Table 3.
Clinical outcomes by hearing loss status.
| Interaction | Stroke/TIA | New dementia | Days in hospital | Fall |
|---|---|---|---|---|
| HR (95% CI) | RR (95% CI) | |||
| Age, y | <0.001 | <0.001 | 0.98 | 0.14 |
| 18–<50 | 1.82 (1.71, 1.93) | 2.57 (2.23, 2.96) | 0.96 (0.64, 1.45) | 1.12 (0.90, 1.38) |
| 50–<70 | 1.17 (1.13, 1.20) | 1.45 (1.37, 1.52) | 1.09 (1.06, 1.12) | 1.01 (0.91, 1.12) |
| ≥70 | 1.03 (1.01, 1.05) | 1.10 (1.08, 1.12) | 0.95 (0.88, 1.04) | 1.01 (0.94, 1.08) |
| Sex | 0.002 | <0.001 | <0.001 | 0.01 |
| Female | 1.21 (1.19, 1.24) | 1.31 (1.28, 1.35) | 1.03 (0.88, 1.21) | 1.40 (1.29, 1.51) |
| Male | 1.27 (1.25, 1.30) | 1.54 (1.50, 1.58) | 1.87 (1.61, 2.18) | 1.16 (1.07, 1.27) |
| Number of morbidities | <0.001 | <0.001 | <0.001 | <0.001 |
| 0 | 2.24 (2.12, 2.38) | 3.57 (3.23, 3.94) | 1.33 (1.02, 1.73) | 1.58 (1.33, 1.87) |
| 1–3 | 1.36 (1.33, 1.39) | 1.80 (1.75, 1.85) | 1.39 (1.21, 1.59) | 1.43 (1.33, 1.53) |
| ≥4 | 1.06 (1.04, 1.08) | 1.18 (1.15, 1.20) | 1.16 (1.09, 1.23) | 1.11 (1.06, 1.16) |
HR, Hazard ratio; RR, Rate ratio; TIA, Transient ischemic attack.
All models treat HL as a time-varying covariate, and thus events occurring any time within the study period prior to HL were included in the no-HL group and after hearing loss in the HL group.
HR or RR (95% CI) are reported as appropriate. Stroke/TIA and new dementia (in those without dementia) were regressed on time-varying hearing loss using Weibull regression and interacted with one of five potential modifiers (age categorized as 18–<50, 50–<70, and ≥70 years, sex, number of comorbidities categorized as 0, 1–3, or ≥4 morbidities) with shared frailty modelled in either 5-year age and sex groups, or simply age or sex groups, depending on the modifier. Length of hospital stays and falls were regressed on time-varying hearing loss using negative binomial regression and interacted with one of five potential modifiers where the clustered sandwich estimator was used for variance estimation in the 5-year age and sex groups, just age, or just sex groups.
The models were further adjusted for time-varying rural status, material deprivation quintiles, obesity, hypertension, chronic pain, depression, chronic pulmonary disease, diabetes mellitus, hypothyroidism, osteoporosis, gout, stroke or TIA, fragility fractures, heart failure, cancer, asthma, alcohol misuse, coronary artery disease, atrial fibrillation, irritable bowel syndrome, rheumatic diseases, epilepsy, dementia, schizophrenia, inflammatory bowel disease, multiple sclerosis, severe constipation, peripheral artery disease, Parkinson's disease, psoriasis, severe chronic kidney disease, peptic ulcer disease, and chronic liver disease.
Bold values indicate statistically significant at p < 0.05.