Table 3 Input parameters for %diag_test macro

data

name of input dataset

truthvar

name of reference/truth/gold standard variable e.g., abbott_plasma_vl

truthcutvalue

cutoff value to use to categorize values of test variable as having disease or not e.g., 1,000 so that if value ≥ 1,000 then presence of disease, otherwise absence of disease

testvarlist

list of diagnostic test variable(s) separated by space e.g., vdbs_vl mdbs_vl ddbs_vl

testcutvalue

cutoff value to use to categorize values of test variable as having disease or not e.g., 1,000 so that if value ≥ 1,000 then presence of disease, otherwise absence of disease

domain

(optional) domain variable for sub-population analysis

domainvalue

(optional) value of domain/sub-population of interest (should be numeric). Required if domain is specified

condition

(optional) any conditional statements to create and or fine-tune the final analysis dataset specified using one IF statement

outputdir

path for directory/folder where output is saved

tablename

short name of output table

tabletitle

title of output table

surveyname

(optional) abbreviation for survey/study to be included in the output

decimalpoints

(optional) desired number of decimal places for each estimated measure (default = 1)

alpha

(optional) desired level of significance (default = 0.05, for 95% confidence intervals)

missvaluelabel

(optional) value label for missing values. If missing data have a format, it should be provided, otherwise macro assumes the default format “.”

varmethod

(optional) method for computing confidence intervals. Included here are: “Normal” (default) for normal distribution approximations, “Wilson” (most desired) for Wilson score and “Exact” for exact binomial approximation

print

variable for displaying/suppressing the output table on the output window which takes the values (NO = suppress output, YES = show output)