Introduction

Intrauterine devices (IUDs) are one of the most effective forms of reversible contraception, with typical 1-year failure rates of <1%. They are convenient with a positive safety profile and high rates of user continuation and satisfaction. Despite these appealing characteristics, use of IUDs in the United States continues to lag behind use of less effective methods such as oral contraceptive pills and condoms. This may in part be due to women’s knowledge of IUDs. Contraceptive knowledge plays a role in women’s contraceptive use. Women who are aware of the actual effectiveness of various contraceptive methods are more likely to choose the IUD. Conversely, women who are misinformed about the safety of IUDs may be less likely to use them. Past studies have revealed that many common concerns, including ones about infection, infertility, and cancer due to IUD, are based on inaccurate and outdated information. These concerns and knowledge gaps stress the importance of accurate information for patients who are considering contraceptive options.

Many patients report using the Internet to supplement the health information provided by health care professionals. For example, in 2012, 72% of adults reported looking for health information online in the past year, and of these 77% sought health information via a search engine. Currently, there are a number of web sites designed for the lay public that provide information about contraception, and a recent study showed that web–based contraceptive education increased contraception knowledge. Another recently published study showed that, among young adults, the Internet was the second most popular source of information about contraception after health professionals.

To provide effective contraceptive education, web–based resources must contain accurate, comprehensive, and up-to-date information. One study of the information available about the IUD on the Internet found that while most sites stated the IUD was effective and long lasting, fewer than half of the sites stated that it was safe, and more than half incorrectly stated that the IUD increased the risk of pelvic inflammatory disease (PID), ectopic pregnancy, and infertility. This study was performed more than a decade ago. Since then, the breadth of health information available on the Internet has greatly increased. Improved knowledge about the accuracy of Internet content could be useful for health care providers when performing contraceptive counseling. In addition, knowledge about which web sites provide accurate and comprehensive information could identify reliable resources for patients. The objective of this study was to describe and evaluate the quality of available Internet-based health information about the IUD.

Materials and Methods

We performed a search for information about the IUD using the most commonly used search engines, Google, Yahoo, Bing, and Ask, and the following terms: “birth control,” “contraception,” “contraceptive method,” “pregnancy prevention,” “intrauterine device,” “IUD,” “intrauterine system,” “intrauterine contraception,” “Mirena,” and “Paragard.” Skyla was not available when the study was designed and therefore not included in the search terms. All web searches were completed in August 2013 using a web browser with no stored browsing history so as not to generate personalized results. We updated the search in October 2015 to ensure that the previously coded web sites were still available and had similar content. We included in our initial sample all web sites identified on the first 3 pages of returns except paid advertisements and duplicate sites. Prior studies have shown that 62% of users will click on a result on the first page of search results and 90% of users are unlikely to go beyond the first 3 pages of search results.

The initially identified web sites were reviewed by a member of the research team for possible inclusion. Sites were eligible if they contained educational information about contraception that was intended for a lay audience. Duplicate sites, non-US sites, and nonaccessible sites were not included. Web sites were also excluded if the site contained nonrelevant content such as information for health care professionals; news articles; dictionary definitions; information about only emergency, male, or nonhuman contraception; or were focused on advocacy. Finally, sites were excluded if they were primarily forums, personal blogs, or paid advertisements.

We developed a 56-item instrument to evaluate the informational content of the eligible web sites based on current guidelines and a prior study reporting common misconceptions. The initial item assessed whether the site contained information about contraception in general and then specifically about IUDs. If the web site contained information about IUDs, additional items asked about whether the web site correctly identified the types of available IUDs, the mechanism of action of the different IUDs, and the advantages, disadvantages, contraindications, and risks of the IUD. The coding instrument contained separate items for the mechanism of the copper IUD and the levonorgestrel intrauterine system. There were also questions about the advantages and disadvantages that were specific to the IUD. If information about risks was presented, the instrument asked if and how the risks were presented and quantified. The instrument also specifically surveyed whether the web site contained inaccurate information about the IUD such as “increases the risk of ectopic pregnancy” and “decreases fertility.” The correct responses were determined from 2 evidence-based references. We considered a stated risk of PID other than in the first month following insertion to be incorrect. To assess the usability of the web site, we used a 5-point Likert scale to quantify the coder’s overall impression of the site from basic to comprehensive, inaccurate to accurate information, and whether the interface was user friendly. The coder’s impression of the overall valence of the site as negative or positive toward IUDs was also recorded. Information about the type of organization hosting the web site was collected. All data were entered directly into REDCap, a secure, Internet–based application designed to support data capture for research studies. Institutional review board approval was not required as there were no human subjects involved in the research.

The first 30 web sites were independently coded by 2 research assistants (S.C. and M.K.). Intercoder agreement and kappa were calculated for each item regarding the accuracy of the content. Coding was reviewed by the entire research team and minor discrepancies resolved by consensus. In cases of major discrepancies, the item was modified and the sites were recoded and rereviewed. The remaining web sites (n = 75) were coded by 1 of the 2 trained coders. We calculated frequencies and proportions to describe the results. We also determined an accuracy score for the web site content by subtracting the inaccurate score from the accurate score. The accurate score was calculated by assigning 1 point for every accurate item coded and could have a possible maximum score of 44. The inaccurate score was calculated by assigning 1 point for every inaccurate or missing item coded and could have a possible maximum score of 22. The range of possible accuracy scores was –22 to 44. Statistical analyses were performed using software (STATA 11; Stata Corp, College Station, TX).

Materials and Methods

We performed a search for information about the IUD using the most commonly used search engines, Google, Yahoo, Bing, and Ask, and the following terms: “birth control,” “contraception,” “contraceptive method,” “pregnancy prevention,” “intrauterine device,” “IUD,” “intrauterine system,” “intrauterine contraception,” “Mirena,” and “Paragard.” Skyla was not available when the study was designed and therefore not included in the search terms. All web searches were completed in August 2013 using a web browser with no stored browsing history so as not to generate personalized results. We updated the search in October 2015 to ensure that the previously coded web sites were still available and had similar content. We included in our initial sample all web sites identified on the first 3 pages of returns except paid advertisements and duplicate sites. Prior studies have shown that 62% of users will click on a result on the first page of search results and 90% of users are unlikely to go beyond the first 3 pages of search results.

The initially identified web sites were reviewed by a member of the research team for possible inclusion. Sites were eligible if they contained educational information about contraception that was intended for a lay audience. Duplicate sites, non-US sites, and nonaccessible sites were not included. Web sites were also excluded if the site contained nonrelevant content such as information for health care professionals; news articles; dictionary definitions; information about only emergency, male, or nonhuman contraception; or were focused on advocacy. Finally, sites were excluded if they were primarily forums, personal blogs, or paid advertisements.

We developed a 56-item instrument to evaluate the informational content of the eligible web sites based on current guidelines and a prior study reporting common misconceptions. The initial item assessed whether the site contained information about contraception in general and then specifically about IUDs. If the web site contained information about IUDs, additional items asked about whether the web site correctly identified the types of available IUDs, the mechanism of action of the different IUDs, and the advantages, disadvantages, contraindications, and risks of the IUD. The coding instrument contained separate items for the mechanism of the copper IUD and the levonorgestrel intrauterine system. There were also questions about the advantages and disadvantages that were specific to the IUD. If information about risks was presented, the instrument asked if and how the risks were presented and quantified. The instrument also specifically surveyed whether the web site contained inaccurate information about the IUD such as “increases the risk of ectopic pregnancy” and “decreases fertility.” The correct responses were determined from 2 evidence-based references. We considered a stated risk of PID other than in the first month following insertion to be incorrect. To assess the usability of the web site, we used a 5-point Likert scale to quantify the coder’s overall impression of the site from basic to comprehensive, inaccurate to accurate information, and whether the interface was user friendly. The coder’s impression of the overall valence of the site as negative or positive toward IUDs was also recorded. Information about the type of organization hosting the web site was collected. All data were entered directly into REDCap, a secure, Internet–based application designed to support data capture for research studies. Institutional review board approval was not required as there were no human subjects involved in the research.

The first 30 web sites were independently coded by 2 research assistants (S.C. and M.K.). Intercoder agreement and kappa were calculated for each item regarding the accuracy of the content. Coding was reviewed by the entire research team and minor discrepancies resolved by consensus. In cases of major discrepancies, the item was modified and the sites were recoded and rereviewed. The remaining web sites (n = 75) were coded by 1 of the 2 trained coders. We calculated frequencies and proportions to describe the results. We also determined an accuracy score for the web site content by subtracting the inaccurate score from the accurate score. The accurate score was calculated by assigning 1 point for every accurate item coded and could have a possible maximum score of 44. The inaccurate score was calculated by assigning 1 point for every inaccurate or missing item coded and could have a possible maximum score of 22. The range of possible accuracy scores was –22 to 44. Statistical analyses were performed using software (STATA 11; Stata Corp, College Station, TX).