Skip to main content
. 2011 Dec 22;11:61. doi: 10.1186/1471-2466-11-61

Table 2.

Attrition of patients using the inclusion/exclusion criteria and assignment to three treatment groups before 1:1:1 matching process.

Criteria Removed Remaining
Number of patients MarketScan database (1/2005-6/2008) NA 49,042,666

Have at least one COPD Study Drug between Jan 1, 2006 - Jun 30, 2007) 48,879,492 163,174

Omit patients with missing or zero quantity or days supply values on any COPD study prescription claim 2,785 160,389

Require at least one diagnosis for COPD (490.xx-492.xx, 494.xx, 496.xx) in any diagnosis field at any time in the data. 44,560 115,829

Omit patients with RX = 0 on any claims medical during the study period; this is MarketScan's indicator that the pharmacy claims were incomplete or unavailable. 690 115,139

Omit patients with a diagnosis of asthma (ICD-9-CM 493.xx), cystic fibrosis (277.0x), or respiratory tract cancer (160.x-164.x) at any time during the available data. 51,149 63,990

Require 12 months pre/12 months post index continuous eligibility 19,871 44,119

Include only patients aged 40 or older on index date 205 43,914

Have at least 30 days per quarter coverage for index prescription 17,669 26,245

Have at least one prescription for tiotropium, formoterol or salmeterol with treatment patterns that were in keeping with one of the three treatment groups (LABA or LAMA, LABA and LAMA or LABA, LAMA and ICS therapy) 9,689 16,556

Omit patients with missing PLANTYPs in CCAE enrollment data 55 16,501

Omit patients with negative dollar amounts in the NETPAY field. 644 15,857

TOTAL 15,857


Treatment group distribution prior to propensity match: n %

 LABA or LAMA group 8,334 52.6

 LABA and LAMA group 592 3.7

 LABA, LAMA, and ICS group 6,931 43.7