Impact of loss to follow-up on survival estimation for cystic fibrosis

Abstract

Purpose: Deaths among those lost to follow-up (LTFU) in the Cystic Fibrosis Foundation Patient Registry (CFFPR) are critically important to the epidemiology of cystic fibrosis (CF). Unreported deaths could impact estimates of survival if LTFU is associated with disease trajectory.

Methods: We characterized the LTFU population (1986-2017) from the CFFPR and identified deaths via linkage with the National Death Index (NDI). Median predicted survival age and conditional survival were estimated with and without additional deaths and person-time from the NDI.

Results: Of the 10,582 individuals LTFU in the CFFPR, 2,460 (23.2%) matched to an NDI death record. Individuals who died after LTFU with a CF diagnosis were 43% female, 91% White/non-Hispanic, 59% had advanced CF lung disease based on last CFFPR recorded forced expiratory volume in one second (FEV₁) %predicted <40 and 18% were post-lung transplant. Median predicted survival age during the most recent period available, 2013-2017, increased from 44.3 years (95% CI: 43.2, 45.7) to 45.8 years (95% CI 44.5, 47.1) with the inclusion of NDI data.

Conclusions: Inclusion of deaths and additional person-time among those LTFU changed the point estimate of median predicted survival for most time periods and increased the point estimate from 2009 onwards.

Keywords: Cystic fibrosis; Loss to follow-up; Registry data; Survival analysis.