Search

The sample size of a randomized controlled trial is typically chosen in order for frequentist operational characteristics to be retained. For normally distributed outcomes, an assumption for the variance needs to be made which is usually based on limited prior information. Especially in the case of small populations, the prior information might consist of only one small pilot study. A Bayesian approach formalizes the aggregation of prior information on the variance with newly collected data. The uncertainty surrounding prior estimates can be appropriately modelled by means of prior distributions. Furthermore, within the Bayesian paradigm, quantities such as the probability of a conclusive trial are directly calculated. However, if the postulated prior is not in accordance with the true variance, such calculations are not trustworthy. In this work we adapt previously suggested methodology to facilitate sample size re-estimation. In addition, we suggest the employment of power priors in order for operational characteristics to be controlled. [ABSTRACT FROM AUTHOR]

Background: Rapid development and implementation of vaccines constituted a crucial step in containing the COVID-19 pandemic. A comprehensive understanding of physiological responses to these vaccines is important to build trust in medicine., Objective: This study aims to investigate temporal dynamics before and after COVID-19 vaccination in 4 physiological parameters as well as the duration of menstrual cycle phases., Methods: In a prospective trial, 17,825 adults in the Netherlands wore a medical device on their wrist for up to 9 months. The device recorded their physiological signals and synchronized with a complementary smartphone app. By means of multilevel quadratic regression, we examined changes in wearable-recorded breathing rate, wrist skin temperature, heart rate, heart rate variability, and objectively assessed the duration of menstrual cycle phases in menstruating participants to assess the effects of COVID-19 vaccination., Results: The recorded physiological signals demonstrated short-term increases in breathing rate and heart rate after COVID-19 vaccination followed by a prompt rebound to baseline levels likely reflecting biological mechanisms accompanying the immune response to vaccination. No sex differences were evident in the measured physiological responses. In menstruating participants, we found a 0.8% decrease in the duration of the menstrual phase following vaccination., Conclusions: The observed short-term changes suggest that COVID-19 vaccines are not associated with long-term biophysical issues. Taken together, our work provides valuable insights into continuous fluctuations of physiological responses to vaccination and highlights the importance of digital solutions in health care., International Registered Report Identifier (irrid): RR2-10.1186/s13063-021-05241-5., (©Andjela Markovic, Vladimir Kovacevic, Timo B Brakenhoff, Duco Veen, Paul Klaver, Marianna Mitratza, George S Downward, Diederick E Grobbee, Maureen Cronin, Brianna M Goodale, COVID-19 Remote Early Detection (COVID-RED) consortium. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 31.07.2024.)

Considering sex as a biological variable in modern digital health solutions, we investigated sex-specific differences in the trajectory of four physiological parameters across a COVID-19 infection. A wearable medical device measured breathing rate, heart rate, heart rate variability, and wrist skin temperature in 1163 participants (mean age = 44.1 years, standard deviation [SD] = 5.6; 667 [57%] females). Participants reported daily symptoms and confounders in a complementary app. A machine learning algorithm retrospectively ingested daily biophysical parameters to detect COVID-19 infections. COVID-19 serology samples were collected from all participants at baseline and follow-up. We analysed potential sex-specific differences in physiology and antibody titres using multilevel modelling and t-tests. Over 1.5 million hours of physiological data were recorded. During the symptomatic period of infection, men demonstrated larger increases in skin temperature, breathing rate, and heart rate as well as larger decreases in heart rate variability than women. The COVID-19 infection detection algorithm performed similarly well for men and women. Our study belongs to the first research to provide evidence for differential physiological responses to COVID-19 between females and males, highlighting the potential of wearable technology to inform future precision medicine approaches., Competing Interests: The authors have read the journal’s policy and have the following competing interests: Lorenz Risch, and Martin Risch are key shareholders of the Dr Risch Medical Laboratory. David Conen has received consulting fees from Roche Diagnostics, outside of the current work. Andjela Markovic, Vladimir Kovacevic, Martina Rothenbühler, Brianna Goodale and Maureen Cronin are past employees of Ava AG. Brianna Goodale and Timo Brakenhoff are current employees of Julius Clinical BV. Billy Franks is a former employee of Julius Clinical BV and now an employee of Haleon. Paul Klaver and Duco Veen are former employees of Julius Clinical BV. Marianna Mitratza is a current employee of P95 CVBA. There are no patents, products in development or marketed products associated with this research to declare. These competing interests do not alter our adherence to PLOS ONE policies on sharing data and materials., (Copyright: © 2024 Grossmann et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Background With the increase of highly portable, wireless, and low-cost ultrasound devices and automatic ultrasound acquisition techniques, an automated interpretation method requiring only a limited set of views as input could make preliminary cardiovascular disease diagnoses more accessible. In this study, we developed a deep learning method for automated detection of impaired left ventricular (LV) function and aortic valve (AV) regurgitation from apical 4-chamber ultrasound cineloops and investigated which anatomical structures or temporal frames provided the most relevant information for the deep learning model to enable disease classification. Methods and Results Apical 4-chamber ultrasounds were extracted from 3554 echocardiograms of patients with impaired LV function (n=928), AV regurgitation (n=738), or no significant abnormalities (n=1888). Two convolutional neural networks were trained separately to classify the respective disease cases against normal cases. The overall classification accuracy of the impaired LV function detection model was 86%, and that of the AV regurgitation detection model was 83%. Feature importance analyses demonstrated that the LV myocardium and mitral valve were important for detecting impaired LV function, whereas the tip of the mitral valve anterior leaflet, during opening, was considered important for detecting AV regurgitation. Conclusions The proposed method demonstrated the feasibility of a 3-dimensional convolutional neural network approach in detection of impaired LV function and AV regurgitation using apical 4-chamber ultrasound cineloops. The current study shows that deep learning methods can exploit large training data to detect diseases in a different way than conventionally agreed on methods, and potentially reveal unforeseen diagnostic image features.

Objectives: We investigated machinelearningbased identification of presymptomatic COVID-19 and detection of infection-related changes in physiology using a wearable device., Design: Interim analysis of a prospective cohort study., Setting, Participants and Interventions: Participants from a national cohort study in Liechtenstein were included. Nightly they wore the Ava-bracelet that measured respiratory rate (RR), heart rate (HR), HR variability (HRV), wrist-skin temperature (WST) and skin perfusion. SARS-CoV-2 infection was diagnosed by molecular and/or serological assays., Results: A total of 1.5 million hours of physiological data were recorded from 1163 participants (mean age 44±5.5 years). COVID-19 was confirmed in 127 participants of which, 66 (52%) had worn their device from baseline to symptom onset (SO) and were included in this analysis. Multi-level modelling revealed significant changes in five (RR, HR, HRV, HRV ratio and WST) device-measured physiological parameters during the incubation, presymptomatic, symptomatic and recovery periods of COVID-19 compared with baseline. The training set represented an 8-day long instance extracted from day 10 to day 2 before SO. The training set consisted of 40 days measurements from 66 participants. Based on a random split, the test set included 30% of participants and 70% were selected for the training set. The developed long short-term memory (LSTM) based recurrent neural network (RNN) algorithm had a recall (sensitivity) of 0.73 in the training set and 0.68 in the testing set when detecting COVID-19 up to 2 days prior to SO., Conclusion: Wearable sensor technology can enable COVID-19 detection during the presymptomatic period. Our proposed RNN algorithm identified 68% of COVID-19 positive participants 2 days prior to SO and will be further trained and validated in a randomised, single-blinded, two-period, two-sequence crossover trial. Trial registration number ISRCTN51255782; Pre-results., Competing Interests: Competing interests: LR, and MR are key shareholders of the Dr Risch Medical Laboratory. DC has received consulting fees from Roche Diagnostics, outside of the current work. MR, DL, VK, AM, MC, and BMG are employed by Ava AG. The other authors have no financial or personal conflicts of interest to declare., (© Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

Containing the COVID-19 pandemic requires rapidly identifying infected individuals. Subtle changes in physiological parameters (such as heart rate, respiratory rate, and skin temperature), discernible by wearable devices, could act as early digital biomarkers of infections. Our primary objective was to assess the performance of statistical and algorithmic models using data from wearable devices to detect deviations compatible with a SARS-CoV-2 infection. We searched MEDLINE, Embase, Web of Science, the Cochrane Central Register of Controlled Trials (known as CENTRAL), International Clinical Trials Registry Platform, and ClinicalTrials.gov on July 27, 2021 for publications, preprints, and study protocols describing the use of wearable devices to identify a SARS-CoV-2 infection. Of 3196 records identified and screened, 12 articles and 12 study protocols were analysed. Most included articles had a moderate risk of bias, as per the National Institute of Health Quality Assessment Tool for Observational and Cross-Sectional Studies. The accuracy of algorithmic models to detect SARS-CoV-2 infection varied greatly (area under the curve 0·52-0·92). An algorithm's ability to detect presymptomatic infection varied greatly (from 20% to 88% of cases), from 14 days to 1 day before symptom onset. Increased heart rate was most frequently associated with SARS-CoV-2 infection, along with increased skin temperature and respiratory rate. All 12 protocols described prospective studies that had yet to be completed or to publish their results, including two randomised controlled trials. The evidence surrounding wearable devices in the early detection of SARS-CoV-2 infection is still in an early stage, with a limited overall number of studies identified. However, these studies show promise for the early detection of SARS-CoV-2 infection. Large prospective, and preferably controlled, studies recruiting and retaining larger and more diverse populations are needed to provide further evidence., Competing Interests: Declaration of interests MM, BMG, VK, BF, DV, DEG, MC, and GSD received grants from Innovative Medicines Initiative 2 Joint Undertaking (number 101005177), during the conduct of the study. BMG reports consulting fees and employment from Ava Science, support for attending meetings and travel from Ava Aktiengesellschaft (AG), a patent application from Ava AG (P24892CH00) filed with the Swiss Federal Institute of Intellectual Property for System and Method for Pre-Symptomatic and/or Asymptomatic Detection of a Human Viral or Bacterial Infection based on pilot data from the COVID-RED clinical study, and consultancy for Falcon Health and TheraB Medical, outside the submitted work. VK reports employment from Ava Science and Ava AG, during the conduct of this study. TBB, BF, DV, and DEG report employment from Julius Clinical Research, during the conduct of the study. MC reports employment from Ava AG during the conduct of the study. GSD reports a grant from Health Holland, outside the submitted work. All other authors declare no competing interests., (Copyright © 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

Objectives: It is currently thought that most-but not all-individuals infected with SARS-CoV-2 develop symptoms, but the infectious period starts on average 2 days before the first overt symptoms appear. It is estimated that pre- and asymptomatic individuals are responsible for more than half of all transmissions. By detecting infected individuals before they have overt symptoms, wearable devices could potentially and significantly reduce the proportion of transmissions by pre-symptomatic individuals. Using laboratory-confirmed SARS-CoV-2 infections (detected via serology tests [to determine if there are antibodies against the SARS-CoV-2 in the blood] or SARS-CoV-2 infection tests such as polymerase chain reaction [PCR] or antigen tests) as the gold standard, we will determine the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for the following two algorithms to detect first time SARS-CoV-2 infection including early or asymptomatic infection: • The algorithm using Ava bracelet data when coupled with self-reported Daily Symptom Diary data (Wearable + Symptom Data Algo; experimental condition) • The algorithm using self-reported Daily Symptom Diary data alone (Symptom Only Algo; control condition) In addition, we will determine which of the two algorithms has superior performance characteristics for detecting SARS-CoV-2 infection including early or asymptomatic infection as confirmed by SARS-CoV-2 virus testing., Trial Design: The trial is a randomized, single-blinded, two-period, two-sequence crossover trial. The study will start with an initial learning phase (maximum of 3 months), followed by period 1 (3 months) and period 2 (3 months). Subjects entering the study at the end of the recruitment period may directly start with period 1 and will not be part of the learning phase. Each subject will undergo the experimental condition (the Wearable + Symptom Data Algo) in either period 1 or period 2 and the control condition (Symptom Only Algo) in the other period. The order will be randomly assigned, resulting in subjects being allocated 1:1 to either sequence 1 (experimental condition first) or sequence 2 (control condition first). Based on demographics, medical history and/or profession, each subject will be stratified at baseline into a high-risk and normal-risk group within each sequence., Participants: The trial will be conducted in the Netherlands. A target of 20,000 subjects will be enrolled. Based on demographics, medical history and/or profession, each subject will be stratified at baseline into a high-risk and normal-risk group within each sequence. This results in approximately 6500 normal-risk individuals and 3500 high-risk individuals per sequence. Subjects will be recruited from previously studied cohorts as well as via public campaigns and social media. All data for this study will be collected remotely through the Ava COVID-RED app, the Ava bracelet, surveys in the COVID-RED web portal and self-sampling serology and PCR kits. More information on the study can be found in www.covid-red.eu . During recruitment, subjects will be invited to visit the COVID-RED web portal. After successfully completing the enrolment questionnaire, meeting eligibility criteria and indicating interest in joining the study, subjects will receive the subject information sheet and informed consent form. Subjects can enrol in COVID-RED if they comply with the following inclusion and exclusion criteria: Inclusion criteria: • Resident of the Netherlands • At least 18 years old • Informed consent provided (electronic) • Willing to adhere to the study procedures described in the protocol • Must have a smartphone that runs at least Android 8.0 or iOS 13.0 operating systems and is active for the duration of the study (in the case of a change of mobile number, the study team should be notified) • Be able to read, understand and write Dutch Exclusion criteria: • Previous positive SARS-CoV-2 test result (confirmed either through PCR/antigen or antibody tests; self-reported) • Current suspected (e.g. waiting for test result) COVID-19 infection or symptoms of a COVID-19 infection (self-reported) • Participating in any other COVID-19 clinical drug, vaccine or medical device trial (self-reported) • Electronic implanted device (such as a pacemaker; self-reported) • Pregnant at the time of informed consent (self-reported) • Suffering from cholinergic urticaria (per the Ava bracelet's user manual; self-reported) • Staff involved in the management or conduct of this study INTERVENTION AND COMPARATOR: All subjects will be instructed to complete the Daily Symptom Diary in the Ava COVID-RED app daily, wear their Ava bracelet each night and synchronize it with the app each day for the entire period of study participation. Provided with wearable sensor and/or self-reported symptom data within the last 24 h, the Ava COVID-RED app's underlying algorithms will provide subjects with a real-time indicator of their overall health and well-being. Subjects will see one of three messages, notifying them that no seeming deviations in symptoms and/or physiological parameters have been detected; some changes in symptoms and/or physiological parameters have been detected and they should self-isolate; or alerting them that deviations in their symptoms and/or physiological parameters could be suggestive of a potential COVID-19 infection and to seek additional testing. We will assess the intraperson performance of the algorithms in the experimental condition (Wearable + Symptom Data Algo) and control conditions (Symptom Only Algo). Note that both algorithms will also instruct to seek testing when any SARS-CoV-2 symptoms are reported in line with those defined by the Dutch national institute for public health and the environment 'Rijksinstituut voor Volksgezondheid en Milieu' (RIVM) guidelines., Main Outcomes: The trial will evaluate the use and performance of the Ava COVID-RED app and Ava bracelet, which uses sensors to measure breathing rate, pulse rate, skin temperature and heart rate variability for the purpose of early and asymptomatic detection and monitoring of SARS-CoV-2 in general and high-risk populations. Using laboratory-confirmed SARS-CoV-2 infections (detected via serology tests, PCR tests and/or antigen tests) as the gold standard, we will determine the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for each of the following two algorithms to detect first-time SARS-CoV-2 infection including early or asymptomatic infection: the algorithm using Ava bracelet data when coupled with the self-reported Daily Symptom Diary data and the algorithm using self-reported Daily Symptom Diary data alone. In addition, we will determine which of the two algorithms has superior performance characteristics for detecting SARS-CoV-2 infection including early or asymptomatic infection as confirmed by SARS-CoV-2 virus testing. The protocol contains an additional twenty secondary and exploratory objectives which address, among others, infection incidence rates, health resource utilization, symptoms reported by SARS-CoV-2-infected participants and the rate of breakthrough and asymptomatic SARS-CoV-2 infections among individuals vaccinated against COVID-19. PCR or antigen testing will occur when the subject receives a notification from the algorithm to seek additional testing. Subjects will be advised to get tested via the national testing programme and report the testing result in the Ava COVID-RED app and a survey. If they cannot obtain a test via the national testing programme, they will receive a nasal swab self-sampling kit at home, and the sample will be tested by PCR in a trial-affiliated laboratory. In addition, all subjects will be asked to take a capillary blood sample at home at baseline (between month 0 and 3.5 months after the start of subject recruitment), at the end of the learning phase (month 3; note that this sampling moment is skipped if a subject entered the study at the end of the recruitment period), period 1 (month 6) and period 2 (month 9). These samples will be used for SARS-CoV-2-specific antibody testing in a trial-affiliated laboratory, differentiating between antibodies resulting from a natural infection and antibodies resulting from COVID-19 vaccination (as vaccination will gradually be rolled out during the trial period). Baseline samples will only be analysed if the sample collected at the end of the learning phase is positive, or if the subject entered the study at the end of the recruitment period, and samples collected at the end of period 1 will only be analysed if the sample collected at the end of period 2 is positive. When subjects obtain a positive PCR/antigen or serology test result during the study, they will continue to be in the study but will be moved into a so-called COVID-positive mode in the Ava COVID-RED app. This means that they will no longer receive recommendations from the algorithms but can still contribute and track symptom and bracelet data. The primary analysis of the main objective will be executed using the data collected in period 2 (months 6 through 9). Within this period, serology tests (before and after period 2) and PCR/antigen tests (taken based on recommendations by the algorithms) will be used to determine if a subject was infected with SARS-CoV-2 or not. Within this same time period, it will be determined if the algorithms gave any recommendations for testing. The agreement between these quantities will be used to evaluate the performance of the algorithms and how these compare between the study conditions., Randomization: All eligible subjects will be randomized using a stratified block randomization approach with an allocation ratio of 1:1 to one of two sequences (experimental condition followed by control condition or control condition followed by experimental condition). Based on demographics, medical history and/or profession, each subject will be stratified at baseline into a high-risk and normal-risk group within each sequence, resulting in approximately equal numbers of high-risk and normal-risk individuals between the sequences., Blinding (masking): In this study, subjects will be blinded to the study condition and randomization sequence. Relevant study staff and the device manufacturer will be aware of the assigned sequence. The subject will wear the Ava bracelet and complete the Daily Symptom Diary in the Ava COVID-RED app for the full duration of the study, and they will not know if the feedback they receive about their potential infection status will only be based on the data they entered in the Daily Symptom Diary within the Ava COVID-RED app or based on both the data from the Daily Symptom Diary and the Ava bracelet., Numbers to Be Randomized (sample Size): A total of 20,000 subjects will be recruited and randomized 1:1 to either sequence 1 (experimental condition followed by control condition) or sequence 2 (control condition followed by experimental condition), taking into account their risk level. This results in approximately 6500 normal-risk and 3500 high-risk individuals per sequence., Trial Status: Protocol version: 3.0, dated May 3, 2021. Start of recruitment: February 19, 2021. End of recruitment: June 3, 2021. End of follow-up (estimated): November 2021 TRIAL REGISTRATION: The Netherlands Trial Register on the 18 th of February, 2021 with number NL9320 ( https://www.trialregister.nl/trial/9320 ) FULL PROTOCOL: The full protocol is attached as an additional file, accessible from the Trials website (Additional file 1). In the interest in expediting dissemination of this material, the familiar formatting has been eliminated; this letter serves as a summary of the key elements of the full protocol., (© 2021. The Author(s).)

Objectives: To compare methods to adjust for confounding by disease severity during multicenter intervention studies in ICU, when different disease severity measures are collected across centers., Design: In silico simulation study using national registry data., Setting: Twenty mixed ICUs in The Netherlands., Subjects: Fifty-five-thousand six-hundred fifty-five ICU admissions between January 1, 2011, and January 1, 2016., Interventions: None., Measurements and Main Results: To mimic an intervention study with confounding, a fictitious treatment variable was simulated whose effect on the outcome was confounded by Acute Physiology and Chronic Health Evaluation IV predicted mortality (a common measure for disease severity). Diverse, realistic scenarios were investigated where the availability of disease severity measures (i.e., Acute Physiology and Chronic Health Evaluation IV, Acute Physiology and Chronic Health Evaluation II, and Simplified Acute Physiology Score II scores) varied across centers. For each scenario, eight different methods to adjust for confounding were used to obtain an estimate of the (fictitious) treatment effect. These were compared in terms of relative (%) and absolute (odds ratio) bias to a reference scenario where the treatment effect was estimated following correction for the Acute Physiology and Chronic Health Evaluation IV scores from all centers. Complete neglect of differences in disease severity measures across centers resulted in bias ranging from 10.2% to 173.6% across scenarios, and no commonly used methodology-such as two-stage modeling or score standardization-was able to effectively eliminate bias. In scenarios where some of the included centers had (only) Acute Physiology and Chronic Health Evaluation II or Simplified Acute Physiology Score II available (and not Acute Physiology and Chronic Health Evaluation IV), either restriction of the analysis to Acute Physiology and Chronic Health Evaluation IV centers alone or multiple imputation of Acute Physiology and Chronic Health Evaluation IV scores resulted in the least amount of relative bias (0.0% and 5.1% for Acute Physiology and Chronic Health Evaluation II, respectively, and 0.0% and 4.6% for Simplified Acute Physiology Score II, respectively). In scenarios where some centers used Acute Physiology and Chronic Health Evaluation II, regression calibration yielded low relative bias too (relative bias, 12.4%); this was not true if these same centers only had Simplified Acute Physiology Score II available (relative bias, 54.8%)., Conclusions: When different disease severity measures are available across centers, the performance of various methods to control for confounding by disease severity may show important differences. When planning multicenter studies, researchers should make contingency plans to limit the use of or properly incorporate different disease measures across centers in the statistical analysis.

Background: When profiling health care providers, adjustment for case-mix is essential. However, conventional risk adjustment methods may perform poorly, especially when provider volumes are small or events rare. Propensity score (PS) methods, commonly used in observational studies of binary treatments, have been shown to perform well when the amount of observations and/or events are low and can be extended to a multiple provider setting. The objective of this study was to evaluate the performance of different risk adjustment methods when profiling multiple health care providers that perform highly protocolized procedures, such as coronary artery bypass grafting., Methods: In a simulation study, provider effects estimated using PS adjustment, PS weighting, PS matching, and multivariable logistic regression were compared in terms of bias, coverage and mean squared error (MSE) when varying the event rate, sample size, provider volumes, and number of providers. An empirical example from the field of cardiac surgery was used to demonstrate the different methods., Results: Overall, PS adjustment, PS weighting, and logistic regression resulted in provider effects with low amounts of bias and good coverage. The PS matching and PS weighting with trimming led to biased effects and high MSE across several scenarios. Moreover, PS matching is not practical to implement when the number of providers surpasses three., Conclusions: None of the PS methods clearly outperformed logistic regression, except when sample sizes were relatively small. Propensity score matching performed worse than the other PS methods considered., Competing Interests: Declaration of conflicting interests:The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Background: When profiling multiple health care providers, adjustment for case-mix is essential to accurately classify the quality of providers. Unfortunately, misclassification of provider performance is not uncommon and can have grave implications. Propensity score (PS) methods have been proposed as viable alternatives to conventional multivariable regression. The objective was to assess the outlier classification performance of risk adjustment methods when profiling multiple providers., Methods: In a simulation study based on empirical data, the classification performance of logistic regression (fixed and random effects), PS adjustment, and three PS weighting methods was evaluated when varying parameters such as the number of providers, the average incidence of the outcome, and the percentage of outliers. Traditional classification accuracy measures were considered, including sensitivity and specificity., Results: Fixed effects logistic regression consistently had the highest sensitivity and negative predictive value, yet a low specificity and positive predictive value. Of the random effects methods, PS adjustment and random effects logistic regression performed equally well or better than all the remaining PS methods for all classification accuracy measures across the studied scenarios., Conclusions: Of the evaluated PS methods, only PS adjustment can be considered a viable alternative to random effects logistic regression when profiling multiple providers in different scenarios.

Objectives: In medical research, covariates (e.g., exposure and confounder variables) are often measured with error. While it is well accepted that this introduces bias and imprecision in exposure-outcome relations, it is unclear to what extent such issues are currently considered in research practice. The objective was to study common practices regarding covariate measurement error via a systematic review of general medicine and epidemiology literature., Study Design and Setting: Original research published in 2016 in 12 high impact journals was full-text searched for phrases relating to measurement error. Reporting of measurement error and methods to investigate or correct for it were quantified and characterized., Results: Two hundred and forty-seven (44%) of the 565 original research publications reported on the presence of measurement error. 83% of these 247 did so with respect to the exposure and/or confounder variables. Only 18 publications (7% of 247) used methods to investigate or correct for measurement error., Conclusions: Consequently, it is difficult for readers to judge the robustness of presented results to the existence of measurement error in the majority of publications in high impact journals. Our systematic review highlights the need for increased awareness about the possible impact of covariate measurement error. Additionally, guidance on the use of measurement error correction methods is necessary., (Copyright © 2018 Elsevier Inc. All rights reserved.)

With the increased use of data not originally recorded for research, such as routine care data (or 'big data'), measurement error is bound to become an increasingly relevant problem in medical research. A common view among medical researchers on the influence of random measurement error (i.e. classical measurement error) is that its presence leads to some degree of systematic underestimation of studied exposure-outcome relations (i.e. attenuation of the effect estimate). For the common situation where the analysis involves at least one exposure and one confounder, we demonstrate that the direction of effect of random measurement error on the estimated exposure-outcome relations can be difficult to anticipate. Using three example studies on cardiovascular risk factors, we illustrate that random measurement error in the exposure and/or confounder can lead to underestimation as well as overestimation of exposure-outcome relations. We therefore advise medical researchers to refrain from making claims about the direction of effect of measurement error in their manuscripts, unless the appropriate inferential tools are used to study or alleviate the impact of measurement error from the analysis.

Background: Research on prognostic prediction models frequently uses data from routine healthcare. However, potential misclassification of predictors when using such data may strongly affect the studied associations. There is no doubt that such misclassification could lead to the derivation of suboptimal prediction models. The extent to which misclassification affects the validation of existing prediction models is currently unclear.We aimed to quantify the amount of misclassification in routine care data and its effect on the validation of the existing risk prediction model. As an illustrative example, we validated the CHA2DS2-VASc prediction rule for predicting mortality in patients with atrial fibrillation (AF)., Methods: In a prospective cohort in general practice in the Netherlands, we used computerized retrieved data from the electronic medical records of patients known with AF as index predictors. Additionally, manually collected data after scrutinizing all complete medical files were used as reference predictors. Comparing the index with the reference predictors, we assessed misclassification in individual predictors by calculating Cohen's kappas and other diagnostic test accuracy measures. Predictive performance was quantified by the c-statistic and by determining calibration of multivariable models., Results: In total, 2363 AF patients were included. After a median follow-up of 2.7 (IQR 2.3-3.0) years, 368 patients died (incidence rate 6.2 deaths per 100 person-years). Misclassification in individual predictors ranged from substantial (Cohen's kappa 0.56 for prior history of heart failure) to minor (kappa 0.90 for a history of type 2 diabetes). The overall model performance was not affected when using either index or reference predictors, with a c-statistic of 0.684 and 0.681, respectively, and similar calibration., Conclusion: In a case study validating the CHA2DS2-VASc prediction model, we found substantial predictor misclassification in routine healthcare data with only limited effect on overall model performance. Our study should be repeated for other often applied prediction models to further evaluate the usefulness of routinely available healthcare data for validating prognostic models in the presence of predictor misclassification., Competing Interests: All data extracted from the electronic patient files were de-identified by a “trusted third party.” This study complied with the data protection law in the Netherlands. The medical ethics committee of the University Medical Centre Utrecht, the Netherlands, judged the CAFe study protocol as exempt from the review as it was conducted outside the criteria for the Medical Research Involving Human Subjects Act (WMO). Participating general practitioners provided written informed consent.Not applicableGJG is supported by a VENI grant from The Netherlands Organization for Scientific Research (ZONMW 016.166.030). RHHG is supported by a VIDI grant from The Netherlands Organization for Scientific Research (ZONMW 917.16.430). All other authors declare that they have no competing interests.Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.