WHO Environmental Noise Guidelines for the European Region: A Systematic Review of Transport Noise Interventions and Their Impacts on Health

This paper describes a systematic review (1980–2014) of evidence on effects of transport noise interventions on human health. The sources are road traffic, railways, and air traffic. Health outcomes include sleep disturbance, annoyance, cognitive impairment of children and cardiovascular diseases. A conceptual framework to classify noise interventions and health effects was developed. Evidence was thinly spread across source types, outcomes, and intervention types. Further, diverse intervention study designs, methods of analyses, exposure levels, and changes in exposure do not allow a meta-analysis of the association between changes in noise level and health outcomes, and risk of bias in most studies was high. However, 43 individual transport noise intervention studies were examined (33 road traffic; 7 air traffic; 3 rail) as to whether the intervention was associated with a change in health outcome. Results showed that many of the interventions were associated with changes in health outcomes irrespective of the source type, the outcome or intervention type (source, path or infrastructure). For road traffic sources and the annoyance outcome, the expected effect-size can be estimated from an appropriate exposure–response function, though the change in annoyance in most studies was larger than could be expected based on noise level change.


Intervention
17. (prevention or preventive or prevent or preventative or preventing or intervening or intervention* or mitigation or measures or reduction or reducing or reduce or improving or minimizing or program* or campaign* or project* or policy or policies or strategy* or guidelines or directive* or community response or public health response) Design No  [60] Change in annoyance from a road surface intervention This is a study into the effect of a road surface change on noise level, annoyance and behaviour in adjacent residents. The study included a before/ after measurement of 138 respondents at 12 sites and a follow up after 6 months. The paper does not meet the criteria because interventions reported only a change in drive-by levels of vehicles ranging from -7 to +6dBA changes. LAeq24 levels are reported only partially. Annoyance was measured using the ISO standard 10 point scale, and the general community reaction scale of Job et al, 2001. Results show a considerable decrease in annoyance with 7dB reduction and increases even with increase in noise of 1 dB.

Supplementary File 4
Modelled Outcomes of Hypothetical Interventions

Modelled Outcomes
Three of the individual studies identified through the search, two for road traffic and one for aircraft noise, modelled the outcomes from hypothetical interventions. While studies of this type do not provide evidence of the effect of interventions, and hence are not included in the body of this report, they effectively provide important information of the likely extent and magnitude of change in outcomes. Such modelling constitutes a sensitivity analyses to potential interventions, which can assist in the allocation of resources for interventions, and also could assist in the design of future intervention studies. Scenario analysis may be of particular relevance to local authorities and to other implementation agencies. The results of modelling of hypothetical interventions in these three studies are reported in this section. It should be noted that the authors did not conduct a comprehensive search for such studies.

Road Traffic Noise: Modelled Changes in Exposure/Effect
While there were no individual studies that reported change in the noise exposure of a specific population of interest, two studies modelled the effect of hypothetical interventions, reporting either modelled change in exposure of populations of interest or modelled change in their health outcomes.
Summary: Information from modelled road traffic noise interventions, Table S4. The two available studies modelled exposure of urban populations, and one modelled the percentage of the urban population that was highly annoyed -based on this exposure estimate. The modelling involved a combination of hypothetical interventions. One study focussed on interventions of traffic speed and/or traffic volume reductions. Results of the interventions were reported as the percentage of change (increase and decrease) at population levels, within the 5 dB exposure bands, for each intervention. Reductions in the percentages of the population exposed to Lden greater than 70dB ranged from −2% to −7.2% depending on the hypothetical intervention tested.
The second study examined interventions involving hypothetical combinations of quiet tires, roads and quiet cars for three EU cities. Results of the interventions were reported as the percentage of decrease at population levels, for the percentages of the population highly annoyed. Reductions in the percentages highly annoyed ranged from -1% to -7% depending on the hypothetical intervention tested. Combined interventions were shown to be more effective than any single intervention. The study took porous road surfaces into account.

Included Paper
Summary: Information from modelled aircraft noise intervention, Table S5 below. For the relatively small airport at Pisa, the study modelled the aircraft noise exposure of the urban populations. It also modelled the number of people who would be highly annoyed, or highly sleep disturbed, under five different hypothetical mitigation strategies regarding aircraft operations. It is not the particular strategies at this airport, or the estimates generated, that are of interest in this review, but a demonstration of the ability to estimate likely consequences, in health outcome terms, of a variety of environmental noise interventions.            Studies   Table S16. Assessment of the risk of bias in studies in Table 3.  Table S18. Assessment of the risk of bias in studies in Table 5.  Table S19. Assessment of the risk of bias in studies in Table 6.  Table S20. Assessment of the risk of bias in studies in Table 7.  Table S21. Assessment of the risk of bias in studies in Table 8.  Table 9.  Table S23. Assessment of the risk of bias in study in Table 10.  Table S24. Assessment of the risk of bias in studies in Table 11.  Table S26. Assessment of the risk of bias in studies in Table 13.  Table S28. Assessment of the risk of bias in studies in Table 15.  Table S29. Assessment of the risk of bias in studies in Table 17.

Hospital Noise and PLD/Music Venues/Other Sources Interventions
For interventions for some noise sources and for some settings, specific subpopulations were considered: viz. patients in hospitals, and (primarily) young people who use personal listening devices (PLDs) or attend music events. For the hospital subpopulation, sources were all sounds heard in a hospital ward. For the subpopulation of adolescents, the noise exposure was the sound delivered to the users' ears through the headphones of personal listening devices, or the exposure experienced when attending music events or similar. Table S30 shows the number of individual studies considered within each group; Table S31 lists the studies excluded on full-text reading.  (3), 800-809. [106] Interventions included mult-faceted sleep-promoting interventions including some that would have reduced noise exposure. However these were not in terms of changes in levels. Perceived sleep quality was measured as were ratings of noise. ICU quality improvements were associated with significant reductions in perceived nighttime noise levels and a substantial decrease in delirium/coma. Paper excluded as little specific information on noise levels or effects of interventions on levels. Inadequate objective recording of noise disturbance factors and noise levelswhich were the outcome meausres of the noise intervention program. Study reported amount of REM sleep in 'patients' sleeping subject to ICU noise level exposures. Randomised control trial with 70 subjects. Subjects exposed to ICU noise experienced less REM and shorter REM than control group. Study excluded because used health volunteers Studies effect of ICU noise, and peak-reduced ICU noise, on sleep. Sleep registered with polysomnography. IC noise led to more fragmented sleep, more arousals and more time awake. Effects of reduced maxima were minor.

Monsén
Excluded because study was on healthy subjects.

Evidence: Hospital Noise
Two individual studies on hospital noise interventions met the inclusion criteria. The sources were the sounds that were internally generated in hospital wards, particularly intensive care units or similar, such as equipment, alarms, doors, voices etc. The outcomes reported were those for patients, often intensive-care patients, in hospital wards, for sleep disturbance, cardiovascular and other effects.

Outcome: Sleep Disturbance
Summary: Evidence from path interventions, Table S32 This study reports the effect of the wearing, by ICU patients, of ear plugs to reduce noise exposure. The intervention was effective in reducing intensive care delirium and, after the first night of sleep in the ICU, improving the patients' perception of their own sleep.
This study ruled out confounding by matching patient groups on a range of demographic factors, lifestyle, illness, and environmental factors. Risk of bias was assessed as low.

Outcome: Cardiovascular Effects
Summary: Information from source interventions, Table S34 This study reports the effect of hospital noise reduction by use of noise absorbing tiles in an intensive Coronary Care Unit, focussing on the effect of the intervention on hospitalized myocardial patients. Heart Rate, Blood Pressure Pulse amplitude as well as perceptions were measured for two groups: one under good acoustic conditions (following the intervention) and one under bad acoustic conditions. The intervention resulted in a significant physiological effect (change in pulse amplitude) as well as several changes in perceptions of staff, attitude, etc. Remarkable was the utilization of objective physiological response to measure the effect of the intervention.

Evidence: PLD/Music Venues/Other Sources
Seven individual studies on Personal Listening Devices (PLDs), attendance at music venues, and participation in other recreational activities, where there was risk of hearing damage and/or tinnitus, met the inclusion criteria. For all studies, the interventions were aimed at children or adolescents, to change hearing damage risk behaviour, or knowledge of risk. The outcome assessed in all intervention studies was (change in) knowledge of, and behaviours towards, hearing damage risk. There were no objectively measured outcomes.
Note that all interventions examined in this section were of Type E (interventions directed at changes in knowledge or behaviour) and do not include a change in noise level exposure. These studies were not required to meet the general rule for all other individual studies of reporting a change in noise levels in order to be included.

Outcome: Knowledge/Attitude/Behaviour
Summary: Evidence from behavioural interventions, Table S36. The studies all sought evidence on the effectiveness of some form of educational program/campaign on children, adolescents or college students on their perceptions and knowledge of the risk of high levels of noise -generally but not exclusively from PLD sources or from attendance at music events -and on their actual or intended changes to hearing damage risk behaviours including avoidance, frequency or durations of exposures, including regeneration periods when in high noise, or playback levels. Most studies found a significant effect of change in knowledge or behaviour, but at least one author questions if the effects will persist. There is no assessment of the risk of bias in studies in Table S36.

Summary
It is noted that there were few studies for PLD/music venue and hospital settings. Table S37 provides an overview of the observed magnitude of change in health outcome as a result of these interventions. Nearly all entries in Table S37 show that most of interventions led to a change in the aggregate health outcome of those who experienced the intervention (asterisk shown in the YES column), irrespective of the source type and irrespective of the type of intervention.
Two of the available studies of PLD/music venue sources suggest that behavioural/educational interventions for young people with respect to hearing risk may not be sustainable. Table S37. Summary of evidence from the individual intervention studies: Hospital sources and PLD/Music Venue sources.

Number of Papers
Evidence that health outcome changed?