2. Methods
This review uses a modified version of the five stage methodological framework for conducting scoping reviews developed by Arksey and O'Malley1. The five steps are identifying the research question or questions, identifying relevant studies and other documents, selecting studies and documents to be included, charting the data, and collating, summarizing and reporting the results.
| Identifying the research questions |
The following research questions guided the review:
|
| Identifying relevant studies and documents |
| For indexed, peer-reviewed literature, three separate searches were conducted.
Primary search: For articles dealing with health and health-related outcomes of wind turbines, a comprehensive search strategy was adopted. A search of all articles published in English between 1980 and September 1, 2009 was conducted using the following databases: MEDLINE, EMBASE, Global Health, PsychINFO, and ISI Web of Science. Search strategies for each database were developed using natural language text words and controlled vocabulary terms specific to each database. For the first four databases, preliminary exploratory searches revealed a limited number of articles on wind turbines generally, so the search terms were kept deliberately broad: wind turbine(s) or wind farm(s). For the final database, these terms proved too broad, and thus several filtering terms were used to limit the results broadly to health-related outcomes: health, annoyance, stress, sleep or noise. In addition, hand searches were conducted of several relevant journals: Bulletin of Science, Technology & Society; Journal of Low Frequency Noise, Vibration and Active Control; Noise & Health; Journal of the Acoustical Society of America, and Wind Energy. The reference lists of retrieved articles were also searched for possibly relevant articles that had been missed. Studies eligible for inclusion were journal articles, books or book chapters from academic publishers and peer-reviewed conference papers. Following the same methods, a second search was conducted for the period September 1, 2009 to May 1, 2012 to ensure the most up-to-date studies were included in this final report. Secondary searches: For articles dealing with the technical aspects of wind turbines, and with health and noise more generally, two searches were conducted using the same databases for all articles published in English between 1980 and September 1, 2009. For the technical articles, search terms were modified to select studies related to the technical aspects of wind turbines relevant to the review: noise, infrasound, low frequency noise or sound, vibration, amplitude modulation, sound, sound levels, sound or noise models. For the noise and health articles, search terms relating to the relationship between noise and health were used: noise and health, noise and annoyance, noise and stress, infrasound and health, low frequency noise and health, noise and annoyance and health. The intention with these searches was not to be exhaustive, but rather to find articles that would provide context to the literature identified in the primary search. These searches were meant to compliment articles already revealed by the primary search. Studies eligible for inclusion were journal articles, books or book chapters from academic publishers and peer-reviewed conference papers. These searches were not repeated for the second time period, although new articles indentified in the updated primary search were considered for inclusion. For non-peer-reviewed sources, the majority of which are not indexed, we used several search strategies. For articles dealing with health and health-related outcomes of wind turbines, eligible sources included: self-published research studies, conference abstracts, review articles, policy papers, opinion articles, presentations, news articles, websites, blogs and visual media. The expanded range of eligible study types reflects a change in the traditional definition of grey literature, which included scientific and technical reports, government documents, theses, conference abstracts and working papers from research groups or committees. With the increased use of the internet, the definition of grey literature has broadened, since almost anyone can publish his or her own work by posting a document on the internet. To access these sources, several preliminary Google searches were conducted using the following search terms: wind turbines and health, wind farms and health, wind turbines and annoyance, wind farms and annoyance, wind turbine syndrome, vibroacoustic disease and wind turbines. Several community groups maintain indexes of publications on the subject of wind turbines and health, and these indexes were consulted: Industrial Wind Action Group, Wind Watch, Wind Concerns Ontario (now called the Ontario Wind Resistance) and an earlier version of Windturbinesyndrome.com. For Canadian news sources, the Canadian Newsstand Index was also consulted using the search terms wind turbine(s) or wind farm(s) and health. While identified documents were retrieved from the publishing source where possible, in some cases they were unobtainable (e.g., court depositions, self-published reports) and in these cases the most recent version of the document was obtained using Google. |
| Selecting studies for inclusion in the review |
Studies from the above searches were sorted into five categories that naturally developed:
Studies in the first category were the result of the primary search. To be included, studies had to contain a direct discussion of the health impact of wind turbines, and present new or newly synthesized health information in some detail. Studies in the second and fourth categories were the result of the Internet searches. For inclusion in the second category, studies had to contain a direct discussion of the health impact of wind turbines, and present new or newly synthesized health information in some detail. Unsystematic reports of adverse event reports with speculated causes, or summaries of information on wind turbines and health readily available in the original sources were not included. The intention was to include all reports commonly referenced by community groups concerned about wind turbines. The fourth category was intended to include a representative collection of community group writings about the health impacts of wind turbines. In addition, selections from news reports were used to create as broad a list of reported symptoms as possible from as many regions and countries as possible. Studies in the third and fifth categories were the result of the secondary searches. For inclusion in the third category, studies had to contain a discussion of technical aspects of wind turbines relevant to the health outcomes. For inclusion in the fifth category, studies had to contain a discussion of noise and health that was relevant to wind turbines and health. For these two categories, the intention was not to be exhaustive, but to include studies frequently referenced by studies in the first two categories, and studies that provided contextual information for studies in the first two categories. Studies identified in the fifth category were reviewed by two researchers, and those determined to be relevant by one or both researchers were included. Studies in all categories were recorded in a database created using Zotero 3. |
| Charting the data |
| Data from each document were abstracted into a chart developed by the research team. The information captured included basic bibliographic information, background information, technical information, issues and concerns, results pertaining to health, other results and policy implications or suggestions. For research articles, methods were recorded in a dedicated section. For studies in categories one, two and five, data abstraction was done by one author and reviewed by the second author independently. |
| Collating, summarizing and reporting the results |
| The results from categories one and two are presented in the main body of this report, together with a discussion of the results and limitations of the literature. All material from these categories is included. Information from categories three, four and five is presented only to provide context in the discussion of health-related effects, and was used to inform conclusions and recommendations. |