Risk Factors of Second Surgery for Adjacent Segment Disease Following Anterior Cervical Discectomy and Fusion

All second ACDF surgery after one year from the first ACDF were identified as a consequence of ASD that required another surgery. A multivariate competing risk survival model, Kaplan-Meier survivorship, and average time to events were calculated.

2.1. Data source and ethical concerns The study used the National Health Insurance Research Database (NHIRD), a national database containing 26 million administered insurants accumulated from January 1997 to December 2013, provided by the National Health Research Institutes of Taiwan. Due to the unique social-welfare health insurance system operated by the government, it is mandatory for every resident in Taiwan to be enrolled. Thus, the NHIRD has covered 99% of the population since its launch in 1996. The monopolistic national health insurance also offers unrestricted access to any healthcare provider of the patients' choices. The statistics, therefore, gathered by the NHIRD represent a sound cohort for investigation of the natural course and subsequent management of diseases. It is particularly good for studies that need longitudinal observation for repeat treatment (i.e. surgery for ASD) because it allows for the capture of events, even though they may occur in multiple or different institutes and hospitals. The universal coverage and comprehensive follow-up provided a valuable chance to study ASD.

All the personal information had already undergone a de-identification and encryption process. Individual and hospital identifiers are unique to the research database and researchers therefore cannot trace individual patients or health service providers.

Inclusion criteria

• patients who had been admitted for their first ACDF surgery during the 16-year span, began on January 1st, 1998 till the end of 2013, recorded in the NHIRD

Exclusion criteria

• any re-operation (i.e. ACDF) within 365 days from the indexed date of the first ACDF were excluded.