using this variable	string

variable-details

Contains the details for the variables defined in a variables sheet

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable	string
dummyVariable	The name of the dummy variable. Not valid if the variable is continuous.	string
typeEnd	The type of the variable	category	cat: Categorical type cont: Continuous type
databaseStart	The starting databases whose harmonization details this row contains	string
variableStart	The names of the variable for each starting database	string
typeStart	The type of the starting variables	category	cat: Categorical type cont: Continuous type N/A: Used mainly for derived variables when there are many start variables
recEnd	The value that the variable will be harmonized to. Can be a number or an interval.	string
catLabel	The description for the category.	string
catLabelLong	A longer description for the category.	string
numValidCat	The number of valid categories for the variables.	number
units	The units for the variable. Only for continuous variables.	string
recStart	The value of the starting variable to harmonize from. Can be a single value or an interval.	string
catStartLabel	The description for the category of the start variable.	string
variableStartShortLabel	A short description of the starting variables	string
variableStartLabel	A long description of the starting variables	string
notes	Any issues/problems that users of this variable should consider. Can also include any problems encountered during harmonization.	string

lookup

Contains the score range that each bin belongs to

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
catValue	The bin number	number
range	The range if score values for this bin. Should use the mathematical range notation.	string

descriptive

Contains descriptive information, for example mean, median etc. for variables

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
variable	The variable whose descriptive statistics the row contains	string
catValue	The value of the category whose descriptive statistics the row contains. N/A for continuous variables.	number
n	The number of individuals used to calculate the row’s statistics	number
proportion	The proportion of all the individuals in the study used to calculate this row’s statistics	number
median	The median statistic	number

model-export

Contains the list of all files that are part of the export for an algorithm

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
fileType	The type of file	category	variables: A variables file variable-details: A variable details file descriptive: A descriptive lookup file descriptive-bins: A descriptive bins file model-export: A model export file model-steps: A model steps dummy: A dummy file center: A centering file rcs: An RCS file interaction: An interaction file fine-and-gray: A fine and gray file cox: A cox file survival_bins: File with survival data for the different bins lookup: Contains the range of score values for each biin in a survival algorithm validate: Contains the validation rules for variables in the algorithm tables: Defines a tables model export file
filePath	The path to the file relative to the model export file	string

model-steps

Contains the steps to score an algorithm

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
step	The type of step to perform	category	dummy: Step used to create dummy variables center: Step used to create centered variables rcs: Steps used to create spline terms interaction: Step used to create interaction terms fine-and-gray: Step used to calculate the outcome of a fine and gray model cox: Step used to calculate the outcome of a cox proportional hazards model simple-model: Step used to calculate the outcome of a simple model logistic-regression: Step used to calculate the outcome of a logistic regression model
fileType	The type of file referenced in the file path column	category	N/A: Missing since certain steps don’t need to specify the file type beta-coefficients: A beta coefficients file for a fine and gray model or a cox model baseline-hazards: The baseline hazards for a fine and gray model or a cox model
filePath	The path to the file relative to the model location of the model steps file	string
notes	Any notes for the future	string

dummy

Contains the variables to convert to dummy variables

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
origVariable	The name of the variable to dummy	string
catValue	The category value in the original variable which the dummy variable represents	string
dummyVariable	The name of the dummy variable	string

center

Contains the variables to create centered variables

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
origVariable	The name of the variable to center	string
centerValue	The value to center with	string
centeredVariable	The name of the new centered variable	string
centeredVariableType	The type of the new centered variable	category	cat: Categorical type cont: Continuous type

interaction

Contains the interaction variables for an algorithm

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
interactionVariable	The name of the interaction variable	string
interactingVariables	The names of the variables that are part of this interaction variable	string
interactionVariableType	The statistical type of the interaction variable	category	cat: Categorical type cont: Continuous type

rcs

Contains the RCS variables for an algorithm

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable which will be splined	string
rcsVariables	The names of the spline variables in correct order	string
knots	The knot values to use seperated by a semi-colon	string

validate

Contains the validation rules for variables in the algorithm

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable to validate	string
rule	The type of validation rule to apply to a variable	category	type: Validate the data type of the variable range: Validate that the variable is within a range of values. Meant for continuous variables. allowed: Validate that the variable contains a value from a set of valid values. Meant for categorical variables. nullable: Whether missing values are allowed for a variable
value	The value to use when applying the validation	string
error_handle	How to handle failed validations	category	error: Failed validations for the variable should throw an error, stopping the scoring process warning: Failed validations for the variable should log a warning and continue the scoring process truncate: Failed validations for the variable should log a warning, truncate the failed variable value, and continue scoring process
error_replace	Value to replace variables that fail validations whose errorHandle value is warning	string
location	Which step in the scoring process the validation should be used	string

beta-coefficients

Contains the coefficients for regression models

Algorithm type(s): cox, fine-and-gray, logistic-regression

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable whose beta coefficient the row contains. If this is the coefficient for the intercept then use the name Intercept	string
coefficient	The beta coefficient	number
type	The statistical type of the variable	category	cat: Categorical type cont: Continuous type

baseline-hazards

Contains the baseline hazards to use with a cox proportional hazards model or a fine and grey model

Algorithm type(s): cox, fine-and-gray

Columns

Column Name	Description	Type	Category Values
time	The time upto which the baseline hazard should be used	number
baselineHazard	The baseline hazard value	number

survival_bins

Contains the survival data for the different bins in a survival algorithm

Algorithm type(s): survival

Columns

Column Name	Description	Type	Category Values
catValue	The bin number	number

lookup

Contains the range of score values for each bin in a survival algorithm

Algorithm type(s): survival

Columns

Column Name	Description	Type	Category Values
catValue	The bin number	number
range	The range if score values for this bin. Should use the mathematical range notation.	string

tables

Contains the list of tables referenced in the variable and variable detail files

Algorithm type(s): all

Columns

Column Name	Description	Type	Category Values
tableName	The name of the table	string
tablePath	The path to the table relative to this file	string

simple-model

Contains the metadata for a simple model

Algorithm type(s): simple-model

Columns

Column Name	Description	Type	Category Values
name	The name of the metadata	category	outputVariableName: The name of the output variable for the model. Should be defined in the variables and variable details sheets.
value	The value of the metadata	string

logistic-regression

Contains the beta coefficients for a logistic regression model

Algorithm type(s): logistic-regression

Columns

Column Name	Description	Type	Category Values
variable	The name of the variable whose beta coefficient the row contains. If this is the coefficient for the intercept then use the name Intercept	string
coefficient	The beta coefficient	number
type	The statistical type of the variable	category	cat: Categorical type cont: Continuous type