A new cohort study published in Occupational and Environmental Medicine, reports that long-term exposure to aircraft noise, particularly during the night, is associated with incident hypertension and potentially, also, cardiovascular events.
Title & Date of Publication of Journal
Is aircraft noise exposure associated with cardiovascular disease and hypertension? Results from a cohort study in Athens, Greece.
Published: 13 June 2017
Journal: Occupational and Environmental Medicine
Study’s main claims – and are they supported by the data
The paper appears to support the claim that ‘incidence of hypertension is associated with higher exposure to aircraft noise at night’ however this result should be interpreted cautiously due to a number of methodological limitations.
This conclusion is based on an observational study. In the study data, participants who lived closer to the airport and therefore experienced increased aircraft noise were on average more likely to be diagnosed with high blood pressure and heart flutter. However this does not necessarily mean that the increased noise exposure caused the health problems: participants closer to the airport may have been exposed to increased air pollution, or may have poorer social-economic circumstances linked to lower accommodation prices: it could be that these or other factors led to the increase in health problems.
This conclusion is based on only a very small number of cases, only 71 new cases of hypertension, 68 cases in total of cardiac arrhythmia and even smaller numbers of heart attack and stroke. The authors base their results on statistical models that include 8 additional factors (age, gender, BMI etc) and it is questionable as to whether these models are appropriate when the number of health events is so low.
The results of this study are expressed as odds ratios. The press release incorrectly interprets the odds ratio of 1.69 as a ‘69% heightened risk of the condition’- odds ratios are commonly misinterpreted as risk differences, but an odds ratio of 1.69 can only be interpreted as a 69% increase in the odds of the condition, which is not the same as a 69% increase in the risk. When an event is rare the odds ratio and relative risk is similar, but this may not be the case here as the rate of hypertension is quite high. Therefore, the odds ratio is likely to overestimate the increase in risk in this case (the actual converted risk from the odds ratio cannot be calculated from the material available as the figures for the control group are not there).
The researchers have looked at a large number of outcomes and exposures simultaneously –It is unclear as to whether any of outcome/exposure pairs were chosen in advance as being the most important, and the ‘most significant’ results were highlighted in the abstract. When looking at a large number of questions in this way without a clear single primary hypothesis, the probability of false conclusions is very high.
This study followed participants up over 7-9 years. Early data on exposure and other variables was available without the need to rely on participant recall which may be poor.
Data on noise levels was obtained using participant addresses and health data was obtained by BP measurement and questionnaires administered in the participant home by specially trained staff. These methods of data collection appear reliable.
The conclusions seem to match conclusions from other observational studies.
The conclusion is based on observational data so cannot demonstrate causally that aircraft noise caused increased high blood pressure, heart flutter, or stroke. There may have been other factors that led to increased health problems.
Of the 780 individuals who participated in the original cross sectional study, only 420 (54%) participated in the follow up. Some of the original participants died, or moved away from the area. It is also likely that response rate was related to noise exposure with more participants responding from the group with the highest level of noise exposure. The authors have not explored what impact this large amount of missing data may have had on their conclusions.
This study only included participants close to Athens International Airport which began operation in 2001. Many participants will have lived in the area since before the airport was built and may have negative feelings about it. Participants who live close to airports may have their own concerns about the relationship between aircraft noise and health risks. This may mean that people who live closest to the airport are more likely to take part in the study and more likely to self-report health problems than those who live further away. The study should be replicated in other settings before firm conclusions can be made.
The press release states ‘Around half of the participants (just under 49%) were exposed to more than 55 dB of daytime aircraft noise’ – it appears that this figure should be 44.8% based on the information in Table 2.
Any specific expertise relevant to studied paper (beyond statistical)?
Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.