Promises and Pitfalls in the Use of "Big Data" for Clinical Research

Relationship between internet research data of oral neoplasms and public health programs in the European Union

BACKGROUND

Tobacco and alcohol are the main risk factors for oral squamous cell carcinoma, the low survival rate of which is a public health problem. European-wide health policies (a prevention campaign, tobacco packaging) have been put in place to inform the population of the risks associated with consumption. Due to the increase in smoking among women, the incidence of this disease remains high. The identification of internet research data on the population could help to measure the impact of and better position these preventive measures. The objective was to analyze a potential temporal association between public health programs and interest in oral cancers on the internet in the European Union (EU).

METHODS

A search of data from Google ©, Wikipedia © and Twitter © users in 28 European countries relating to oral cancer between 2004 and 2019 was completed. Bibliometric analysis of press and scientific articles over the same period was also performed. The association between these data and the introduction of public health programs in Europe was studied.

RESULTS

There was a temporal association between changes in tobacco packaging and a significant increase in internet searches for oral cancer in seven countries. Unlike national policies and ad campaigns, the European awareness program Make Sense has had no influence on internet research. There was an asymmetric correlation in internet searches between publications on oral cancer from scientific articles or "traditional" media (weak association) and those from internet media such as Twitter © or Wikipedia © (strong association).

CONCLUSION

Our work highlights seven areas around which oral cancer awareness in Europe could be refocused, such as a change in the communication of health warnings on cigarette packs, the establishment of a more explicit campaign name regarding oral cancer, the involvement of public figures and associations in initiatives to be organized at the local level and the strengthening of awareness of the dangers of tobacco in the development of oral cancer.

Identification of important factors in an inpatient fall risk prediction model to improve the quality of care using EHR and electronic administrative data: A machine-learning approach

BACKGROUND

Inpatient falls, many resulting in injury or death, are a serious problem in hospital settings. Existing falls risk assessment tools, such as the Morse Fall Scale, give a risk score based on a set of factors, but don't necessarily signal which factors are most important for predicting falls. Artificial intelligence (AI) methods provide an opportunity to improve predictive performance while also identifying the most important risk factors associated with hospital-acquired falls. We can glean insight into these risk factors by applying classification tree, bagging, random forest, and adaptive boosting methods applied to Electronic Health Record (EHR) data.

OBJECTIVE

The purpose of this study was to use tree-based machine learning methods to determine the most important predictors of inpatient falls, while also validating each via cross-validation.

MATERIALS AND METHODS

A case-control study was designed using EHR and electronic administrative data collected between January 1, 2013 to October 31, 2013 in 14 medical surgical units. The data contained 38 predictor variables which comprised of patient characteristics, admission information, assessment information, clinical data, and organizational characteristics. Classification tree, bagging, random forest, and adaptive boosting methods were used to identify the most important factors of inpatient fall-risk through variable importance measures. Sensitivity, specificity, and area under the ROC curve were computed via ten-fold cross validation and compared via pairwise t-tests. These methods were also compared to a univariate logistic regression of the Morse Fall Scale total score.

RESULTS

In terms of AUROC, bagging (0.89), random forest (0.90), and boosting (0.89) all outperformed the Morse Fall Scale (0.86) and the classification tree (0.85), but no differences were measured between bagging, random forest, and adaptive boosting, at a p-value of 0.05. History of Falls, Age, Morse Fall Scale total score, quality of gait, unit type, mental status, and number of high fall risk increasing drugs (FRIDs) were considered the most important features for predicting inpatient fall risk.

CONCLUSIONS

Machine learning methods have the potential to identify the most relevant and novel factors for the detection of hospitalized patients at risk of falling, which would improve the quality of patient care, and to more fully support healthcare provider and organizational leadership decision-making. Nurses would be able to enhance their judgement to caring for patients at risk for falls. Our study may also serve as a reference for the development of AI-based prediction models of other iatrogenic conditions. To our knowledge, this is the first study to report the importance of patient, clinical, and organizational features based on the use of AI approaches.

Use of Data Analysis Methods in Dental Publications: Is There Evidence of a Methodological Change?

Objectives: To evaluate how data analysis methods in dental studies have changed in recent years. Methods: A total of 400 articles published in 2010 and 2017 in five dental journals, Journal of Dental Research, Caries Research, Community Dentistry and Oral Epidemiology, Journal of Dentistry, and Acta Odontologica Scandinavica, were analyzed. The study characteristics and the reporting of data analysis techniques were systematically identified. Results: The statistical intensity of the dental journals did not change from 2010 to 2017. Dental researchers did not adopt the data mining, machine learning, or Bayesian approaches advocated in the computer-oriented methodological literature. The determination of statistical significance was the most generally used method for conducting research in both 2010 and 2017. Observational study designs were more common in 2017. Insufficient and incomplete descriptions of statistical methods were still a serious problem. Conclusion: The stabilization of statistical intensity in the literature suggests that papers applying highly computationally complex data analysis methods have not meaningfully contributed to dental research or clinical care. Greater rigor is required in reporting the methods in dental research articles, given the current pervasiveness of failure to describe the basic techniques used.

A review of the influence of periodontal treatment in systemic diseases

The effects and consequences of periodontal diseases might not be confined to the oral cavity. A great body of evidence has arisen supporting the claim demonstrating an association with several systemic conditions and diseases. With different levels of evidence, an association between periodontal disease and cardiovascular disease, diabetes, psoriasis, rheumatoid arthritis, pregnancy outcomes and respiratory diseases has been established. However, the true nature of this association, if it is causal, still remains elusive. For a better understanding of the complex relationships linking different conditions, interventional studies now begin to focus on the possible outcomes of periodontal treatment in relation to the events, symptoms and biomarkers of several systemic disorders, assessing if periodontal treatment has any impact on them, hopefully reducing their severity or prevalence. Therefore, we proceeded to review the recent literature on the subject, attempting to present a brief explanation of the systemic condition or disease, what proposed mechanisms might give biological plausibility to its association with periodontal disease, and finally and more importantly, what data are currently available pertaining to the effects periodontal treatment may have. Raising awareness and discussing the possible benefits of periodontal treatment on overall systemic health is important, in order to change the perception that periodontal diseases are only limited to the oral cavity, and ultimately providing better and comprehensive care to patients.

Swedish Quality Registry for Caries and Periodontal Diseases - a framework for quality development in dentistry

ObjectivesL There is a need for monitoring dental health and healthcare, as support for quality development, allocation of resources and long-term planning of dental care. The aim of this paper is to describe the concept and implementation of the Swedish Quality Registry for Caries and Periodontal Diseases (SKaPa). Materials and methods: The SKaPa receives information by automatic transfer of data daily from electronic patient dental records via secure connections from affiliated dental care organisations (DCOs). The registry stores information about DCOs, dental professionals and patients. Information on a patient level includes personal identifier, gender, age, living area, dental status, risk assessments for caries and periodontitis, and dental care provided. In addition, data generated from a global question on patient-perceived oral health are uploaded. In total, more than 400 variables are transferred to the registry and updated daily. Results: In 2018, all of the 21 public DCOs and the largest private DCO in Sweden were affiliated to SKaPa, representing a total of 1,089 public and 234 private dental clinics. The accumulated amount of information on dental healthcare covers 6.9 million individuals out of the total Swedish population of 10 million. SKaPa produces reports on de-identified data, both cross-sectional and longitudinal. Conclusion: As a nationwide registry based on automatic retrieval of data directly from patient records, SKaPa offers the basis for a new era of systematic evaluation of oral health and quality of dental care. The registry supports clinical and epidemiological research, data mining and external validation of results from randomised controlled trials

What Can Google Inform Us about People's Interests regarding Dental Caries in Different Populations?

The diagnosis or suspicion of dental caries can lead people to seek additional information on the Internet through the use of structured queries in search engine tools. This action generates a considerable volume of data, which can be analyzed to provide a better understanding of the public's behavior linked to the consumption of oral health information. This study aimed to assess the volume and profile of web searches on dental caries-related queries performed by Google users from different countries. The monthly variation of the Search Volume Index (SVI) for dental caries was obtained in Google Trends for the period between January 2004 and September 2016. The validity of SVI data was assessed by their levels of stability and correlation with the disability-adjusted life-years (DALYs) for permanent teeth. In all countries, a trend of an increasing interest of Google users in dental caries issues was revealed by the comparison of the means observed in the predictive models and those in the last 12 months. The interest levels varied throughout the year, with the observation of the highest SVI values in the spring and the lowest in the summer. The most popular queries were markedly associated with symptoms and treatments, with a little interest in prevention. In conclusion, the use of Internet data mining could be helpful in establishing the dental needs of specific population groups in a near real-time, since the web consumption of dental information is increasing in importance and appears to have a direct relation with untreated dental caries.

Big Data in traumatic brain injury; promise and challenges

Traumatic brain injury (TBI) is a spectrum disease of overwhelming complexity, the research of which generates enormous amounts of structured, semi-structured and unstructured data. This resulting big data has tremendous potential to be mined for valuable information regarding the "most complex disease of the most complex organ". Big data analyses require specialized big data analytics applications, machine learning and artificial intelligence platforms to reveal associations, trends, correlations and patterns not otherwise realized by current analytical approaches. The intersection of potential data sources between experimental TBI and clinical TBI research presents inherent challenges for setting parameters for the generation of common data elements and to mine existing legacy data that would allow highly translatable big data analyses. In order to successfully utilize big data analyses in TBI, we must be willing to accept the messiness of data, collect and store all data and give up causation for correlation. In this context, coupling the big data approach to established clinical and pre-clinical data sources will transform current practices for triage, diagnosis, treatment and prognosis into highly integrated evidence-based patient care.

Medical big data: promise and challenges

The concept of big data, commonly characterized by volume, variety, velocity, and veracity, goes far beyond the data type and includes the aspects of data analysis, such as hypothesis-generating, rather than hypothesis-testing. Big data focuses on temporal stability of the association, rather than on causal relationship and underlying probability distribution assumptions are frequently not required. Medical big data as material to be analyzed has various features that are not only distinct from big data of other disciplines, but also distinct from traditional clinical epidemiology. Big data technology has many areas of application in healthcare, such as predictive modeling and clinical decision support, disease or safety surveillance, public health, and research. Big data analytics frequently exploits analytic methods developed in data mining, including classification, clustering, and regression. Medical big data analyses are complicated by many technical issues, such as missing values, curse of dimensionality, and bias control, and share the inherent limitations of observation study, namely the inability to test causality resulting from residual confounding and reverse causation. Recently, propensity score analysis and instrumental variable analysis have been introduced to overcome these limitations, and they have accomplished a great deal. Many challenges, such as the absence of evidence of practical benefits of big data, methodological issues including legal and ethical issues, and clinical integration and utility issues, must be overcome to realize the promise of medical big data as the fuel of a continuous learning healthcare system that will improve patient outcome and reduce waste in areas including nephrology.

Evidence Supports No Relationship between Obstructive Sleep Apnea and Premolar Extraction: An Electronic Health Records Review

OBJECTIVE

A controversy exists concerning the relationship, if any, between obstructive sleep apnea (OSA) and the anatomical position of the anterior teeth. Specifically, there has been speculation that extraction orthodontics and retraction of the anterior teeth contributes to OSA by crowding the tongue and decreasing airway space. This retrospective study utilized electronic medical and dental health records to examine the association between missing premolars and OSA.

METHODS

The sample (n = 5,584) was obtained from the electronic medical and dental health records of HealthPartners in Minnesota. Half of the subjects (n = 2,792) had one missing premolar in each quadrant. The other half had no missing premolars. Cases and controls were paired in a 1:1 match on age range, gender, and body mass index (BMI) range. The outcome was the presence or absence of a diagnosis of OSA confirmed by polysomnography.

RESULTS

Of the subjects without missing premolars, 267 (9.56%) had received a diagnosis of OSA. Of the subjects with four missing premolars, 299 (10.71%) had received a diagnosis of OSA. The prevalence of OSA was not significantly different between the groups (OR = 1.14, p = 0.144).

CONCLUSION

The absence of four premolars (one from each quadrant), and therefore a presumed indicator of past "extraction orthodontic treatment," is not supported as a significant factor in the cause of OSA.