. 2020 Jul 1;15(7):e0234845. doi: 10.1371/journal.pone.0234845

Table 1. Algorithm for deciding PD likelihood of participants, by available sources of PD information.

Likelihood score 1	Likelihood score 2	Likelihood score 3
Tanner score 2–4 (A1)^a	• 1 type of evidence from:Tanner score ≥ 5 (A2)^a • Self-reported PD diagnosis (B)^b • Self-reported PD medication (C)^b	• At least 2 types of evidence from: A2, B and C • At least: PD diagnosis death certificate (D)^c • At least: PD diagnosis HDR (E)^d • At least: PD diagnosis EMR (F)^e

^a assessed in follow-up 3 questionnaire EPIC-NL, baseline and follow-up AMIGO

^b assessed in follow up 1 + 2 questionnaires (open questions) and follow-up 3 questionnaires (PD specific) EPIC-NL, and baseline and follow-up AMIGO

^c assessed from baseline to 31 December 2011 EPIC-NL, baseline to 31 August 2015 AMIGO

^d assessed from baseline to 31 December 2010 EPIC-NL

^e assessed from 2011 until 2013 in AMIGO, ATC and ICPC codes

PD, Parkinson’s Disease; EMR, electronic medical records; HDR, hospital discharge registry; ATC, Anatomical Therapeutic Chemical; ICPC, International Classification of Primary Care.