Evaluating the Dental Caries-Related Information on Brazilian Websites: Qualitative Study

BACKGROUND

Dental caries is the most common chronic oral disease, affecting 2.4 billion people worldwide who on average have 2.11 decayed, missing, or filled teeth. It impacts the quality of life of patients, socially and economically. However, the comprehension of dental caries may be difficult for most people, as it involves a multifactorial etiology with the interplay between the tooth surface, the dental biofilm, dietary fermentable carbohydrates, and genetic and behavioral factors. Therefore, the production of effective materials addressed to the education and counseling of patients for the prevention of dental caries requires a high level of specialization. In this regard, the dental caries-related contents produced by laypersons and their availability on the Internet may be low-quality information.

OBJECTIVE

The aim of this study was to assess the readability and the quality of dental caries-related information on Brazilian websites.

METHODS

A total of 75 websites were selected through Google, Bing, Yahoo!, and Baidu. The websites were organized in rankings according to their order of appearance in each one of the 4 search engines. Furthermore, 2 independent examiners evaluated the quality of websites using the DISCERN questionnaire and the Journal of American Medical Association (JAMA) benchmark criteria. The readability of the websites was assessed by the Flesch Reading Ease adapted to Brazilian Portuguese (FRE-BP). In addition, the information presented on the websites was categorized as etiology, prevention, and treatment of dental caries. The statistical analysis was performed using Spearman rank correlation coefficient, Mann-Whitney U test, hierarchical clustering analysis by Ward minimum variance method, Kruskal-Wallis test, and post hoc Dunn test. P<.05 was considered significant.

RESULTS

The Web contents were considered to be of poor quality by DISCERN (mean 33.48, standard deviation, SD 9.06) and JAMA (mean 1.12, SD 0.97) scores, presenting easy reading levels (FRE-BP: mean 62.93, SD 10.15). The rankings of the websites presented by Google (ρ=-.22, P=.08), Baidu (ρ=-.19, P=.53), Yahoo! (ρ=.22, P=.39), and Bing (ρ=-.36, P=.23) were not correlated with DISCERN scores. Moreover, the quality of websites with health- and nonhealth-related authors was similar (P=.27 for DISCERN and P=.47 for JAMA); however, the pages with a greater variety of dental caries information showed significantly higher quality scores than those with limited contents (P=.009).

CONCLUSIONS

On the basis of this sample, dental caries-related contents available on Brazilian websites were considered simple, accessible, and of poor quality, independent of their authorship. These findings indicate the need for the development of specific policies focused on the stimulus for the production and publication of Web health information, encouraging dentists to guide their patients in searching for recommended oral health websites.

Improving a full-text search engine: the importance of negation detection and family history context to identify cases in a biomedical data warehouse

Objective

The repurposing of electronic health records (EHRs) can improve clinical and genetic research for rare diseases. However, significant information in rare disease EHRs is embedded in the narrative reports, which contain many negated clinical signs and family medical history. This paper presents a method to detect family history and negation in narrative reports and evaluates its impact on selecting populations from a clinical data warehouse (CDW).

Materials and Methods

We developed a pipeline to process 1.6 million reports from multiple sources. This pipeline is part of the load process of the Necker Hospital CDW.

Results

We identified patients with "Lupus and diarrhea," "Crohn's and diabetes," and "NPHP1" from the CDW. The overall precision, recall, specificity, and F-measure were 0.85, 0.98, 0.93, and 0.91, respectively.

Conclusion

The proposed method generates a highly accurate identification of cases from a CDW of rare disease EHRs.

Google Trends can improve surveillance of Type 2 diabetes

Recent studies demonstrate that people are increasingly looking online to assess their health, with reasons varying from personal preferences and beliefs to inability to book a timely appointment with their local medical practice. Records of these activities represent a new source of data about the health of populations, but which is currently unaccounted for by disease surveillance models. This could potentially be useful as evidence of individuals' perception of bodily changes and self-diagnosis of early symptoms of an emerging disease. We make use of the Experian geodemographic Mosaic dataset in order to extract Type 2 diabetes candidate risk variables and compare their temporal relationships with the search keywords, used to describe early symptoms of the disease on Google. Our results demonstrate that Google Trends can detect early signs of diabetes by monitoring combinations of keywords, associated with searches for hypertension treatment and poor living conditions; Combined search semantics, related to obesity, how to quit smoking and improve living conditions (deprivation) can be also employed, however, may lead to less accurate results.

Did online publishers "get it right"? Using a naturalistic search strategy to review cognitive health promotion content on internet webpages

BACKGROUND

One of the most common uses of the Internet is to search for health-related information. Although scientific evidence pertaining to cognitive health promotion has expanded rapidly in recent years, it is unclear how much of this information has been made available to Internet users. Thus, the purpose of our study was to assess the reliability and quality of information about cognitive health promotion encountered by typical Internet users.

METHODS

To generate a list of relevant search terms employed by Internet users, we entered seed search terms in Google Trends and recorded any terms consistently used in the prior 2 years. To further approximate the behaviour of typical Internet users, we entered each term in Google and sampled the first two relevant results. This search, completed in October 2014, resulted in a sample of 86 webpages, 48 of which had content related to cognitive health promotion. An interdisciplinary team rated the information reliability and quality of these webpages using a standardized measure.

RESULTS

We found that information reliability and quality were moderate, on average. Just one retrieved page mentioned best practice, national recommendations, or consensus guidelines by name. Commercial content (i.e., product promotion, advertising content, or non-commercial) was associated with differences in reliability and quality, with product promoter webpages having the lowest mean reliability and quality ratings.

CONCLUSIONS

As efforts to communicate the association between lifestyle and cognitive health continue to expand, we offer these results as a baseline assessment of the reliability and quality of cognitive health promotion on the Internet.

Abyss or Shelter? On the Relevance of Web Search Engines' Search Results When People Google for Suicide

Despite evidence that suicide rates can increase after suicides are widely reported in the media, appropriate depictions of suicide in the media can help people to overcome suicidal crises and can thus elicit preventive effects. We argue on the level of individual media users that a similar ambivalence can be postulated for search results on online suicide-related search queries. Importantly, the filter bubble hypothesis (Pariser, 2011) states that search results are biased by algorithms based on a person's previous search behavior. In this study, we investigated whether suicide-related search queries, including either potentially suicide-preventive or -facilitative terms, influence subsequent search results. This might thus protect or harm suicidal Internet users. We utilized a 3 (search history: suicide-related harmful, suicide-related helpful, and suicide-unrelated) × 2 (reactive: clicking the top-most result link and no clicking) experimental design applying agent-based testing. While findings show no influences either of search histories or of reactivity on search results in a subsequent situation, the presentation of a helpline offer raises concerns about possible detrimental algorithmic decision-making: Algorithms "decided" whether or not to present a helpline, and this automated decision, then, followed the agent throughout the rest of the observation period. Implications for policy-making and search providers are discussed.

Querying EHRs with a Semantic and Entity-Oriented Query Language

While the digitization of medical documents has greatly expanded during the past decade, health information retrieval has become a great challenge to address many issues in medical research. Information retrieval in electronic health records (EHR) should also reduce the difficult tasks of manual information retrieval from records in paper format or computer. The aim of this article was to present the features of a semantic search engine implemented in EHRs. A flexible, scalable and entity-oriented query language tool is proposed. The program is designed to retrieve and visualize data which can support any Conceptual Data Model. The search engine deals with structured and unstructured data, for a sole patient from a caregiver perspective, and for a number of patients (e.g. epidemiology). Several types of queries on a test database containing 2,000 anonymized patients EHRs (i.e. approximately 200,000 records) were tested. These queries were able to accurately treat symbolic, textual, numerical and chronological data.

PubMed search filters for the study of putative outdoor air pollution determinants of disease

OBJECTIVES

Several PubMed search filters have been developed in contexts other than environmental. We aimed at identifying efficient PubMed search filters for the study of environmental determinants of diseases related to outdoor air pollution.

METHODS

We compiled a list of Medical Subject Headings (MeSH) and non-MeSH terms seeming pertinent to outdoor air pollutants exposure as determinants of diseases in the general population. We estimated proportions of potentially pertinent articles to formulate two filters (one 'more specific', one 'more sensitive'). Their overall performance was evaluated as compared with our gold standard derived from systematic reviews on diseases potentially related to outdoor air pollution. We tested these filters in the study of three diseases potentially associated with outdoor air pollution and calculated the number of needed to read (NNR) abstracts to identify one potentially pertinent article in the context of these diseases. Last searches were run in January 2016.

RESULTS

The 'more specific' filter was based on the combination of terms that yielded a threshold of potentially pertinent articles ≥40%. The 'more sensitive' filter was based on the combination of all search terms under study. When compared with the gold standard, the 'more specific' filter reported the highest specificity (67.4%; with a sensitivity of 82.5%), while the 'more sensitive' one reported the highest sensitivity (98.5%; with a specificity of 47.9%). The NNR to find one potentially pertinent article was 1.9 for the 'more specific' filter and 3.3 for the 'more sensitive' one.

CONCLUSIONS

The proposed search filters could help healthcare professionals investigate environmental determinants of medical conditions that could be potentially related to outdoor air pollution.

Accuracy of weight loss information in Spanish search engine results on the internet

OBJECTIVE

To systematically assess the quality of online information related to weight loss that Spanish speakers in the U.S. are likely to access.

METHODS

This study evaluated the accessibility and quality of information for websites that were identified from weight loss queries in Spanish and compared this with previously published results in English. The content was scored with respect to five dimensions: nutrition, physical activity, behavior, pharmacotherapy, and surgical recommendations.

RESULTS

Sixty-six websites met eligibility criteria (21 commercial, 24 news/media, 10 blogs, 0 medical/government/university, 11 unclassified sites). Of 16 possible points, mean content quality score was 3.4 (SD = 2.0). Approximately 1.5% of sites scored greater than 8 (out of 12) on nutrition, physical activity, and behavior. Unsubstantiated claims were made on 94% of the websites. Content quality scores varied significantly by type of website (P < 0.0001) with unclassified websites scoring the highest (mean = 6.3, SD = 1.4) and blogs scoring the lowest (mean = 2.2, SD = 1.2). All content quality scores were lower for Spanish websites relative to English websites.

CONCLUSIONS

Weight loss information accessed in Spanish Web searches is suboptimal and relatively worse than weight loss information accessed in English, suggesting that U.S. Spanish speakers accessing weight loss information online may be provided with incomplete and inaccurate information.

A Probabilistic Reformulation of No Free Lunch: Continuous Lunches Are Not Free

No Free Lunch (NFL) theorems have been developed in many settings over the last two decades. Whereas NFL is known to be possible in any domain based on set-theoretic concepts, probabilistic versions of NFL are presently believed to be impossible in continuous domains. This article develops a new formalization of probabilistic NFL that is sufficiently expressive to prove the existence of NFL in large search domains, such as continuous spaces or function spaces. This formulation is arguably more complicated than its set-theoretic variants, mostly as a result of the numerous technical complications within probability theory itself. However, a probabilistic conceptualization of NFL is important because stochastic optimization methods inherently need to be evaluated probabilistically. Thus the present study fills an important gap in the study of performance of stochastic optimizers.

Towards Identifying and Reducing the Bias of Disease Information Extracted from Search Engine Data

The estimation of disease prevalence in online search engine data (e.g., Google Flu Trends (GFT)) has received a considerable amount of scholarly and public attention in recent years. While the utility of search engine data for disease surveillance has been demonstrated, the scientific community still seeks ways to identify and reduce biases that are embedded in search engine data. The primary goal of this study is to explore new ways of improving the accuracy of disease prevalence estimations by combining traditional disease data with search engine data. A novel method, Biased Sentinel Hospital-based Area Disease Estimation (B-SHADE), is introduced to reduce search engine data bias from a geographical perspective. To monitor search trends on Hand, Foot and Mouth Disease (HFMD) in Guangdong Province, China, we tested our approach by selecting 11 keywords from the Baidu index platform, a Chinese big data analyst similar to GFT. The correlation between the number of real cases and the composite index was 0.8. After decomposing the composite index at the city level, we found that only 10 cities presented a correlation of close to 0.8 or higher. These cities were found to be more stable with respect to search volume, and they were selected as sample cities in order to estimate the search volume of the entire province. After the estimation, the correlation improved from 0.8 to 0.864. After fitting the revised search volume with historical cases, the mean absolute error was 11.19% lower than it was when the original search volume and historical cases were combined. To our knowledge, this is the first study to reduce search engine data bias levels through the use of rigorous spatial sampling strategies.

Use of Internet Search Queries to Enhance Surveillance of Foodborne Illness

“Food poisoning” queries were correlated with the number of foodborne illness–related hospital stays.

As a supplement to or extension of methods used to determine trends in foodborne illness over time, we propose the use of Internet search metrics. We compared Internet query data for foodborne illness syndrome–related search terms from the most popular 5 Korean search engines using Health Insurance Review and Assessment Service inpatient stay data for 26 International Classification of Diseases, Tenth Revision, codes for foodborne illness in South Korea during 2010–2012. We used time-series analysis with Seasonal Autoregressive Integrated Moving Average (SARIMA) models. Internet search queries for “food poisoning” correlated most strongly with foodborne illness data (r = 0.70, p<0.001); furthermore, “food poisoning” queries correlated most strongly with the total number of inpatient stays related to foodborne illness during the next month (β = 0.069, SE 0.017, p<0.001). This approach, using the SARIMA model, could be used to effectively measure trends over time to enhance surveillance of foodborne illness in South Korea.

Collection of Medical Original Data with Search Engine for Decision Support

Medicine is becoming more and more complex and humans can capture total medical knowledge only partially. For specific access a high resolution search engine is demonstrated, which allows besides conventional text search also search of precise quantitative data of medical findings, therapies and results. Users can define metric spaces ("Domain Spaces", DSs) with all searchable quantitative data ("Domain Vectors", DSs). An implementation of the search engine is online in http://numericsearch.com. In future medicine the doctor could make first a rough diagnosis and check which fine diagnostics (quantitative data) colleagues had collected in such a case. Then the doctor decides about fine diagnostics and results are sent (half automatically) to the search engine which filters a group of patients which best fits to these data. In this specific group variable therapies can be checked with associated therapeutic results, like in an individual scientific study for the current patient. The statistical (anonymous) results could be used for specific decision support. Reversely the therapeutic decision (in the best case with later results) could be used to enhance the collection of precise pseudonymous medical original data which is used for better and better statistical (anonymous) search results.

Evaluating Open-Source Full-Text Search Engines for Matching ICD-10 Codes

This research presents the results of evaluating multiple free, open-source engines on matching ICD-10 diagnostic codes via full-text searches. The study investigates what it takes to get an accurate match when searching for a specific diagnostic code. For each code the evaluation starts by extracting the words that make up its text and continues with building full-text search queries from the combinations of these words. The queries are then run against all the ICD-10 codes until a match indicates the code in question as a match with the highest relative score. This method identifies the minimum number of words that must be provided in order for the search engines choose the desired entry. The engines analyzed include a popular Java-based full-text search engine, a lightweight engine written in JavaScript which can even execute on the user's browser, and two popular open-source relational database management systems.

Search Engine Ranking, Quality, and Content of Web Pages That Are Critical Versus Noncritical of Human Papillomavirus Vaccine

PURPOSE

Online information can influence attitudes toward vaccination. The aim of the present study was to provide a systematic evaluation of the search engine ranking, quality, and content of Web pages that are critical versus noncritical of human papillomavirus (HPV) vaccination.

METHODS

We identified HPV vaccine-related Web pages with the Google search engine by entering 20 terms. We then assessed each Web page for critical versus noncritical bias and for the following quality indicators: authorship disclosure, source disclosure, attribution of at least one reference, currency, exclusion of testimonial accounts, and readability level less than ninth grade. We also determined Web page comprehensiveness in terms of mention of 14 HPV vaccine-relevant topics.

RESULTS

Twenty searches yielded 116 unique Web pages. HPV vaccine-critical Web pages comprised roughly a third of the top, top 5- and top 10-ranking Web pages. The prevalence of HPV vaccine-critical Web pages was higher for queries that included term modifiers in addition to root terms. Compared with noncritical Web pages, Web pages critical of HPV vaccine overall had a lower quality score than those with a noncritical bias (p < .01) and covered fewer important HPV-related topics (p < .001). Critical Web pages required viewers to have higher reading skills, were less likely to include an author byline, and were more likely to include testimonial accounts. They also were more likely to raise unsubstantiated concerns about vaccination.

CONCLUSIONS

Web pages critical of HPV vaccine may be frequently returned and highly ranked by search engine queries despite being of lower quality and less comprehensive than noncritical Web pages.

PlateRunner: A Search Engine to Identify EMR Boilerplates

Medical text contains boilerplated content, an artifact of pull-down forms from EMRs. Boilerplated content is the source of challenges for concept extraction on clinical text. This paper introduces PlateRunner, a search engine on boilerplates from the US Department of Veterans Affairs (VA) EMR. Boilerplates containing concepts should be identified and reviewed to recognize challenging formats, identify high yield document titles, and fine tune section zoning. This search engine has the capability to filter negated and asserted concepts, save and search query results. This tool can save queries, search results, and documents found for later analysis.

Stochastic Leader Gravitational Search Algorithm for Enhanced Adaptive Beamforming Technique

In this paper, stochastic leader gravitational search algorithm (SL-GSA) based on randomized k is proposed. Standard GSA (SGSA) utilizes the best agents without any randomization, thus it is more prone to converge at suboptimal results. Initially, the new approach randomly choses k agents from the set of all agents to improve the global search ability. Gradually, the set of agents is reduced by eliminating the agents with the poorest performances to allow rapid convergence. The performance of the SL-GSA was analyzed for six well-known benchmark functions, and the results are compared with SGSA and some of its variants. Furthermore, the SL-GSA is applied to minimum variance distortionless response (MVDR) beamforming technique to ensure compatibility with real world optimization problems. The proposed algorithm demonstrates superior convergence rate and quality of solution for both real world problems and benchmark functions compared to original algorithm and other recent variants of SGSA.

Employing complex polyhierarchical ontologies and promoting interoperability of i2b2 data systems

I2b2 is in widespread use for managing research data warehouses. It employs reference ontologies as a record index and supports searching for aggregate cases using a pattern match operator on ASCII strings representing the node traversal from root to concept(PATHs). This creates complexities in dissemination and deployment for large polyhierarchical ontologies such as SNOMED CT. We hypothesized that an alternative approach employing transitive closure tables (TC) could lead to more accurate, efficient and interoperable search tools for i2b2. We evaluated search speed, accuracy and interoperability of queries employing each approach. We found both TC-based and PATH-based queries to produce accurate results. However, we observed that TC-based queries involving concepts included in large numbers of paths ran substantially faster than PATH-based queries for the same concept. Oracle query plan resource estimates differed by one to three orders of magnitude for these queries. We conclude that a simplification of dissemination tools for SNOMED CT and revision in the metadata build for i2b2 can effectively employ SNOMED CT with increased efficiency and comparable accuracy. Use of transitive closure tables in metadata can promote network query interoperability.

Interpretation of Coronary Angiograms Recorded Using Google Glass: A Comparative Analysis

INTRODUCTION

Google Glass (Google, Inc) is a voice-activated, hands-free, optical head-mounted display device capable of taking pictures, recording videos, and transmitting data via wi-fi. In the present study, we examined the accuracy of coronary angiogram interpretation, recorded using Google Glass.

METHODS

Google Glass was used to record 15 angiograms with 17 major findings and the participants were asked to interpret those recordings on: (1) an iPad (Apple, Inc); or (2) a desktop computer. Interpretation was compared with the original angiograms viewed on a desktop. Ten physicians (2 interventional cardiologists and 8 cardiology fellows) participated. One point was assigned for each correct finding, for a maximum of 17 points.

RESULTS

The mean angiogram interpretation score for Google Glass angiogram recordings viewed on an iPad or a desktop vs the original angiograms viewed on a desktop was 14.9 ± 1.1, 15.2 ± 1.8, and 15.9 ± 1.1, respectively (P=.06 between the iPad and the original angiograms, P=.51 between the iPad and recordings viewed on a desktop, and P=.43 between the recordings viewed on a desktop and the original angiograms). In a post-study survey, one of the 10 physicians (10%) was "neutral" with the quality of the recordings using Google Glass, 6 physicians (60%) were "somewhat satisfied," and 3 physicians (30%) were "very satisfied."

CONCLUSION

This small pilot study suggests that the quality of coronary angiogram video recordings obtained using Google Glass may be adequate for recognition of major findings, supporting its expanding use in telemedicine.

Using homology relations within a database markedly boosts protein sequence similarity search

Inference of homology from protein sequences provides an essential tool for analyzing protein structure, function, and evolution. Current sequence-based homology search methods are still unable to detect many similarities evident from protein spatial structures. In computer science a search engine can be improved by considering networks of known relationships within the search database. Here, we apply this idea to protein-sequence-based homology search and show that it dramatically enhances the search accuracy. Our new method, COMPADRE (COmparison of Multiple Protein sequence Alignments using Database RElationships) assesses the relationship between the query sequence and a hit in the database by considering the similarity between the query and hit's known homologs. This approach increases detection quality, boosting the precision rate from 18% to 83% at half-coverage of all database homologs. The increased precision rate allows detection of a large fraction of protein structural relationships, thus providing structure and function predictions for previously uncharacterized proteins. Our results suggest that this general approach is applicable to a wide variety of methods for detection of biological similarities. The web server is available at prodata.swmed.edu/compadre.

Developing an Open-Source Bibliometric Ranking Website Using Google Scholar Citation Profiles for Researchers in the Field of Biomedical Informatics

We developed the Biomedical Informatics Researchers ranking website (rank.informatics-review.com) to overcome many of the limitations of previous scientific productivity ranking strategies. The website is composed of four key components that work together to create an automatically updating ranking website: (1) list of biomedical informatics researchers, (2) Google Scholar scraper, (3) display page, and (4) updater. The site has been useful to other groups in evaluating researchers, such as tenure and promotions committees in interpreting the various citation statistics reported by candidates. Creation of the Biomedical Informatics Researchers ranking website highlights the vast differences in scholarly productivity among members of the biomedical informatics research community.

A recommender system for medical imaging diagnostic

The large volume of data captured daily in healthcare institutions is opening new and great perspectives about the best ways to use it towards improving clinical practice. In this paper we present a context-based recommender system to support medical imaging diagnostic. The system relies on data mining and context-based retrieval techniques to automatically lookup for relevant information that may help physicians in the diagnostic decision.

Does query expansion limit our learning? A comparison of social-based expansion to content-based expansion for medical queries on the internet

Searching for medical information online is a common activity. While it has been shown that forming good queries is difficult, Google's query suggestion tool, a type of query expansion, aims to facilitate query formation. However, it is unknown how this expansion, which is based on what others searched for, affects the information gathering of the online community. To measure the impact of social-based query expansion, this study compared it with content-based expansion, i.e., what is really in the text. We used 138,906 medical queries from the AOL User Session Collection and expanded them using Google's Autocomplete method (social-based) and the content of the Google Web Corpus (content-based). We evaluated the specificity and ambiguity of the expansion terms for trigram queries. We also looked at the impact on the actual results using domain diversity and expansion edit distance. Results showed that the social-based method provided more precise expansion terms as well as terms that were less ambiguous. Expanded queries do not differ significantly in diversity when expanded using the social-based method (6.72 different domains returned in the first ten results, on average) vs. content-based method (6.73 different domains, on average).

Constrained novelty search: a study on game content generation

Novelty search is a recent algorithm geared toward exploring search spaces without regard to objectives. When the presence of constraints divides a search space into feasible space and infeasible space, interesting implications arise regarding how novelty search explores such spaces. This paper elaborates on the problem of constrained novelty search and proposes two novelty search algorithms which search within both the feasible and the infeasible space. Inspired by the FI-2pop genetic algorithm, both algorithms maintain and evolve two separate populations, one with feasible and one with infeasible individuals, while each population can use its own selection method. The proposed algorithms are applied to the problem of generating diverse but playable game levels, which is representative of the larger problem of procedural game content generation. Results show that the two-population constrained novelty search methods can create, under certain conditions, larger and more diverse sets of feasible game levels than current methods of novelty search, whether constrained or unconstrained. However, the best algorithm is contingent on the particularities of the search space and the genetic operators used. Additionally, the proposed enhancement of offspring boosting is shown to enhance performance in all cases of two-population novelty search.