“…To estimate the stability of outcomes, we conducted sensitivity analyses by assessing the effects of removing low-quality publications (Francey et al, 2012 ; Chang and Choi, 2014 ; Gu et al, 2015 ; Vona et al, 2015 ; Bademci et al, 2016 ; Mehta et al, 2016 ; Sommen et al, 2016 ; Baux et al, 2017 ; Zazo Seco et al, 2017 ; Cabanillas et al, 2018 ; Guan et al, 2018 ; Marková et al, 2018 ; Sheppard et al, 2018 ; Back et al, 2019 ; Budde et al, 2020 ; Downie et al, 2020 ; Kim et al, 2020 ; Safka Brozkova et al, 2020 ). The summary prevalence of DFNB16 in non-GJB2 HL patients was 3.98% (95% CI: 0.0260–0.0606, I 2 = 70%, p < 0.01), which stabilized the findings in the range of that of crude analysis.…”