Statistical biases due to anonymization evaluated in an open clinical dataset from COVID-19 patients.

Anonymization has the potential to foster the sharing of medical data. State-of-the-art methods use mathematical models to modify data to reduce privacy risks. However, the degree of protection must be balanced against the impact on statistical properties. We studied an extreme case of this trade-off: the statistical validity of an open medical dataset based on the German National Pandemic Cohort Network (NAPKON), which was prepared for publication using a strong anonymization procedure. Descriptive statistics and results of regression analyses were compared before and after anonymization of multiple variants of the original dataset. Despite significant differences in value distributions, the statistical bias was found to be small in all cases. In the regression analyses, the median absolute deviations of the estimated adjusted odds ratios for different sample sizes ranged from 0.01 [minimum = 0, maximum = 0.58] to 0.52 [minimum = 0.25, maximum = 0.91]. Disproportionate impact on the statistical properties of data is a common argument against the use of anonymization. Our analysis demonstrates that anonymization can actually preserve validity of statistical results in relatively low-dimensional data.

Transvenous Phrenic Nerve Stimulation for Treatment of Central Sleep Apnea: Five-Year Safety and Efficacy Outcomes.

BACKGROUND: The remede System Pivotal Trial was a prospective, multi-center, randomized trial demonstrating transvenous phrenic nerve stimulation (TPNS) therapy is safe and effectively treats central sleep apnea (CSA) and improves sleep architecture and daytime sleepiness. Subsequently, the remede System was approved by FDA in 2017. As a condition of approval, the Post Approval Study (PAS) collected clinical evidence regarding long-term safety and effectiveness in adults with moderate to severe CSA through five years post implant. METHODS: Patients remaining in the Pivotal Trial at the time of FDA approval were invited to enroll in the PAS and consented to undergo sleep studies (scored by a central laboratory), complete the Epworth Sleepiness Scale (ESS) questionnaire to assess daytime sleepiness, and safety assessment. All subjects (treatment and former control group) receiving active therapy were pooled; data from both trials were combined for analysis. RESULTS: Fifty-three of the original 151 Pivotal Trial patients consented to participate in the PAS and 52 completed the 5-year visit. Following TPNS therapy, the apnea-hypopnea index (AHI), central-apnea index (CAI), arousal index, oxygen desaturation index, and sleep architecture showed sustained improvements. Comparing 5 years to baseline, AHI and CAI decreased significantly (AHI baseline median 46 events/hour vs 17 at 5 years; CAI baseline median 23 events/hour vs 1 at 5 years), though residual hypopneas were present. In parallel, the arousal index, oxygen desaturation index and sleep architecture improved. The ESS improved by a statistically significant median reduction of 3 points at 5 years. Serious adverse events related to implant procedure, device or delivered therapy were reported by 14% of patients which include 16 (9%) patients who underwent a pulse generator reposition or lead revision (primarily in the first year). None of the events caused long-term harm. No unanticipated adverse device effects or related deaths occurred through 5 years. CONCLUSION: Long-term TPNS safely improves CSA, sleep architecture and daytime sleepiness through 5 years post implant. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT01816776.