Table A42.
Summary of findings table for the association between aircraft noise exposure and the incidence of diabetes.
| Question | Does Exposure to Aircraft Noise Increase the Risk of Diabetes | |||
|---|---|---|---|---|
| People | Adult population (men and women) | |||
| Setting | Residential setting: people living in Stockholm (Sweden) | |||
| Outcome | The incidence of diabetes | |||
| Summary of findings | RR per 10 dB increase in aircraft noise level (LDEN) | 0.99 (95% CI: 0.47–2.09) | ||
| Number of participants (# studies) | 5156 (1) | |||
| Number of cases | 159 | |||
| Rating | Adjustment to rating | |||
| Quality assessment | Starting rating | 1 cohort # | 4 (high) | |
| Factors decreasing confidence | Risk of bias | Limited a | No downgrading | |
| Inconsistency | NA b | No downgrading | ||
| Indirectness | None c | No downgrading | ||
| Imprecision | Serious d | Downgrading | ||
| Publication bias | NA e | No downgrading | ||
| Factors increasing confidence | Strength of association | NA f | No upgrading | |
| Exposure-response gradient | No evidence of an exposure-response gradient f | No upgrading | ||
| Possible confounding | No conclusions can be drawn g | No upgrading | ||
| Overall judgement of quality of evidence | 2 (low) h | |||
# Since we have a cohort study, we start at 4 (high evidence; a The loss-to-follow-up was estimated as > 20%; b Since only one study is available, this criterion is not applicable; c The study assessed population, exposure and outcome of interest; d Although the number of cases was large, the 95% CI was judged as not sufficiently narrow; e Since the results of only one study were available it was not possible to test for publication bias or small study bias; f The evaluated study found that the risk of diabetes decreased when air traffic noise level increased (RR per 10 dB < 1). No evidence of an exposure-response gradient was found: the evaluated study found an non-significant effect size of 0.99 per 10 dB; g We were not able to draw any conclusions whether possible residual confounders or biases would reduce our effect estimate; h We graded the overall quality of the evidence to be “moderate” (3). Since only one study was available, we downgraded the overall level of evidence to “low” (2).