Accuracy of pain recall in children.

BACKGROUND AND AIM: Chronic abdominal pain (AP) is common in children. Recall of symptoms is used clinically to determine management, to assess treatment progress, and in drug studies to assess outcomes. Limited data exist on accuracy of AP recall in children. The aim of the present study was to assess ability to accurately recall AP in children. METHODS: The study was a secondary analysis of data obtained from a double-blind, randomized, placebo-controlled trial, evaluating amitriptyline in children with functional gastrointestinal disorders. Children ages 8 to 17 years with AP predominant functional gastrointestinal disorders based on Rome II criteria were recruited from 6 centers. Those with evidence of organic disease were excluded. Patients maintained AP diary daily for 1 month (presence, frequency, and intensity). At the end of the study, patients reported the number of days of AP during previous month. Agreement between daily pain reports and recalled pain was assessed. Univariate analysis was conducted with Spearman rank correlations. RESULTS: We recruited 63 children (45 girls, mean age 12.8 years). Sixteen percent children had perfect agreement on number of days of AP. Fifty-four percent of children recalled fewer episodes of pain. The average number of days with AP by recall was 17.7/month, whereas by diary it was 23.5/month (P = 0.001). Correlation between patient recall of the last week of symptoms (r = 0.47) was no better than correlation between recall of the last 30 days of symptoms (r = 0.48). On comparing AP recall versus various pain intensities, reported AP did not reflect only AP of greater severity. Higher correlation of recall of symptoms was seen in children 11 years or younger (r = 0.59) as compared with children older than 11 years (r = 0.26). CONCLUSIONS: Few children can accurately recall the episodes of AP. Children commonly recall a lower frequency of AP than that assessed by prospective diary reports. Reported recall does not reflect a shorter recollection period. Recall is not related to intensity of pain. Adolescents have worse recall of symptoms.

Curating Cyberbullying Datasets: a Human-AI Collaborative Approach.

Cyberbullying is the use of digital communication tools and spaces to inflict physical, mental, or emotional distress. This serious form of aggression is frequently targeted at, but not limited to, vulnerable populations. A common problem when creating machine learning models to identify cyberbullying is the availability of accurately annotated, reliable, relevant, and diverse datasets. Datasets intended to train models for cyberbullying detection are typically annotated by human participants, which can introduce the following issues: (1) annotator bias, (2) incorrect annotation due to language and cultural barriers, and (3) the inherent subjectivity of the task can naturally create multiple valid labels for a given comment. The result can be a potentially inadequate dataset with one or more of these overlapping issues. We propose two machine learning approaches to identify and filter unambiguous comments in a cyberbullying dataset of roughly 19,000 comments collected from YouTube that was initially annotated using Amazon Mechanical Turk (AMT). Using consensus filtering methods, comments were classified as unambiguous when an agreement occurred between the AMT workers' majority label and the unanimous algorithmic filtering label. Comments identified as unambiguous were extracted and used to curate new datasets. We then used an artificial neural network to test for performance on these datasets. Compared to the original dataset, the classifier exhibits a large improvement in performance on modified versions of the dataset and can yield insight into the type of data that is consistently classified as bullying or non-bullying. This annotation approach can be expanded from cyberbullying datasets onto any classification corpus that has a similar complexity in scope.

How reliable are the Rome III criteria for the assessment of functional gastrointestinal disorders in children?

OBJECTIVES: Functional gastrointestinal disorders (FGIDs) are common in children. Diagnosis of these conditions is based on the pediatric Rome criteria. In the past, we have shown that there was low inter-rater reliability (IRR) among pediatric gastroenterologists using the Rome II criteria. Since then, a new version of the criteria has been issued. The reliability of the Rome III criteria has not been established. METHODS: A total of 10 pediatric gastroenterologist specialists and 10 pediatric gastroenterology fellows were provided with 20 clinical vignettes and a list of 17 possible diagnoses (all pediatric categories of the Rome criteria plus "none of the above" or "not enough information") and instructed to select one or more diagnosis for each vignette. RESULTS: The average percentage of agreement among the raters was 50% for the pediatric gastroenterologists and 45% for the pediatric gastroenterology fellows. The inter-rater percentage of agreement per clinical case was >50% in only 7 out of 20 (35%) vignettes for the gastroenterologists and only 6 out of 20 (30%) cases for the fellows. The inter-rater percentage of agreement was <25% in 2 out of 20 (10%) vignettes for the gastroenterologists and 4 out of 20 (20%) vignettes for the fellows. The κ coefficient was 0.45 for the specialists (P<0.0001) and 0.39 for the fellows (P<0.0001). In a subanalysis of the groups of pain and constipation-related disorders, the inter-rater percentage of agreement per clinical case ranged between 27 and 100% (mean 57%, κ=0.37, P<0.0001) for the gastroenterologists and between 36 and 80% (mean 52%, κ=0.33, P<0.0001) for the fellows in the constipation subgroup. The inter-rater percentage of agreement per clinical case for the pain subgroup ranged between 22 and 80% (mean 48%, κ=0.36, P<0.0001) for the gastroenterologists and 22 and 62% (mean 39%, κ=0.29, P<0.0001) for the fellows in the pain subgroup. The κ coefficient for specialists with expertise in FGIDs was 0.37 (P<0.0001) and for those with expertise in other gastroenterology conditions was 0.53 (P<0.0001). CONCLUSIONS: The IRR among pediatric gastroenterologists and fellows was found to be fair to moderate for the Rome III criteria. Only slight to fair agreement between raters existed for important subcategories of pain and constipation. The results from our current study are almost similar to that of the IRR study done for the Rome II criteria. This indicates the need for further refinement of the Rome criteria to make them more encompassing and user friendly.

A prospective school-based study of abdominal pain and other common somatic complaints in children.

OBJECTIVE: To determine the prevalence and impact of pediatric abdominal pain (AP). STUDY DESIGN: Prospective cohort study (12/2005-06/2006), with gastrointestinal and other symptoms assessed weekly. Anxiety, depression, functional disability, quality of life, somatization, coping, school absenteeism and medical care were assessed in 237 students in the third through eighth grades (11.8 years; 134 girls) from 2 public schools weekly. Complete data were obtained on 4606 of 5175 (89%) possible questionnaires. RESULTS: Seventy-two percent of children reported >1 somatic symptom weekly, and 45% of children reported >1 gastrointestinal symptom weekly. The weekly prevalence of AP was 38%, and 90% of children reported AP at least once. AP persisted >4 consecutive weeks in 52% of children and was associated with higher anxiety (P < .001) and depression (P < .001) scores and worse quality of life (P < .001). Twenty-three percent of children missed school for AP (average, 2.3 days), and 10% of parents of those children missed work (average, 1.9 days). Presence of AP (P < .001) was independently associated with school absences. Four children (2%) sought medical attention. CONCLUSIONS: AP is common in school-age children and is associated with worse quality of life, psychological co-morbidities, school absenteeism, and parental work absences.

A prospective community-based study of gastroenterological symptoms in school-age children.

INTRODUCTION: Current knowledge on the prevalence of common gastrointestinal (GI) problems is based on office-based samples or retrospective questionnaires, leading to possible bias and inaccurate estimates. AIMS: To examine the prevalence of GI symptoms in school-age children. IMPORTANCE: This is the first American prospective community study intended to assess the prevalence of common GI symptoms in children. PATIENTS AND METHODS: All fourth- and fifth-grade students from a middle-size urban school were invited to participate in a prospective cohort pilot study by completing confidential weekly surveys. The survey included a set of 8 age-appropriate, validated questions assessing the presence and severity of abdominal pain, constipation, diarrhea, nausea, vomiting, chest pain, headaches and limb pain. RESULTS: A total of 48 children (32 boys and 16 girls) participated in the study. Data were obtained for 16 weeks on 690 (90.5%) of 768 possible children per week. Children reported at least 1 symptom in 544 (70%) children per week. Headaches were the most common complaint (55% [range, 40%-72%] children per week). Sixty percent of children (range, 46%-89%) reported at least 1 GI symptom weekly. The overall prevalence of GI symptoms was the following: abdominal pain 46% (range, 28%-72%), nausea 28% (range, 17%-59%), constipation 18% (range, 7%-39%), diarrhea 17% (range, 11%-24%) and vomiting 5% (range, 0%-13%). None of the children missed school because of persistent GI symptoms during the study. DISCUSSION: The investigation demonstrates the feasibility of prospective school studies in children. The high prevalence of GI symptoms not interfering with school attendance in children underscores the benign nature of such symptoms in most children. CONCLUSION: Gastrointestinal symptoms are common somatic complaints among school-age children.