Session 2: Thesauri and Algorithms

Friday, 19 June, 2015 - 09:45 to 11:00
Conference room: 

An ontology to support evidence-based Functional Behavioral Assessments

The term Functional Behavioral Assessment (FBA) refers to a set of multiple assessment strategies aimed to identify specific antecedents and consequences that maintain a target behaviour (Du Paul, 1996). In the context of FBA, standards and protocols to support the data exchange between researchers, health professionals and therapist are under-represented. Moreover, there is a need for sharing common actions and assessment practices in order to improve the application of the FBA. To meet these goals, we propose the definition of an FBA ontology as a tool to describe three specific aspects related to individual, behavioural and assessment data in different contexts of everyday life such as school, family and social environments. The individual data include: diagnoses, medications, school information, discipline referrals and other events, thus providing a comprehensive overview of the individuals and the network of support people with specific roles that collaborates on the individual’s care. The behavioural data include the description of an individual´s target behaviour, and information about places and settings in which the behaviour occurs. The assessment gathers structural behavioural data collections, according to systematic direct observation, and in compliance to the ABC model (antecedents, behavior, and consequences) commonly used in the Cognitive Behavioral Therapy (CBT) for identifying behavioural functions, and designing intervention plans. This data is relevant in supporting statistical analysis in order to evaluate the efficacy of the behavioural treatments. Moreover, the FBA ontology provides a complete model that enables integration and interlinking with other Linked Open Data datasets and repositories thus supporting the sharing of appropriate resources such as behavioural patterns, effective intervention strategies, and behavioural treatments. Finally, the ontology provides the basis for the designing of software applications to support the functional assessment processes. This ontology has been applied in the framework of the WHAAM (Web Health Application for ADHD Monitoring) project, aimed to promote the FBA approach to the behavioural treatment of ADHD children. References: DuPaul, G. J., & Ervin, R. A. (1996). Functional assessment of behaviors related to attention-deficit/hyperactivity disorder: Linking assessment to intervention design. Behavior Therapy, 27(4)

A Mobile-Friendly Vocabulary for ADHD

Attention Deficit Hyperactivity Disorder (ADHD) is one of the most common disorders that affect children. There is a debate among scientific community about whether ADHD constitutes a risk factor with consequences in adult life or it is a distinguished syndrome. Given that there is not an ADHD acceptable biological marker, its diagnosis is based on behavioral symptoms. This study aims at developing a mobile-friendly vocabulary for ADHD that combines DSM-IV, ICD-10 and Vanderbilt ADHD Diagnostic Teacher Rating Scale in the most sufficient way. More precisely, we propose a list of ADHD symptoms that are the minimum number of symptoms that should be diagnosed on an ADHD sufferer. The proposed vocabulary could promote the early diagnosis from parents, teachers as well as health professionals but also stimulate the public awareness on ADHD.

A generic ABCDE algorithm for medical image interpretation

An approach to the process of interpreting medical images is presented. The algorithm applies for both Pathology and Radiology images. This simple mnemonic (ABCDE) can be used as a memory aid determining the order in which diagnosis should be approached. First, before having the slide under the microscope or the radiology image on diaphanoscope we need to make sure that it is correctly labelled and prepared (Correct). We also need to know or gather all available information concerning the patient (A, Available Information). Once we come across the image, two fundamental questions should be answered: which part of the body does the image concern and -where it applies- if the image is adequate (B, Body & Being Adequate). Next, we proceed to answer if there is neoplasia or not (C, Cancer). We then either form a differential diagnosis list or we end up with our final diagnosis (D, Diagnosis), which is followed by the writing of the report (E, Exhibit). These sequential steps (Correct ABCDE) followed as an ad hoc procedure by most pathologists and radiologists, are important in order to achieve a complete and clear diagnosis and report, which is intended to support optimal clinical practice. This ABCDE concept is a generic standard approach which is not limited to specific specimens and can help improve both diagnoses and the quality of the final reports.

Finding the best web medical content for different learner categories

In the age of Internet where any kind of information can be easily found online, it is becoming increasingly evident that more and more people use the World Wide Web to seek health and medical information for understanding and learning. Different users have diverse needs, even when searching for the same topic. This is certainly true in healthcare, where a patient, a physician or a health executive might look for information on the same topic but have different necessities and bring different levels of reading ability and prior knowledge together with a different vocabulary. Generic search engines (like Google, Bing or Yahoo) work on the whole web but make generic searches often overloading the user with the provided amount of information. Moreover, they are not able to provide specific information to different types of users. On the other hand, specific search engines, such as PubMed or Quertle, work only on medical literature (mostly PubMed). They provide extracts from medical journals that are mainly useful for medical researchers and experts but do not consider all the information contained in the web that can often provide additional insights to the specific research domain being explored. Internet users looking for medical information would greatly benefit from a search engine that provides them with the ‘right’ information they are looking for without getting ‘lost’ with the amount and quality of information that Internet provides. To this end, we have developed a web search engine that ‘drives’ the search path of different learner categories by classifying the web pages on the basis of their level of health information and used language. In particular, we divide the web pages retrieved by a generic search engine in four categories: high level of medical content for experts, high level of medical content for consumers, low level of medical content and no medical content. We have implemented this system and carried out some experiments that show the effectiveness of our system when compared with the results carried out through human analysis.
Single session