Mobile electronic versus paper case report forms in clinical trials: a randomized controlled trial

An Overview of Data Management in Human Subjects Research

Research studies generate data in various forms. Data can be quantitative or qualitative. Research involving human subjects requires protection of data to ensure privacy. Various regulations and local policies need to be followed to ensure data security. Data management plans are critical for effective data stewardship and include details plan on data collection, management, storage, and formatting. This paper will review data collection tools, data security strategies, file management, data storage, government regulations, prepping data for analysis, reference management, and file management.

ISO/TS 21564:2019- based Evaluation of a Semantic Map between Variables in the ISARIC Freestanding Follow Up Survey and ORCHESTRA Studies

The COVID-19 pandemic has led to tremendous investment in clinical studies to generate much-needed knowledge on the prevention, diagnosis, treatment and long-term effects of the disease. Case report forms, comprised of questions and answers (variables), are commonly used to collect data in clinical trials. Maximizing the value of study data depends on data quality and on the ability to easily pool and share data from several sources. ISARIC, in collaboration with the WHO, has created a case report form that is available for use by the scientific community to collect COVID-19 trial data. One of such research initiatives collecting and analyzing multi-country and multi-cohort COVID-19 study data is the Horizon 2020 project ORCHESTRA. Following the ISO/TS 21564:2019 standard, a mapping between five ORCHESTRA studies' variables and the ISARIC Freestanding Follow-Up Survey elements was created. Measures of correspondence of shared semantic domain of 0 (perfect match), 1 (fully inclusive match), 2 (partial match), 4 (transformation required) or 4* (not present in ORCHESTRA) as compared to the target code system, ORCHESTRA study variables, were assigned to each of the elements in the ISARIC FUP case report form (CRF) which was considered the source code system. Of the ISARIC FUP CRF's variables, around 34% were found to show an exact match with corresponding variables in ORCHESTRA studies and about 33% showed a non-inclusive overlap. Matching variables provided information on patient demographics, COVID-19 testing, hospital admission and symptoms. More in-depth details are covered in ORCHESTRA variables with regards to treatment and comorbidities. ORCHESTRA's Long-Term Sequelae and Fragile population studies' CRFs include 32 and 27 variables respectively which were evaluated as a perfect match to variables in the ISARIC FUP CRF. Our study serves as an example of the kind of maps between case report form variables from different research projects needed to link ongoing COVID-19 research efforts and facilitate collaboration and data sharing. To enable data aggregation across two data systems, the information they contain needs to be connected through a map to determine compatibility and transformation needs. Combining data from various clinical studies can increase the power of analytical insights.

Rationale and design of a multicenter randomized clinical trial of vestibulodynia: understanding pathophysiology and determining appropriate treatments (vestibulodynia: UPDATe)

BACKGROUND

Limited data are available to establish evidence-based management protocols for vestibulodynia (VBD), a chronic vulvar pain condition that affects approximately 14 million women in the U.S. For the purposes of the study, our group subdivided VBD subtypes that may benefit from different types of treatment: 1) VBD peripheral (VBD-p), characterized by pain localized to the vulvar vestibule and 2) VBD central (VBD-c), characterized by VBD alongside one or more other chronic overlapping pain conditions (e.g. irritable bowel syndrome, temporomandibular disorder, and fibromyalgia syndrome) that affect remote body regions. Here, we describe the rationale and design of an NIH-funded multicenter clinical trial comparing the effectiveness of topical and/or systemic medication for alleviating pain and normalizing pain- relevant biomarkers among women with VBD-p and VBD-c.

METHODS

Participants will be randomly assigned to one of four parallel arms: peripheral treatment with 5% lidocaine + 0.5 mg/ml 0.02% oestradiol compound cream + oral placebo pill, 2) central treatment with the tricyclic antidepressant nortriptyline + placebo cream, 3) combined peripheral cream and central pill treatments, or 4) placebo cream and placebo pill. The treatment phase will last 16 weeks, with outcome measures and biomarkers assessed at 4 time points (0, 8, 16, and 24 weeks). First, we will compare the efficacy of treatments in alleviating pain using standardized tampon insertion with a numeric rating scale and self-reported pain on the short form McGill Pain Questionnaire. Next, we will compare the efficacy of treatments in improving perceived physical, mental, and sexual health using standardized questionnaires. Finally, we will measure cytokines and microRNAs in local vaginal and circulating blood samples using multiplex assays and RNA sequencing, and determine the ability of these biomarkers to predict treatment response.

CONCLUSION

This is the first multicenter randomized controlled trial to evaluate the efficacy of peripherally and centrally acting medications currently used in clinical practice for treating unique VBD subtypes based on distinct clinical and biological signatures.

ADMINISTRATIVE INFORMATION

Vestibulodynia UPDATe is a multi-centre, two-by-two factorial designed randomized, double-blind, placebo-controlled trial registered at clinical trials.gov (NCT03844412). This work is supported by the R01 HD096331 awarded to Drs. Nackley, Rapkin, Geller and Carey by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD).Key messagesPeripheral lidocaine and oestradiol and centrally-targeted nortriptyline medications are used for the treatment of pain in women with VBD, but there is a lack of data from well-powered RCTs.This two-by-two factorial RCT will test the efficacy of these medications in VBD subtypes characterized by distinct clinical characteristics and biomarker profiles.We hope that results will provide clinicians with scientific evidence of therapeutic efficacy in distinct VBD subtypes in an effort to direct and optimize treatment approaches.

The association between research funding status and clinical research papers’ citation impact in Japan: A cross-sectional bibliometric study

Introduction

Studies have not sufficiently clarified the differences in citation impact between funded and non-funded clinical research papers. Hence, this study seeks to evaluate the relation between research funding status and clinical research papers’ citation impact in different research fields using multiple evaluation indices.

Methods

In this cross-sectional bibliometric study, clinical research papers published by core clinical research hospitals in Japan were compared retrospectively in terms of times cited (TC), category normalized citation impact (CNCI), citation percentile (CP), journal impact factor (JIF), the Software to Identify, Manage, and Analyze Scientific Publications (SIGAPS) category, and whether they were the funded clinical research. The association between research funding status or the SIGAPS category and CNCI ≥ 2 was analyzed using logistic regression analysis.

Results

11 core clinical research hospitals published 553 clinical research papers, of which 120 were non-funded and 433 were funded (public institution-funded and industry-funded). The study found that funded clinical research papers (public institution-funded and industry-funded) had significantly higher TC, CNCI, CP, and JIF than non-funded ones [TC: 8 (3–17) vs. 14 (8–31), p < 0.001; CNCI: 0.53 (0.19–0.97) vs. 0.87 (0.45–1.85), p < 0.001; CP: 51.9 (24.48–70.42) vs. 66.7 (40.53–88.01), p < 0.001; JIF: 2.59 (1.90–3.84) vs. 2.93 (2.09–4.20) p = 0.008], while the proportion of A or B rank clinical research papers of the SIGAPS category was not significantly different between the two groups (30.0 vs. 34.9%, p = 0.318). In the logistic regression analysis, having a CNCI ≥ 2 was significantly associated with research funding (public institution-funded and industry-funded) and publication in A or B rank journals of the SIGAPS category [research funding: Estimate 2.169, 95% confidence interval (CI) 1.153–4.083, p = 0.016; SIGAPS category A/B: Estimate 6.126, 95% CI 3.889–9.651, p < 0.001].

Conclusion

Analysis via multiple indicators including CNCI and the SIGAPS category, which allows for a comparison of the papers’ citation impact in different research fields, found a positive relation between research funding status and the citation impact of clinical research papers.

The efficacy and safety of Jin's three-needle therapy vs. placebo acupuncture on anxiety symptoms in patients with post-stroke anxiety: A study protocol for a randomized controlled trial

BACKGROUND

A large number of clinical RCTs have verified that Jin's three-needle therapy (JTNT) has a great contribution to promoting the function of paralyzed limbs and relieving anxiety disorders for patients with post-stroke anxiety (PSA). However, there is still a lack of sham needle control, and its placebo effect cannot be ruled out. This study firstly verifies the real effectiveness of JTNT. Besides, the changes in serum indexes on the hypothalamic-pituitary-adrenal axis (HPA axis) are observed dynamically by the Enzyme-Linked ImmunoSorbent Assay (ELISA). The activation of different brain regions by JTNT is recorded using resting functional magnetic resonance imaging (rs-fMRI). Therefore, we can provide more practical and powerful evidence-based medical evidence for clinical decisions.

METHOD

This is a 16 week parallel, single-blind, random, controlled trial, including baseline, 4 weeks of treatment, and 12 weeks of follow-up. A total of 114 participants will be randomly divided into three groups in the proportion of 1:1:1. Participants will receive Jin's three-needle therapy in the active acupuncture group and accept sham needle treatment in the sham acupuncture group. In the waitlist control group, patients will not receive any acupuncture treatment. Outcomes cover three types of indicators, including scale indicators, serum indicators, and imaging indicators. The primary outcome is the change in the performance of anxiety symptoms, which is estimated by the 14-item Hamilton Anxiety Rating Scale (HAMA-14) and the 7-item Generalized Anxiety Disorder scale (GAD-7). Secondary outcomes are physical recovery and daily quality of life, which are evaluated by the National Institute of Health stroke scale (NIHSS) and the Modified Barthel Index Score (MBI Scale). Therefore, the assessment of the scale is carried out at baseline, 2nd, 4th, 8, 12, and 16 weeks. Adrenocorticotropin and cortisol will be quantitatively detected by ELISA at baseline and 4 weeks after treatment. In addition, regional homogeneity analysis (ReHo) will be used to record the activity of brain regions at baseline and 4 weeks after intervention.

DISCUSSION

The study aims to provide high-quality clinical evidence on the effectiveness and safety of JTNT for patients with PSA. In addition, this trial explores a possible mechanism of JTNT for patients with PSA.

CLINICAL TRIAL REGISTRATION

Chinese Clinical Trial Registry, identifier [ChiCTR2200058992].

The Successful Synchronized Orchestration of an Investigator-Initiated Multicenter Trial Using a Clinical Trial Management System and Team Approach: Design and Utility Study

BACKGROUND

As the cost of clinical trials continues to rise, novel approaches are required to ensure ethical allocation of resources. Multisite trials have been increasingly utilized in phase 1 trials for rare diseases and in phase 2 and 3 trials to meet accrual needs. The benefits of multisite trials include easier patient recruitment, expanded generalizability, and more robust statistical analyses. However, there are several problems more likely to arise in multisite trials, including accrual inequality, protocol nonadherence, data entry mistakes, and data integration difficulties.

OBJECTIVE

The Biostatistics & Data Science department at the University of Kansas Medical Center developed a clinical trial management system (comprehensive research information system [CRIS]) specifically designed to streamline multisite clinical trial management.

METHODS

A National Institute of Child Health and Human Development-funded phase 3 trial, the ADORE (assessment of docosahexaenoic acid [DHA] on reducing early preterm birth) trial fully utilized CRIS to provide automated accrual reports, centralize data capture, automate trial completion reports, and streamline data harmonization.

RESULTS

Using the ADORE trial as an example, we describe the utility of CRIS in database design, regulatory compliance, training standardization, study management, and automated reporting. Our goal is to continue to build a CRIS through use in subsequent multisite trials. Reports generated to suit the needs of future studies will be available as templates.

CONCLUSIONS

The implementation of similar tools and systems could provide significant cost-saving and operational benefit to multisite trials.

TRIAL REGISTRATION

ClinicalTrials.gov NCT02626299; https://tinyurl.com/j6erphcj.

Usability Evaluation of an Offline Electronic Data Capture App in a Prospective Multicenter Dementia Registry (digiDEM Bayern): Mixed Method Study

BACKGROUND

Digital registries have been shown to provide an efficient way of gaining a better understanding of the clinical complexity and long-term progression of diseases. The paperless method of electronic data capture (EDC) during a patient interview saves both time and resources. In the prospective multicenter project "Digital Dementia Registry Bavaria (digiDEM Bayern)," interviews are also performed on site in rural areas with unreliable internet connectivity. It must be ensured that EDC can still be performed in such a context and that there is no need to fall back on paper-based questionnaires. In addition to a web-based data collection solution, the EDC system REDCap (Research Electronic Data Capture) offers the option to collect data offline via an app and to synchronize it afterward.

OBJECTIVE

The aim of this study was to evaluate the usability of the REDCap app as an offline EDC option for a lay user group and to examine the necessary technology acceptance of using mobile devices for data collection. The feasibility of the app-based offline data collection in the digiDEM Bayern dementia registry project was then evaluated before going live.

METHODS

An exploratory mixed method design was employed in the form of an on-site usability test with the "Thinking Aloud" method combined with an online questionnaire including the System Usability Scale (SUS). The acceptance of mobile devices for data collection was surveyed based on five categories of the technology acceptance model.

RESULTS

Using the "Thinking Aloud" method, usability issues were identified and solutions were accordingly derived. Evaluation of the REDCap app resulted in a SUS score of 74, which represents "good" usability. After evaluating the technology acceptance questionnaire, it can be concluded that the lay user group is open to mobile devices as interview tools.

CONCLUSIONS

The usability evaluation results show that a lay user group generally agree that data collecting partners in the digiDEM project can handle the REDCap app well. The usability evaluation provided statements about positive aspects and could also identify usability issues relating to the REDCap app. In addition, the current technology acceptance in the sample showed that heterogeneous groups of different ages with diverse experiences in handling mobile devices are also ready for the use of app-based EDC systems. Based on these results, it can be assumed that the offline use of an app-based EDC system on mobile devices is a viable solution for collecting data in a decentralized registry-based research project.

A look to the future: Pandemic-induced digital technologies in vascular surgery

Like many areas of medicine, vascular surgery has been transformed by the COVID-19 (coronavirus disease 2019) pandemic. Public health precautions to minimize disease transmission have led to reduced attendance at hospitals and clinics in elective and emergency settings; fewer face-to-face and hands-on clinical interactions; and increased reliance on telemedicine, virtual attendance, investigations, and digital therapeutics. However, a “silver lining” to the COVID-19 pandemic may be the mainstream acceptance and acceleration of telemedicine, remote monitoring, digital health technology, and three-dimensional technologies, such as three-dimensional printing and virtual reality, by connecting health care providers to patients in a safe, reliable, and timely manner, and supplanting face-to-face surgical simulation and training. This review explores the impact of these changes in the delivery of vascular surgical care.

Adaptation of a quality improvement approach to implement eScreening in VHA healthcare settings: innovative use of the Lean Six Sigma Rapid Process Improvement Workshop

BACKGROUND

The Veterans Health Administration (VHA) developed a comprehensive mobile screening technology (eScreening) that provides customized and automated self-report health screening via mobile tablet for veterans seen in VHA settings. There is agreement about the value of health technology, but limited knowledge of how best to broadly implement and scale up health technologies. Quality improvement (QI) methods may offer solutions to overcome barriers related to broad scale implementation of technology in health systems. We aimed to develop a process guide for eScreening implementation in VHA clinics to automate self-report screening of mental health symptoms and psychosocial challenges.

METHODS

This was a two-phase, mixed methods implementation project building on an adapted quality improvement method. In phase one, we adapted and conducted an RPIW to develop a generalizable process guide for eScreening implementation (eScreening Playbook). In phase two, we integrated the eScreening Playbook and RPIW with additional strategies of training and facilitation to create a multicomponent implementation strategy (MCIS) for eScreening. We then piloted the MCIS in two VHA sites. Quantitative eScreening pre-implementation survey data and qualitative implementation process "mini interviews" were collected from individuals at each of the two sites who participated in the implementation process. Survey data were characterized using descriptive statistics, and interview data were independently coded using a rapid qualitative analytic approach.

RESULTS

Pilot data showed overall satisfaction and usefulness of our MCIS approach and identified some challenges, solutions, and potential adaptations across sites. Both sites used the components of the MCIS, but site 2 elected not to include the RPIW. Survey data revealed positive responses related to eScreening from staff at both sites. Interview data exposed implementation challenges related to the technology, support, and education at both sites. Workflow and staffing resource challenges were only reported by site 2.

CONCLUSIONS

Our use of RPIW and other QI methods to both develop a playbook and an implementation strategy for eScreening has created a testable implementation process to employ automated, patient-facing assessment. The efficient collection and communication of patient information have the potential to greatly improve access to and quality of healthcare.

Translation and validation of an extended German version of ID Migraine™ as a migraine screening tool

Background and purpose:

Diagnosing a patient with headache as a migraineur is critical for state-of-the-art migraine management. Screening tools are imperative means to improve the diagnostic yield in the primary care settings and specialized clinics. This study aims to translate and assess the diagnostic accuracy of a German version of the ID Migraine™ as a widely used and efficient screening instrument.

Methods:

The Functional Assessment of Chronic Illness Therapy translation methodology was used to translate the original three-item ID Migraine™, including a fourth question for aura, from the English language into the German language. Diagnostic accuracy of the German ID Migraine™ and predictors of false screening results were assessed among patients presenting to a headache outpatient clinic of a tertiary care center in Germany over a 6-month period.

Results:

The translation procedure yielded a harmonized German ID Migraine™ and its diagnostic accuracy was assessed in 105 patients (80 female, 46.5 ± 17.2 years of age), including 79 patients (75.2%) with migraine. The three-item German ID Migraine™ provides a sensitivity of 99%, specificity of 68%, and positive and negative predictive values of 90% and 95%, respectively, using a cutoff of ≥2. Positive and negative predictive values in a general headache population are estimated to be 74% and 98%, respectively. The aura question identified 18 out of 20 migraineurs with aura.

Conclusions:

The German ID Migraine™ is an accurate screening tool for migraine even in a challenging population of a specialized outpatient clinic. Its diagnostic accuracy indicates a potential utility for screening in primary health care.

Applicability of care quality indicators for women with low-risk pregnancies planning hospital birth: a retrospective study of medical records

Practices for planned birth among women with low-risk pregnancies vary by birth setting, medical professional, and organizational system. Appropriate monitoring is essential for quality improvement. Although sets of quality indicators have been developed, their applicability has not been tested. To improve the quality of childbirth care for low-risk mothers and infants in Japanese hospitals, we developed 35 quality indicators using existing clinical guidelines and quality indicators. We retrospectively analysed data for 347 women in Japan diagnosed with low-risk pregnancy in the second trimester, admitted between April 2015 and March 2016. We obtained scores for 35 quality indicators and evaluated their applicability, i.e., feasibility, improvement potential, and reliability (intra- and inter-rater reliability: kappa score, positive and negative agreement). The range of adherence to each indicator was 0-95.7%. We identified feasibility concerns for six indicators with over 25% missing data. Two indicators with over 90% adherence showed limited potential for improvement. Three indicators had poor kappa scores for intra-rater reliability, with positive/negative agreement scores 0.94/0.33, 0.33/0.95, and 0.00/0.97, respectively. Two indicators had poor kappa scores for inter-rater reliability, with positive/negative agreement scores 0.25/0.92 and 0.68/0.61, respectively. The findings indicated that these 35 care quality indicators for low-risk pregnant women may be applicable to real-world practice, with some caveats.

Electronic Data Capture Versus Conventional Data Collection Methods in Clinical Pain Studies: Systematic Review and Meta-Analysis

BACKGROUND

The most commonly used means to assess pain is by patient self-reported questionnaires. These questionnaires have traditionally been completed using paper-and-pencil, telephone, or in-person methods, which may limit the validity of the collected data. Electronic data capture methods represent a potential way to validly, reliably, and feasibly collect pain-related data from patients in both clinical and research settings.

OBJECTIVE

The aim of this study was to conduct a systematic review and meta-analysis to compare electronic and conventional pain-related data collection methods with respect to pain score equivalence, data completeness, ease of use, efficiency, and acceptability between methods.

METHODS

We searched the Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica Database (EMBASE), and Cochrane Central Register of Controlled Trials (CENTRAL) from database inception until November 2019. We included all peer-reviewed studies that compared electronic (any modality) and conventional (paper-, telephone-, or in-person-based) data capture methods for patient-reported pain data on one of the following outcomes: pain score equivalence, data completeness, ease of use, efficiency, and acceptability. We used random effects models to combine score equivalence data across studies that reported correlations or measures of agreement between electronic and conventional pain assessment methods.

RESULTS

A total of 53 unique studies were included in this systematic review, of which 21 were included in the meta-analysis. Overall, the pain scores reported electronically were congruent with those reported using conventional modalities, with the majority of studies (36/44, 82%) that reported on pain scores demonstrating this relationship. The weighted summary correlation coefficient of pain score equivalence from our meta-analysis was 0.92 (95% CI 0.88-0.95). Studies on data completeness, patient- or provider-reported ease of use, and efficiency generally indicated that electronic data capture methods were equivalent or superior to conventional methods. Most (19/23, 83%) studies that directly surveyed patients reported that the electronic format was the preferred data collection method.

CONCLUSIONS

Electronic pain-related data capture methods are comparable with conventional methods in terms of score equivalence, data completeness, ease, efficiency, and acceptability and, if the appropriate psychometric evaluations are in place, are a feasible means to collect pain data in clinical and research settings.

The Video intervention to Inspire Treatment Adherence for Life (VITAL Start): protocol for a multisite randomized controlled trial of a brief video-based intervention to improve antiretroviral adherence and retention among HIV-infected pregnant women in Malawi

BACKGROUND

Improving maternal antiretroviral therapy (ART) retention and adherence is a critical challenge facing prevention of mother-to-child transmission (PMTCT) of HIV programs. There is an urgent need for evidence-based, cost-effective, and scalable interventions to improve maternal adherence and retention that can be feasibly implemented in overburdened health systems. Brief video-based interventions are a promising but underutilized approach to this crisis. We describe a trial protocol to evaluate the effectiveness and implementation of a standardized educational video-based intervention targeting HIV-infected pregnant women that seeks to optimize their ART retention and adherence by providing a VITAL Start (Video intervention to Inspire Treatment Adherence for Life) before committing to lifelong ART.

METHODS

This study is a multisite parallel group, randomized controlled trial assessing the effectiveness of a brief facility-based video intervention to optimize retention and adherence to ART among pregnant women living with HIV in Malawi. A total of 892 pregnant women living with HIV and not yet on ART will be randomized to standard-of-care pre-ART counseling or VITAL Start. The primary outcome is a composite of retention and adherence (viral load < 1000 copies/ml) 12 months after starting ART. Secondary outcomes include assessments of behavioral adherence (self-reported adherence, pharmacy refill, and tenofovir diphosphate concentration), psychosocial impact, and resource utilization. We will also examine the implementation of VITAL Start via surveys and qualitative interviews with patients, partners, and health care workers and conduct cost-effectiveness analyses.

DISCUSSION

This is a robust evaluation of an innovative facility-based video intervention for pregnant women living with HIV, with the potential to improve maternal and infant outcomes.

TRIAL REGISTRATION

ClinicalTrials.gov, NCT03654898. Registered on 31 August 2018.

Implementing eScreening technology in four VA clinics: a mixed-method study

Background

Technology-based self-assessment (TB-SA) benefits patients and providers and has shown feasibility, ease of use, efficiency, and cost savings. A promising TB-SA, the VA eScreening program, has shown promise for the efficient and effective collection of mental and physical health information. To assist adoption of eScreening by healthcare providers, we assessed technology-related as well as individual- and system-level factors that might influence the implementation of eScreening in four diverse VA clinics.

Methods

This was a mixed-method, pre-post, quasi-experimental study originally designed as a quality improvement project. The clinics were selected to represent a range of environments that could potentially benefit from TB-SA and that made use of the variety eScreening functions. Because of limited resources, the implementation strategy consisted of staff education, training, and technical support as needed. Data was collected using pre- and post-implementation interviews or focus groups of leadership and clinical staff, eScreening usage data, and post-implementation surveys. Data was gathered on: 1) usability of eScreening; 2) knowledge about and acceptability and 3) facilitators and barriers to the successful implementation of eScreening.

Results

Overall, staff feedback about eScreening was positive. Knowledge about eScreening ranged widely between the clinics. Nearly all staff felt eScreening would fit well into their clinical setting at pre-implementation; however some felt it was a poor fit with emergent cases and older adults at post-implementation. Lack of adequate personnel support and perceived leadership support were barriers to implementation. Adequate training and technical assistance were cited as important facilitators. One clinic fully implemented eScreening, two partially implemented, and one clinic did not implement eScreening as part of normal practice after 6 months as measured by usage data and self-report. Organizational engagement survey scores were higher among clinics with full or partial implementation and low in the clinic that did not implement.

Conclusions

Despite some added work load for some staff and perceived lack of leadership support, eScreening was at least partially implemented in three clinics. The technology itself posed no barriers in any of the settings. An implementation strategy that accounts for increased work burden and includes accountability may help in future eScreening implementation efforts.

Note. This abstract was previously published (e.g., Annals of Behavioral Medicine 53: S1–S842, 2019).

Electronic supplementary material

The online version of this article (10.1186/s12913-019-4436-z) contains supplementary material, which is available to authorized users.

Improving the Efficacy of the Data Entry Process for Clinical Research With a Natural Language Processing-Driven Medical Information Extraction System: Quantitative Field Research

BACKGROUND

The growing interest in observational trials using patient data from electronic medical records poses challenges to both efficiency and quality of clinical data collection and management. Even with the help of electronic data capture systems and electronic case report forms (eCRFs), the manual data entry process followed by chart review is still time consuming.

OBJECTIVE

To facilitate the data entry process, we developed a natural language processing-driven medical information extraction system (NLP-MIES) based on the i2b2 reference standard. We aimed to evaluate whether the NLP-MIES-based eCRF application could improve the accuracy and efficiency of the data entry process.

METHODS

We conducted a randomized and controlled field experiment, and 24 eligible participants were recruited (12 for the manual group and 12 for NLP-MIES-supported group). We simulated the real-world eCRF completion process using our system and compared the performance of data entry on two research topics, pediatric congenital heart disease and pneumonia.

RESULTS

For the congenital heart disease condition, the NLP-MIES-supported group increased accuracy by 15% (95% CI 4%-120%, P=.03) and reduced elapsed time by 33% (95% CI 22%-42%, P<.001) compared with the manual group. For the pneumonia condition, the NLP-MIES-supported group increased accuracy by 18% (95% CI 6%-32%, P=.008) and reduced elapsed time by 31% (95% CI 19%-41%, P<.001).

CONCLUSIONS

Our system could improve both the accuracy and efficiency of the data entry process.

Outcome prediction of out-of-hospital cardiac arrest with presumed cardiac aetiology using an advanced machine learning technique

BACKGROUND

Outcome prediction for patients with out-of-hospital cardiac arrest (OHCA) has the possibility to detect patients who could have been potentially saved. Advanced machine learning techniques have recently been developed and employed for clinical studies. In this study, we aimed to establish a prognostication model for OHCA with presumed cardiac aetiology using an advanced machine learning technique.

METHODS AND RESULTS

Cohort data from a prospective multi-centre cohort study for OHCA patients transported by an ambulance in the Kanto area of Japan between January 2012 and March 2013 (SOS-KANTO 2012 study) were analysed in this study. Of 16,452 patients, data for OHCA patients aged ≥18 years with presumed cardiac aetiology were retrieved, and were divided into two groups (training set: n = 5718, between January 1, 2012 and December 12, 2012; test set: n = 1608, between January 1, 2013 and March 31, 2013). Of 421 variables observed during prehospital and emergency department settings, 35 prehospital variables, or 35 prehospital and 18 in-hospital variables, were used for outcome prediction of 1-year survival using a random forest method. In validation using the test set, prognostication models trained with 35 variables, or 53 variables for 1-year survival showed area under the receiver operating characteristics curve (AUC) values of 0.943 (95% CI [0.930, 0.955]) and 0.958 (95% CI [0.948, 0.969]), respectively.

CONCLUSIONS

The advanced machine learning technique showed favourable prediction capability for 1-year survival of OHCA with presumed cardiac aetiology. These models can be useful for detecting patients who could have been potentially saved.

A Review of Clinical Data Management Systems Used in Clinical Trials

BACKGROUND

A clinical data management system is a software supporting the data management process in clinical trials. In this system, the effective support of clinical data management dimensions leads to the increased accuracy of results and prevention of diversion in clinical trials. The aim of this review article was to investigate the dimensions of data management in clinical data management systems.

METHODS

This study was conducted in 2017. The used databases included Web of Science, Scopus, Science Direct, ProQuest, Ovid Medline and PubMed. The search was conducted over a period of 10 years from 2007 to 2017. The initial number of studies was 101 reaching 19 in the final stage. The final studies were described and compared in terms of the year, country and dimensions of the clinical data management process in clinical trials.

RESULTS

The research findings indicated that none of the systems completely supported the data management dimensions in clinical trials. Although these systems were developed for supporting the clinical data management process, they were similar to electronic data capture systems in many cases. The most significant dimensions of data management in such systems were data collection or entry, report, validation, and security maintenance.

CONCLUSION

Seemingly, not sufficient attention has been paid to automate all dimensions of the clinical data management process in clinical trials. However, these systems could take positive steps towards changing the manual processes of clinical data management to electronic processes.