Bone-anchored hearing device surgery: a 3- to 6-year follow-up with life table and worst-case scenario calculation

Published studies on complications for the bone-anchored hearing devices can present different outcomes. Unfortunately, data are often difficult to interpret, compare, or use in meta-analysis owing to missing data. This study presents a sensitivity analysis of presenting cohort data, which considers incomplete follow-up data. An example is given using life table calculations and worst-case scenarios for implant losses and revision surgeries after installing BAHA.

A retrospective case review of all patients consecutively operated on for BAHA between 2005 and 2007 with a 1-stage procedure. Life table calculations were used to handle incomplete follow-up data, including 50% of patients lost to follow-up as failures. In addition, a worst-case scenario was calculated, in which 100% of patients lost to follow-up were considered to be failures.

A total of 138 patients were operated on at Sahlgrenska, with a follow-up time of 3 to 5 years. Within the study cohort, 9 patients (6.5%) lost the implant. Reasons for incomplete follow-up data were as follows: "lost to follow-up" (n = 4), "not using BAHA" (n = 3), and "patient was deceased" (n = 2). Life table calculation gave a calculated implant loss of 9.8%. The worst-case scenario gave an implant loss frequency of 13%. Revision surgery was performed in 3 (2.2%) of 138 patients in the study cohort. With life table and worst-case scenario calculations, the frequency of revision surgery was 6.2% and 10.1%, respectively.

Incomplete follow-up data can theoretically affect outcome data considerably. Therefore, outcome data should also be presented with life table and worst-case scenario. This would enhance the possibility to interpret and compare data.