Dataset Information

Counterintuitive results from observational data: a case study and discussion.

ABSTRACT: OBJECTIVE:To explore the issue of counterintuitive data via analysis of a representative case in which the data obtained was unexpected and inconsistent with current knowledge. We then discuss the issue of counterintuitive data while developing a framework for approaching such findings. DESIGN:The case study is a retrospective analysis of a cohort of coronary artery bypass graft (CABG) patients. Regression was used to examine the association between perceived pain in the intensive care unit (ICU) and selected outcomes. SETTING:Medical Information Mart for Intensive Care-III, a publicly available, de-identified critical care patient database. PARTICIPANTS:844 adult patients from the database who underwent CABG surgery and were extubated within 24?hours after ICU admission. OUTCOMES:30?day mortality, 1?year mortality and hospital length of stay (LOS). RESULTS:Increased pain levels were found to be significantly associated with reduced mortality at 30 days and 1?year, and shorter hospital LOS. A one-point increase in mean pain level was found to be associated with a reduction in the odds of 30?day and 1?year mortality by a factor of 0.457 (95% CI 0.304 to 0.687, p<0.01) and 0.710 (95% CI 0.571 to 0.881, p<0.01) respectively, and a 0.916 (95% CI -1.159 to -0.673, p<0.01) day decrease in hospital LOS. CONCLUSION:The finding of an association between increased pain and improved outcomes was unexpected and clinically counterintuitive. In an increasingly digitised age of medical big data, such results are likely to become more common. The reliability of such counterintuitive results must be carefully examined. We suggest several issues to consider in this analytic process. If the data is determined to be valid, consideration must then be made towards alternative explanations for the counterintuitive results observed. Such results may in fact indicate that current clinical knowledge is incomplete or not have been firmly based on empirical evidence and function to inspire further research into the factors involved.

SUBMITTER: Doty E

PROVIDER: S-EPMC6502001 | biostudies-literature | 2019 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Counterintuitive results from observational data: a case study and discussion.

Doty Erik E Stone David J DJ McCague Ned N Celi Leo Anthony LA

BMJ open 20190505 5

<h4>Objective</h4>To explore the issue of counterintuitive data via analysis of a representative case in which the data obtained was unexpected and inconsistent with current knowledge. We then discuss the issue of counterintuitive data while developing a framework for approaching such findings.<h4>Design</h4>The case study is a retrospective analysis of a cohort of coronary artery bypass graft (CABG) patients. Regression was used to examine the association between perceived pain in the intensive ...[more]

PMID: 31061037

Similar Datasets

Project description:BackgroundWith the spread of COVID-19, significant concerns have been raised about the potential increased risk for electronic cigarette (e-cigarette) users for COVID-19 infection and related syndromes. Social media is an increasingly popular source for health information dissemination and discussion, and can affect health outcomes.ObjectiveThis study aims to identify the topics in the public vaping discussion in COVID-19-related Twitter posts in order to get insight into public vaping-related perceptions, attitudes and concerns, and to discern possible misinformation and misconceptions around vaping in the COVID-19 pandemic.MethodsUsing the tweets ID database maintained by Georgia State University's Panacea Lab, we downloaded the tweets related to COVID-19 from March 11, 2020, when the World Health Organization declared COVID-19 a pandemic, to February 12, 2021. We used R to analyze the tweets that contained a list of 79 keywords related to vaping. After removing duplicates and tweets created by faked accounts or bots, the final data set consisted of 11,337 unique tweets from 7,710 different users. We performed the latent Dirichlet allocation (LDA) algorithm for topic modeling and carried out a sentiment analysis.ResultsDespite fluctuations, the number of daily tweets was relatively stable (average number of daily tweets = 33.4) with a sole conspicuous spike happening on a few days after August 11, 2020 when a research team published findings that teenagers and young adults who vape face a much higher risk of COVID-19 infection than their peers who do not vape. Topic modeling generated 8 topics: linkage between vaping and risk of COVID-19 infection, vaping pneumonia and the origin of COVID-19, vaping and spread of COVID-19, vaping regulation, calling for quitting vaping, protecting youth, similarity between e-cigarette or vaping-associated lung injury (EVALI) and COVID-19, and sales information. Daily sentiment scores showed that the public sentiment was predominantly negative, but became slightly more positive over the course of the study time period.ConclusionsWhile some content in the public discourse on vaping before the COVID-19 pandemic continued in Twitter posts during the COVID-19 time period, new topics emerged. We found a substantial amount of anti-vaping discussion and dominantly negative sentiment around vaping during COVID-19, a sharp contrast to the predominantly pro-vaping voice on social media in the pre-COVID-19 period. Continued monitoring of social media conversations around vaping is needed, and the public health community may consider using social media platforms to actively convey scientific information around vaping and vaping cessation.

Project description:BackgroundTwitter is increasingly used to estimate disease prevalence, but such measurements can be biased, due to both biased sampling and inherent ambiguity of natural language.ObjectiveWe characterized the extent of these biases and how they vary with disease.MethodsWe correlated self-reported prevalence rates for 22 diseases from Experian's Simmons National Consumer Study (n=12,305) with the number of times these diseases were mentioned on Twitter during the same period (2012). We also identified and corrected for two types of bias present in Twitter data: (1) demographic variance between US Twitter users and the general US population; and (2) natural language ambiguity, which creates the possibility that mention of a disease name may not actually refer to the disease (eg, "heart attack" on Twitter often does not refer to myocardial infarction). We measured the correlation between disease prevalence and Twitter disease mentions both with and without bias correction. This allowed us to quantify each disease's overrepresentation or underrepresentation on Twitter, relative to its prevalence.ResultsOur sample included 80,680,449 tweets. Adjusting disease prevalence to correct for Twitter demographics more than doubles the correlation between Twitter disease mentions and disease prevalence in the general population (from .113 to .258, P <.001). In addition, diseases varied widely in how often mentions of their names on Twitter actually referred to the diseases, from 14.89% (3827/25,704) of instances (for stroke) to 99.92% (5044/5048) of instances (for arthritis). Applying ambiguity correction to our Twitter corpus achieves a correlation between disease mentions and prevalence of .208 ( P <.001). Simultaneously applying correction for both demographics and ambiguity more than triples the baseline correlation to .366 ( P <.001). Compared with prevalence rates, cancer appeared most overrepresented in Twitter, whereas high cholesterol appeared most underrepresented.ConclusionsTwitter is a potentially useful tool to measure public interest in and concerns about different diseases, but when comparing diseases, improvements can be made by adjusting for population demographics and word ambiguity.

Project description:Rationale: "Target trial emulation" has been proposed as an observational method to answer comparative effectiveness questions, but it has rarely been attempted concurrently with a randomized clinical trial (RCT).Objectives: We tested the hypothesis that blinded analysts applying target trial emulation to existing observational data could predict the results of an RCT.Methods: PreVent (Preventing Hypoxemia with Manual Ventilation during Endotracheal Intubation) was a multicenter RCT examining the effects of positive-pressure ventilation during tracheal intubation on oxygen saturation and severe hypoxemia. Analysts unaware of PreVent's results used patient-level data from three previous trials evaluating airway management interventions to emulate PreVent's eligibility criteria, randomization procedure, and statistical analysis. After PreVent's release, results of this blinded observational analysis were compared with those of the RCT. Difference-in-differences estimates for comparison of treatment effects between the observational analysis and the PreVent trial are reported on the absolute scale.Results: Using observational data, we were able to emulate PreVent's randomization procedure to produce balanced groups for comparison. The lowest oxygen saturation during intubation was higher in the positive-pressure ventilation group than the no positive-pressure ventilation group in the observational analysis (n = 360; mean difference = 1.8%; 95% confidence interval [CI] = -1.0 to 4.6) and in the PreVent trial (n = 401; mean difference = 3.9%; 95% CI = 1.4 to 6.4), though the observational analysis could not exclude no difference. Difference-in-differences estimates comparing treatment effects showed reasonable agreement for lowest oxygen saturation between the observational analysis and the PreVent trial (mean difference = -2.1%; 95% CI = -5.9 to 1.7). Positive-pressure ventilation resulted in lower rates of severe hypoxemia in both the observational analysis (risk ratio = 0.60; 95% CI = 0.38 to 0.93) and in the PreVent trial (risk ratio = 0.48; 95% CI = 0.30 to 0.77). The absolute reduction in the incidence of severe hypoxemia with positive-pressure ventilation was similar in the observational analysis (9.4%) and the PreVent trial (12.0%), though the difference between these estimates had wide CIs (mean difference = 2.5%; 95% CI = -8.0 to 13.6%).Conclusions: Applying target trial emulation methods to existing observational data for the evaluation of a novel intervention produced results similar to those of a randomized trial. These findings support the use of target trial emulation for comparative effectiveness research.

Project description:BACKGROUND:During health emergencies, the people within affected communities ask many questions at a time when professional medics and health agencies are overstretched and struggling to cope. Our previous research has shown that, during the 2014-2015 West Africa Ebola crisis, volunteer-moderated online discussion forums were able to provide peer-to-peer reliable, trustworthy, and well-managed information. We speculated that with the right mix of epistemic and experiential knowledge, such a discussion forum could be set up rapidly during a future serious disease outbreak. OBJECTIVE:The aim of this study was to set up a peer-to-peer health information exchange forum within the shortest time possible after the emergence of a real outbreak of a serious infectious disease. An outbreak of Nipah virus in Kerala, India, in May 2018 provided the opportunity to test our theories. METHODS:We initiated a Nipah virus discussion forum on the platform Reddit, recruiting volunteer moderators from within the existing Reddit community. This facilitated posts and comments to the forum from genuine Reddit users. We gathered and analyzed data on the number of posts, comments, page views, and subscribers during the period of May 24 to June 23, 2018, by using the data analysis tools embedded in the Reddit platform. RESULTS:We were able to set up a functioning health information exchange platform by May 24, 2018, within two weeks of the index case and one week of the official World Health Organization verification of a Nipah virus outbreak. Over the following five weeks, the forum received a steady flow of traffic including posts (36) and comments (21) submitted, page views (840), and subscribers (33). On the busiest day, 368 page views were recorded. The forum provided information in the languages spoken in the outbreak region as well as in English on how the virus spreads, symptoms of the disease, and how to take measures to avoid contracting it. Information on government helpline numbers and frequently asked questions was also provided to the community at risk. CONCLUSIONS:The delivery of a fully functional discussion forum within a short space of time during an actual health emergency demonstrates that our suggestion is fully practical. Our theory that Reddit could provide a suitable platform to host such a forum was upheld. This offers great potential for public health communication during future serious disease outbreaks.

Project description:BackgroundPerioperative FLOT (5-fluorouracil, oxaliplatin and docetaxel) has recently become the gold standard treatment for fit patients with operable gastric (GC) or gastroesophageal (GEJ) adenocarcinoma, getting a 5-year overall survival (OS) of 45%, over 23% with surgery alone.MethodsRealFLOT is an Italian, multicentric, observational trial, collecting data from patients with resectable GC or GEJ adenocarcinoma treated with perioperative FLOT. Aim of the study was to describe feasibility and safety of FLOT, pathological complete response rate (pCR), surgical outcomes and overall response rate (ORR) in an unselected real-world population. Additional analyses evaluated the correlation between pCR and survival and the prognostic role of microsatellite instability (MSI) status.ResultsOf 206 patients enrolled that received perioperative FLOT at 15 Italian centers, 124 (60.2%) received at least 4 full-dose cycles, 190 (92.2%) underwent surgery, and 142 (68.9%) started the postoperative phase. Among patients who started the postoperative phase, 105 (51.0%) received FLOT, while 37 (18%) received de-intensified regimens, depending on clinical condition or previous toxicities. pCR was achieved in 7.3% of cases. Safety profile was consistent with literature. Neutropenia was the most common G 3-4 adverse event (AE): 19.9% in the preoperative phase and 16.9% in the postoperative phase. No toxic death was observed and 30-day postoperative mortality rate was 1.0%. ORR was 45.6% and disease control rate (DCR) was 94.2%. Disease-free survival (DFS) and OS were significantly longer in case of pCR (p = 0.009 and p = 0.023, respectively). A trend towards better DFS was observed among MSI-H patients.ConclusionsThese real-world data confirm the feasibility of FLOT in an unselected population, representative of the clinical practice. pCR rate was lower than expected, nevertheless we confirm pCR as a predictive parameter of survival. In addition, MSI-H status seems to be a positive prognostic marker also in patients treated with taxane-containing triplets.

Project description:Minimally invasive lumbar interbody fusion (MILIF) offers potential for reduced operative morbidity and earlier recovery compared with open procedures for patients with degenerative lumbar disorders (DLD). Firm conclusions about advantages of MILIF over open procedures cannot be made because of limited number of large studies of MILIF in a real-world setting. Clinical effectiveness of MILIF in a large, unselected real-world patient population was assessed in this Prospective, monitored, international, multicenter, observational study.To observe and document short-term recovery after minimally invasive interbody fusion for DLD.In a predefined 4-week analysis from this study, experienced surgeons (? 30 MILIF surgeries pre-study) treated patients with DLD by one- or two-level MILIF. The primary study objective was to document patients' short-term post-interventional recovery (primary objective) including back/leg pain (visual analog scale [VAS]), disability (Oswestry Disability Index [ODI]), health status (EQ-5D) and Patient satisfaction.At 4 weeks, 249 of 252 patients were remaining in the study; the majority received one-level MILIF (83%) and TLIF was the preferred approach (94.8%). For one-level (and two-level) procedures, surgery duration was 128 (182) min, fluoroscopy time 115 (154) sec, and blood-loss 164 (233) mL. Time to first ambulation was 1.3 days and time to study-defined surgery recovery was 3.2 days. Patients reported significantly (P < 0.0001) reduced back pain (VAS: 2.9 vs 6.2), leg pain (VAS: 2.5 vs 5.9), and disability (ODI: 34.5% vs 45.5%), and a significantly (P < 0.0001) improved health status (EQ-5D index: 0.61 vs 0.34; EQ VAS: 65.4 vs 52.9) 4 weeks postoperatively. One adverse event was classified as related to the minimally invasive surgical approach. No deep site infections or deaths were reported.For experienced surgeons, MILIF for DLD demonstrated early benefits (short time to first ambulation, early recovery, high patient satisfaction and improved patient-reported outcomes) and low major perioperative morbidity at 4 weeks postoperatively.

Dataset Information

Counterintuitive results from observational data: a case study and discussion.

Publications

Counterintuitive results from observational data: a case study and discussion.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets