Observations that hearing loss is a substantial risk factor for dementia may be accounted for by a common pathology. Mitochondrial oxidative stress and alterations in α-synuclein pathology may be common pathology candidates. Crucially, these candidate pathologies are implicated in Parkinson’s disease (PD). Consequently, hearing loss may be a risk factor for PD. Subsequently, this prospective cohort study of the English Longitudinal Study of Ageing examines whether hearing loss is a risk factor for PD longitudinally. Participants reporting self-reported hearing capabilities and no PD diagnosis prior to entry (n = 14,340) were used. A joint longitudinal and survival model showed that during a median follow up of 10 years (SD = 4.67 years) increased PD risk (p < 0.001), but not self-reported hearing capability (p = 0.402). Additionally, an exploratory binary logistic regression modelling the influence of hearing loss identified using a screening test (n = 4812) on incident PD indicated that neither moderate (p = 0.794), nor moderately severe/severe hearing loss (p = 0.5210), increased PD risk, compared with normal hearing. Whilst discrepancies with prior literature may suggest a neurological link between hearing loss and PD, further large-scale analyses using clinically derived hearing loss are needed.

Hearing impairment is implicated as a risk factor for Parkinson's disease (Parkinson's) incidence, with evidence suggesting that clinically diagnosed hearing loss increases Parkinson's risk 1.5-1.6 fold over 2-5 years follow up. However, the evidence is not unanimous with additional studies observing that self-reported hearing capabilities do not significantly influence Parkinson's incidence. Thus, additional cohort analyses that draw on alternative auditory measures are required to further corroborate the link between Parkinson's and hearing impairment. To determine whether hearing impairment, estimated using a speech-in-noise test (the Digit Triplet Test, DTT), is a risk factor for Parkinson's incidence. This was a pre-registered prospective cohort study using data from the UK Biobank. Data pertaining to 159,395 individuals, who underwent DTT testing and were free from Parkinson's at the point of assessment, were analysed. A Cox Proportional Hazard model, controlling for age, sex and educational attainment was conducted. During a median follow up of 14.24 years, 810 cases of probable Parkinson's were observed. The risk of incident Parkinson's increased with baseline hearing impairment [hazard ratio: 1.57 (95%CI: 1.018, 2.435; P = .041)], indicating 57 % increase in risk for every 10 dB increase in speech-reception threshold (SRT). However, when hearing impairment was categorised in accordance with UK Biobank SRT norms neither 'Insufficient' nor 'Poor' hearing significantly influenced Parkinson's risk compared to 'Normal' hearing. The congruence of these findings with prior research further supports the existence of a relationship between hearing impairment and Parkinson's incidence.

Regression modeling is the workhorse of statistics and there is a vast literature on estimation of the regression function. It has been realized in recent years that in regression analysis the ultimate aim may be the estimation of a level set of the regression function, ie, the set of covariate values for which the regression function exceeds a predefined level, instead of the estimation of the regression function itself. The published work on estimation of the level set has thus far focused mainly on nonparametric regression, especially on point estimation. In this article, the construction of confidence sets for the level set of linear regression is considered. In particular, level upper, lower and two‐sided confidence sets are constructed for the normal‐error linear regression. It is shown that these confidence sets can be easily constructed from the corresponding level simultaneous confidence bands. It is also pointed out that the construction method is readily applicable to other parametric regression models where the mean response depends on a linear predictor through a monotonic link function, which include generalized linear models, linear mixed models and generalized linear mixed models. Therefore, the method proposed in this article is widely applicable. Simulation studies with both linear and generalized linear models are conducted to assess the method and real examples are used to illustrate the method.