Assessment of the quality and variability of health information on chronic pain websites using the DISCERN instrument
- Jatin Kaicker†1,
- Victoria Borg Debono†1,
- Wilfred Dang1,
- Norman Buckley1 and
- Lehana Thabane2, 3Email author
© Kaicker et al; licensee BioMed Central Ltd. 2010
Received: 15 July 2010
Accepted: 12 October 2010
Published: 12 October 2010
The Internet is used increasingly by providers as a tool for disseminating pain-related health information and by patients as a resource about health conditions and treatment options. However, health information on the Internet remains unregulated and varies in quality, accuracy and readability. The objective of this study was to determine the quality of pain websites, and explain variability in quality and readability between pain websites.
Five key terms (pain, chronic pain, back pain, arthritis, and fibromyalgia) were entered into the Google, Yahoo and MSN search engines. Websites were assessed using the DISCERN instrument as a quality index. Grade level readability ratings were assessed using the Flesch-Kincaid Readability Algorithm. Univariate (using alpha = 0.20) and multivariable regression (using alpha = 0.05) analyses were used to explain the variability in DISCERN scores and grade level readability using potential for commercial gain, health related seals of approval, language(s) and multimedia features as independent variables.
A total of 300 websites were assessed, 21 excluded in accordance with the exclusion criteria and 110 duplicate websites, leaving 161 unique sites. About 6.8% (11/161 websites) of the websites offered patients' commercial products for their pain condition, 36.0% (58/161 websites) had a health related seal of approval, 75.8% (122/161 websites) presented information in English only and 40.4% (65/161 websites) offered an interactive multimedia experience. In assessing the quality of the unique websites, of a maximum score of 80, the overall average DISCERN Score was 55.9 (13.6) and readability (grade level) of 10.9 (3.9). The multivariable regressions demonstrated that website seals of approval (P = 0.015) and potential for commercial gain (P = 0.189) were contributing factors to higher DISCERN scores, while seals of approval (P = 0.168) and interactive multimedia (P = 0.244) contributed to lower grade level readability, as indicated by estimates of the beta coefficients.
The overall quality of pain websites is moderate, with some shortcomings. Websites that scored high using the DISCERN questionnaire contained health related seals of approval and provided commercial solutions for pain related conditions while those with low readability levels offered interactive multimedia options and have been endorsed by health seals.
Health information and the Internet
The Internet is an important international electronic network and mass medium for individuals seeking information pertaining to almost any topic, including health information and health care services . Reports by the National Telecommunications and Information Administration in the United States indicate that in 2000, 44% of individuals had access to the Internet . This percentage increased to 54% by 2008, with current usage estimated at over one billion users in the world [3–5]. In 2009, Statistics Canada found that 80% of Canadians over the age of 16 use the Internet for personal reasons with searching for health information reported by 70% of home user, up from 59% in 2007 . Growth in internet usage has led to about 37% of consumers using the electronic network to retrieve health related knowledge . Consumers report convenience, diversity and anonymity of information sources on the Internet as reasons for using it as attractive alternatives to consulting with a clinician . One concern and public health issue is the quality of health information on the Internet. Despite efforts toward standardization, health information on the web remains unregulated and varies substantially in quality, accuracy, and readability .
Issues with online health information
An individual's risk of encountering an online website that is compromised in its quality, accuracy and readability is a combination of two variables: the proportion of inadequate information on the web, and the competency of the population to filter out those sites that lack reliability and validity. It is plausible that a correlation exists between the proportion of websites presenting low quality health information and adverse health outcomes in a target population. Consumer risk can be reduced through the introduction of clear critical appraisal tools and with a standardized website evaluation system for health information, thus improving the ability of users to locate trustworthy sites and to filter the inadequate ones .
Apart from the quality of health information on the web, patients also find many websites presenting health information using highly technical language. Technical presentations may be advantageous for researchers and clinicians; however, this technical language can be overwhelming and confusing, especially if it is not properly explained . Therefore, it is also imperative to systematically assess the presentation of online health information using readability algorithms to ensure that such information is easily assessable to lay audiences.
The Internet and pain management
Chronic pain is a serious health and socioeconomic concern . In the mid 1990s, the National Population Health Survey in Canada showed that about four million Canadians, about 17% of the population at the time, suffered from chronic pain with a negative impact on quality of life . Seventy percent of those reporting pain rated their pain as moderate to severe, intruding upon activities of daily living. Patients with chronic pain visited their family physician more frequently (12.9 versus 3.8 mean visits per year) and spent more time as hospital inpatients (3.9 versus 0.7 days) than those without pain . In the United States, the American Productivity Audit found that over a two-week period in 2001, 13% of the workforce experienced loss in productivity due to chronic pain . This report highlighted that 76% of the production time loss was due to diminished performance of staff suffering from pain, rather than absenteeism .
Many chronic pain patients look to the Internet to learn more about their condition, treatment and prognosis. However, a great deal of online information about pain may be inaccurate in one or more elements. An investigation at a pain clinic in the Netherlands found that patients felt confident about the credibility of the information ascertained from online resources, and as a result only half of the patients discussed the knowledge acquired from websites with physicians . This practice of unconditionally accepting online health information can be harmful as another study, from the University of Western Australia, found that chronic pain information on the Internet is poor in quality . The majority of the 27 websites that were evaluated were rated as fair or poor. This study mentions that though there may be high quality chronic pain content available online that is relevant and reliable, it is difficult to find amongst all the content that is of poor quality . To assist patients in the process of selecting quality sites, health related websites should be judged by both the quality of health information presented and design features that facilitate or impede use. The DISCERN quality index tool is one way to evaluate the reliability and quality of online health information and treatment choices .
The DISCERN questionnaire
Several solutions have been proposed to address the issues of accuracy, quality and reliability of information found on websites. These solutions have included electronic filtering of web-based information, creation of ethical codes of conduct for providers of web-based information (currently done on a voluntary basis) and assessment of websites by health professionals [16, 17]. An additional approach is to design and publicize instruments to assess the reliability and utility of information found on websites. This allows consumers to appraise web-based information themselves . The DISCERN questionnaire is a valid and reliable instrument for analyzing written consumer health information. It is the first standardized quality index of consumer health information that can be used as a critical appraisal tool to evaluate health information by not only health professionals, but also by patients and the general population. This questionnaire was derived systematically with the input of an expert panel, health information providers and patients from a self-help group . Thus, the purpose of this study was to determine the quality and readability of pain management websites and factors that can explain variability in the quality and grade level readability scores.
A list of search terms commonly used by patients suffering from pain was obtained from published literature. These terms included "Pain", "Chronic Pain", "Back pain", "Arthritis", and "Fibromyalgia" . Each of these keywords were entered into three different search engines (Google, Yahoo, and MSN) chosen because of their popularity . The first 20 links reported by each search engine per keyword were evaluated for quality using the DISCERN questionnaire and readability through the Flesch-Kincaid Algorithm [9, 20].
Websites were included in the investigation if they provided detailed information pertaining to the five search terms and treatment options for patients suffering from the conditions. Websites that were unrelated to pain or only provided a list of website links were removed. Sponsored links and banner advertisements were excluded as they are normally ignored . Websites that were considered 'For profit' were included only if they attempted to educate patients about pain conditions, and presented products with scientific evidence. 'For profit' websites were excluded: if their only intention was to sell a product, if the site promised quick and unrealistic dramatic results, made claims that one remedy will cure a variety of illnesses through some miraculous breakthrough, or used excessive sensational writing .
DISCERN questionnaire scoring criteria
The DISCERN questionnaire consists of 16 questions, on a continuous rating scale of 1 to 5, where 1 = definite NO and 5 = definite YES. Any rating in between (such as a 2, 3 or 4) suggests that some of the elements asked of by the question are present to a certain extent. These questions are categorized into three sections . Section 1 (questions 1 to 8) assesses reliability, dependability and trustworthiness of a website; Section 2 (questions 9 to 15) focuses on the quality of information about treatment choices; and Section 3 (question 16), evaluates overall quality rating on a continuous rating scale for the online website with a ratings of 1 = Low to 5 = High. Rating for question 16 is done independently of the rating given for the other previous 15 questions.
The readability of an online webpage refers to the level of reading difficulty found in written passages . Two factors comprise the readability of an online passage: 1) average sentence length in words -- the average of the numeric word count found in a passage and 2) average word length in syllables -- the number of syllables per 100 words. A lower readability score found using the Flesch-Kincaid algorithm indicates an easier reading level. Readability was assessed using the cross-platform, open source Java Flesch 2.0 Software .
The first stage of this investigation involved determining attributes found on websites. A random sample of 25 websites was assessed by the three reviewers to determine the characteristics most commonly present on webpages. It was decided that each subsequent website would be evaluated on potential for commercial gain, website seals of approval (for example, Health on Net), language(s) and multimedia (for example, online videos, audio recordings). The number of websites presenting these characteristics or lacking them was recorded. Each website was also evaluated on quality, as indicated by the DISCERN score tabulated and grade level readability, based on the Flesch-Kincaid algorithm. Overall and characteristic specific average DISCERN and readability scores were also calculated, with means (standard deviation (SD)), presented.
This investigation had three reviewers assess the quality of online health information. VBD and JK assessed the online information from Google, VBD and WD assessed Yahoo, and JK and WD assessed the information found on MSN. Each reviewer independently assessed the same list of websites and generated a DISCERN Score which was averaged and used for statistical purposes. In order to assess the level of agreement in the DISCERN rating scores between the reviewers; the chance corrected agreement (weighted kappa) value was generated .
The effect of each website characteristic on the variability in DISCERN and readability scores were assessed using univariate and multivariable analyses. No a priori hypotheses were generated for the effect of each website characteristic on quality and readability scores. Thus, the analysis in this study is primarily exploratory and meant to generate hypothesis for further investigation in larger studies. Model assumptions were assessed using the Normal Probability PP and QQ plot and by examining the residuals.
Website characteristics were compared independently to DISCERN and readability scores using univariate regression analyses. The characteristics were selected for inclusion in the multivariable analysis using alpha = 0.20 during the univariarate analysis. Characteristics from the univariate analyses that were of interest were further investigated by a multivariable regression using alpha = 0.05. The results were reported as estimates of the model coefficient Beta, corresponding to the 95% Confidence Interval (CI) and associated P-values. All P-values are reported to three decimals places with less than 0.001 reported as P < 0.001. Finally, for this investigation, reviewer agreement on website data abstraction was quantified using the weighted kappa statistic from the MedCalc (Version 11.3) Statistical Software [Mariakerke, Belgium] . Descriptive and regression analyses were carried out using SPSS software Version 17 (Chicago, IL, USA).
Average DISCERN and readability scores
Descriptive table: n = 161 (total number of unique websites included)
Statistic N (% of total number of unique websites)
Potential for commercial gain
Health related seals of approval
English and another language
Summary of statistics of the DISCERN scores and readability
DISCERN score (Maximum Score of 80)
Readability score (grade level)*
Potential for commercial gain
Yes, n = 11
No, n = 150
Health related website seals of approval
Yes, n = 58
No, n = 103
Only English, n = 122
English and another language(s), n = 39
Yes, n = 65
No, n = 96
Factors associated with variability in DISCERN and readability scores
Univariate and multivariable results using total quality DISCERN scores as an outcome
Coefficient (95% CI)
Coefficient (95% CI)
Potential for commercial gain
Health related seals of approval
5.4 (1.1, 9.8)
Language (English or English and another language)
Univariate and multivariable results using readability grade level scores as an outcome
Coefficient (95% CI)
Coefficient (95% CI)
Health related seals of approval
-0.9 (-2.2, 0.4)
-0.7 (-2.0, 0.5)
Potential for commercial gain
Language (English or English and another language)
The Internet has the potential to rapidly provide both patients and health care providers with access to health information. With its growing use there is increasing concern about the quality of online health information, as well as variability amongst health websites [13, 16, 24]. The American Medical Association (AMA) has published guidelines to aid patients in search of health information on the Internet and they also address the variability in patient health literacy, quality of content and access to information online [16, 26]. It is therefore imperative to use a consistent approach to the assessment of online health information commonly found by patients. In this investigation, the DISCERN questionnaire was used to evaluate the quality of websites and the Flesch-Kincaid algorithm was used to assess grade level readability.
The mean DISCERN score for the 161 unique websites was 55.9 of a maximum of 80 with a SD of 3.9. This suggests information of moderate quality, with potentially important but not serious shortcomings. The majority of websites found were above the sixth grade reading level recommended for patient directed literature as a mean Readability Grade Level Score was found to be 10.9 (3.9), an intermediate grade level (see Tables 1 and 2) .
The potential implications of this investigation for clinical practice, patient care and protocols pertaining to online health information are widespread. Although only the relationship between website seals of approval and DISCERN scores was found to be significant, several factors contribute to higher DISCERN and lower readability scores. Beta coefficient values suggest that potential for commercial gain and health related seals of approval result in high DISCERN scores while lower readability is associated with health related seals of approval and interactive multimedia.
Health related seals of approval are attempting to unify the quality of medical information on the Internet and offers the audience an indication of the online provider's commitment to providing quality information . The coefficient beta values found in this investigation for seal of approval in regards to both the DISCERN and the readability scores suggests that it is advantageous for more health websites to strive for a seal of approval certification and for patient/health care providers to seek such sites. Lower readability scores were correlated with websites offering interactive multimedia. This may occur because when developing information for multiple media, information must be reviewed on numerous occasions with the resulting information presented more succinctly. Finally, those websites engaged in selling commercial products have a need to present a clear sales message and therefore may achieve high DISCERN scores. This enables patients (potential consumers) to understand their medical conditions better and make informed decisions for their treatment options. The observations of this study would suggest that patients attempting to obtain health information online should start by selecting websites that are interactive, display seals of approval, are easy to navigate and provide commercial treatment options. We realize that this latter recommendation may be controversial because of the conflict of interest inherent in the use of health related information to promote product sales. On the other hand, it is to the advantage of the commercial entity to have its products used correctly by the appropriate patient population.
In terms of uniqueness, this is the first study that analyzes pain websites using the highly reliable DISCERN tool and also provides novel insight into the variability of quality and readability scores when examining online websites. Previous studies using the DISCERN tool have focused on the quality of websites for pain , low back pain [30, 31], rheumatoid arthritis [32, 33], burn scar management , and treatment for cough in children  have used different tools. Similar to this investigation, these studies repeatedly reported that websites on the Internet are of moderate quality, and are not consistent in keeping pace with new research.
Despite attempts towards the standardization of online health information, only 32.2% of websites reviewed presented with a health related seals of approval. This investigation found health related seals of approval to not only be an important factor in high DISCERN scores, assessing website content and treatment options but also the most significant factor for low grade level readability scores. In addition to health related seals of approval, interactive multimedia options, offered in 40.4% of websites were found to be a significant factor in low readability scores. Finally, while only 6.8% of the websites evaluated offered a commercial solution for patients with pain conditions, that factor contributed most to high DISCERN scores.
Websites are constantly being updated or removed and new ones are emerging, all of which may change the quality of websites found in this study. For practical reasons, only websites in English were considered. It may be advantageous to investigate search engines in other languages to determine if any discrepancies exist in the quality of the online health information. Only the top 20 websites were investigated for each chronic pain search term. It may be worthwhile to compare the quality of online health information from the top 50 websites and not just the top 20 as done here from each search engine, as quality may vary. Finally, the DISCERN instrument is an effective tool for assessing the quality of online information as it pertains to treatment, but less effective in evaluating other website information.
The overall quality of chronic pain websites is moderate, with some shortcomings that need to be addressed. We have found that websites which contain health related seals of approval and offer information for alterative commercial solutions to pain related conditions have higher DISCERN scores. Lower readability levels were again found in websites with health related seals of approval and also interactive multimedia options.
American Medical Association.
The authors would like to thank McMaster University's Department of Anesthesia for their support and guidance throughout the investigation. We thank the reviewers for comments that improved the presentation.
- Eysenbach G, Powell J, Kuss O, Sa ER: Empirical studies assessing the quality of health information for consumers on the world wide web: a systematic review. JAMA. 2002, 20: 2691-2700. 10.1001/jama.287.20.2691.View ArticleGoogle Scholar
- U.S Department of Commerce-National Telecommunications and Information Administration (NTIA). Falling through the Net: Toward Digital Inclusion- A Report on Americans' Access to Technology Tools. [http://www.ntia.doc.gov/ntiahome/fttn00/falling.htm]
- Cline RJ, Haynes KM: Consumer health information seeking on the Internet: the state of the art. Health Educ Res. 2001, 6: 671-692. 10.1093/her/16.6.671.View ArticleGoogle Scholar
- U.S Department of Commerce, National Telecommunications and Information Administration (NTIA). A Nation Online: How Americans Are Expanding Their Use Of The Internet. [http://www.ntia.doc.gov/ntiahome/dn/nationonline_020502.htm]
- Miniwatts Marketing Group, World Internet Users and Population Stats. [http://www.Internetworldstats.com/stats.htm]
- Statistics Canada, Canadian Internet Use Survey-2009. [http://www.statcan.gc.ca]
- Maloney S, Ilic D, Green S: Accessibility, nature and quality of health information on the Internet: a survey on osteoarthritis. Rheumatology. 2005, 3: 382-385. 10.1093/rheumatology/keh498.View ArticleGoogle Scholar
- Pew Internet and American Life Project, The online health care revolution: How the Web helps Americans take better care of themselves. [http://pewinternet.org/Reports/2000/The-Online-Health-Care-Revolution.aspx]
- Moody EM, Clemens KK, Storsley L, Waterman A, Parikh CR, Garg AX, Donor Nephrectomy Outcomes Research (Donor) Network: Improving on-line information for potential living kidney donors. Kidney Int. 2007, 10: 1062-1070. 10.1038/sj.ki.5002168.View ArticleGoogle Scholar
- Henry JL: The need for knowledge translation in chronic pain. Pain Res Manag. 2008, 6: 465-476.View ArticleGoogle Scholar
- Statistics Canada, National Population Health Survey-1994/1995. [http://www.statcan.gc.ca]
- Stewart WF, Ricci JA, Chee E, Morganstein D: Lost productive work time costs from health conditions in the United States: results from the American Productivity Audit. J Occup Environ Med. 2003, 12: 1234-1246.View ArticleGoogle Scholar
- De Boer MJ, Versteegen GJ, van Wijhe M: Patients' use of the Internet for pain-related medical information. Patient Educ Couns. 2007, 1: 86-97. 10.1016/j.pec.2007.05.012.View ArticleGoogle Scholar
- Corcoran TB, Haigh F, Seabrook A, Schug SA: The quality of internet-sourced information for patients with chronic pain is poor. Clin J Pain. 2009, 25: 617-623. 10.1097/AJP.0b013e3181a5b5d5.View ArticlePubMedGoogle Scholar
- Khazaal Y, Chatton A, Cochand S, Coquard O, Fernandez S, Khan R, Billieux J, Zullino D: Brief DISCERN, six questions for the evaluation of evidence-based content of health-related websites. Patient Educ Couns. 2009, 1: 33-37. 10.1016/j.pec.2009.02.016.View ArticleGoogle Scholar
- Berland GK, Elliott MN, Morales LS, Algazy JI, Kravitz RL, Broder MS, Kanouse DE, Muñoz JA, Puyol JA, Lara M, Watkins KE, Yang H, McGlynn EA: Health information on the Internet: accessibility, quality, and readability in English and Spanish. JAMA. 2001, 20: 2612-2621. 10.1001/jama.285.20.2612.View ArticleGoogle Scholar
- Batchelor JM, Ohya Y: Use of the DISCERN instrument by patients and health professionals to assess information resources on treatments for asthma and atopic dermatitis. Allergol Int. 2009, 1: 141-145. 10.2332/allergolint.08-SC-0022.View ArticleGoogle Scholar
- Washington TA, Fanciullo GJ, Sorensen JA, Baird JC: Quality of Chronic Pain Websites. Pain Medicine. 2008, 9: 86-97. 10.1111/j.1526-4637.2008.00419.x.View ArticleGoogle Scholar
- Sullivan D, Nielsen net ratings search engine ratings. [http://searchenginewatch.com]
- Flesch R: A new readability yardstick. J Appl Psychol. 1948, 3: 221-233. 10.1037/h0057532.View ArticleGoogle Scholar
- Food and Drug Administration. How to Evaluate Health Information on the Internet. [http://www.fda.gov]
- Charnock D: The DISCERN Handbook. 1998, Abingdon, Oxford: Radcliffe Medical PressGoogle Scholar
- Flesch-Kincade : Flesch 2.0 [Software]. 2007Google Scholar
- Sim J, Wright CC: The kappa statistic in reliability studies: use, interpretation, and sample size requirements. Phys Ther. 2005, 3: 257-268.Google Scholar
- MedCalc Software: MedCalc (Version 11.2). 1993Google Scholar
- Charnock D, Shepperd S, Needham G, Gann R: DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Community Health. 1999, 2: 105-111. 10.1136/jech.53.2.105.View ArticleGoogle Scholar
- Davis TC, Crouch MA, Wills G, Miller S, Abdehou DM: The gap between patient reading comprehension and the readability of patient education materials. J Fam Pract. 1990, 5: 533-538.Google Scholar
- Boyer C, Selby M, Scherrer JR, Appel RD: The Health On the Net Code of Conduct for medical and health Websites. Comput Biol Med. 1998, 5: 603-610. 10.1016/S0010-4825(98)00037-7.View ArticleGoogle Scholar
- Lorence D, Abraham J: A study of undue pain and surfing: using hierarchical criteria to assess website quality. Health Informatics J. 2008, 3: 155-173. 10.1177/1081180X08092827.View ArticleGoogle Scholar
- Greene DL, Appel AJ, Reinert SE, Palumbo MA: Lumbar disc herniation: evaluation of information on the internet. Spine. 2005, 7: 826-829. 10.1097/01.brs.0000157754.98023.cd.View ArticleGoogle Scholar
- Li L, Irvin E, Guzman J, Bombardier C: Surfing for back pain patients: the nature and quality of back pain information on the Internet. Spine. 2001, 5: 545-557. 10.1097/00007632-200103010-00020.View ArticleGoogle Scholar
- Culver M, Chadwick A: Internet information on rheumatoid arthritis: an evaluation. Musculoskeletal Care. 2005, 1: 33-43. 10.1002/msc.24.View ArticleGoogle Scholar
- Newman MA, Ziebland S, Barker KL: Patients' views of a multimedia resource featuring experiences of rheumatoid arthritis: pilot evaluation of. Health Informatics J. 2009, 2: 147-159. 10.1177/1460458209102974. [http://www.healthtalkonline.org]View ArticleGoogle Scholar
- Bohacek L, Gomez M, Fish JS: An evaluation of internet sites for burn scar management. J Burn Care Rehabil. 2003, 4: 246-251. 10.1097/01.BCR.0000075844.04297.D9.View ArticleGoogle Scholar
- Pandolfini C, Impicciatore P, Bonati M: Parents on the web: risks for quality management of cough in children. Pediatrics. 2000, 1: e1-10.1542/peds.105.1.e1.View ArticleGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1741-7015/8/59/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.