Quality, quantity and harmony: the DataSHaPER approach to integrating data across bioclinical studies

Evaluating the harmonization potential of oral health-related questionnaires in national longitudinal birth and child cohort surveys

BACKGROUND/OBJECTIVES

Effective use of longitudinal study data is challenging because of divergences in the construct definitions and measurement approaches over time, between studies and across disciplines. One approach to overcome these challenges is data harmonization. Data harmonization is a practice used to improve variable comparability and reduce heterogeneity across studies. This study describes the process used to evaluate the harmonization potential of oral health-related variables across each survey wave.

METHODS

National child cohort surveys with similar themes/objectives conducted in the last two decades were selected. The Maelstrom Research Guidelines were followed for harmonization potential evaluation.

RESULTS

Seven nationally representative child cohort surveys were included and questionnaires examined from 50 survey waves. Questionnaires were classified into three domains and fifteen constructs and summarized by age groups. A DataSchema (a list of core variables representing the suitable version of the oral health outcomes and risk factors) was compiled comprising 42 variables. For each study wave, the potential (or not) to generate each DataSchema variable was evaluated. Of the 2100 harmonization status assessments, 543 (26%) were complete. Approximately 50% of the DataSchema variables can be generated across at least four cohort surveys while only 10% (n = 4) variables can be generated across all surveys. For each survey, the DataSchema variables that can be generated ranged between 26% and 76%.

CONCLUSION

Data harmonization can improve the comparability of variables both within and across surveys. For future cohort surveys, the authors advocate more consistency and standardization in survey questionnaires within and between surveys.

Pioneering a Multi-Phase Framework to Harmonize Self-Reported Sleep Data Across Cohorts

STUDY OBJECTIVES

Harmonizing and aggregating data across studies enable pooled analyses that support external validation and enhance replicability and generalizability. However, the multidimensional nature of sleep poses challenges for data harmonization and aggregation. Here we describe and implement our process for harmonizing self-reported sleep data.

METHODS

We established a multi-phase framework to harmonize self-reported sleep data: (1) compile items; (2) group items into domains; (3) harmonize items; and (4) evaluate harmonizability. We applied this process to produce a pooled multi-cohort sample of five United States cohorts plus a separate yet fully harmonized sample from Rotterdam, Netherlands. Sleep and sociodemographic data are described and compared to demonstrate the utility of harmonization and aggregation.

RESULTS

We collected 190 unique self-reported sleep items and grouped them into 15 conceptual domains. Using these domains as guiderails, we developed 14 harmonized items measuring aspects of Satisfaction, Alertness/Sleepiness, Timing, Efficiency, Duration, Insomnia, and Sleep Apnea. External raters determined that 13 of these 14 items had moderate-to-high harmonizability. Alertness/Sleepiness items had lower harmonizability, while continuous, quantitative items (e.g., timing, total sleep time, efficiency) had higher harmonizability. Descriptive statistics identified features that are more consistent (e.g., wake-up time, duration) and more heterogeneous (e.g., time in bed, bedtime) across samples.

CONCLUSIONS

Our process can guide researchers and cohort stewards towards effective sleep harmonization and provides a foundation for further methodological development in this expanding field. Broader national and international initiatives promoting common data elements across cohorts are needed to enhance future harmonization and aggregation efforts.

Longitudinal natural history studies based on real-world data in rare diseases: Opportunity and a novel approach

Growing interest in therapeutic development for rare diseases necessitate a systematic approach to the collection and curation of natural history data that can be applied consistently across this group of heterogenous rare diseases. In this study, we discuss the challenges facing natural history studies for leukodystrophies and detail a novel standardized approach to creating a longitudinal natural history study using existing medical records. Prospective studies are uniquely challenging for rare diseases. Delays in diagnosis and overall rarity limit the timely collection of natural history data. When feasible, prospective studies are often cross-sectional rather than longitudinal and are unlikely to capture pre- or early- symptomatic disease trajectories, limiting their utility in characterizing the full natural history of the disease. Therapeutic development in leukodystrophies is subject to these same obstacles. The Global Leukodystrophy Initiative Clinical Trials Network (GLIA-CTN) comprises of a network of research institutions across the United States, supported by a multi-center biorepository protocol, to map the longitudinal clinical course of disease across leukodystrophies. As part of GLIA-CTN, we developed Standard Operating Procedures (SOPs) that delineated all study processes related to staff training, source documentation, and data sharing. Additionally, the SOP detailed the standardized approach to data extraction including diagnosis, clinical presentation, and medical events, such as age at gastrostomy tube placement. The key variables for extraction were selected through face validity, and common electronic case report forms (eCRF) across leukodystrophies were created to collect analyzable data. To enhance the depth of the data, clinical notes are extracted into "original" and "imputed" encounters, with imputed encounter referring to a historic event (e.g., loss of ambulation 3 months prior). Retrospective Functional Assessments were assigned by child neurologists, using a blinded dual-rater approach and score discrepancies were adjudicated by a third rater. Upon completion of extraction, data source verification is performed. Data missingness was evaluated using statistics. The proposed methodology will enable us to leverage existing medical records to address the persistent gap in natural history data within this unique disease group, allow for assessment of clinical trajectory both pre- and post-formal diagnosis, and promote recruitment of larger cohorts.

Diet patterns associated with cognitive decline: methods to harmonize data from European and US cohort studies

The impact of dietary intake on cognitive outcomes and dementia prevention is a topic of increasing interest. Meta-analyses of observational studies, mostly conducted within US and European populations, have reported benefits of healthy diet patterns on cognitive performance, but results from individual studies have been inconsistent. These inconsistencies are likely due to the diverse methodology used in studies, including different diet and cognitive function assessment instruments, follow-up periods, and analytical methods, which make drawing conclusions relevant to dietary guidance challenging. The objective of this project is to describe a protocol to conduct a retrospective harmonization study on dietary intake and cognitive health using data from European and US studies. The recommendations resulting from the project can be used to support evidence-based synthesis for future iterations of the Dietary Guidelines for Americans or other population-based dietary guidance. Additionally, this study will serve as a harmonization guide for future research on the relationship between diet patterns and cognition. The approach outlined ultimately aims to optimize resources and expedite research efforts for dementia prevention.

Standardizing digital biobanks: integrating imaging, genomic, and clinical data for precision medicine

Advancements in data acquisition and computational methods are generating a large amount of heterogeneous biomedical data from diagnostic domains such as clinical imaging, pathology, and next-generation sequencing (NGS), which help characterize individual differences in patients. However, this information needs to be available and suitable to promote and support scientific research and technological development, supporting the effective adoption of the precision medicine approach in clinical practice. Digital biobanks can catalyze this process, facilitating the sharing of curated and standardized imaging data, clinical, pathological and molecular data, crucial to enable the development of a comprehensive and personalized data-driven diagnostic approach in disease management and fostering the development of computational predictive models. This work aims to frame this perspective, first by evaluating the state of standardization of individual diagnostic domains and then by identifying challenges and proposing a possible solution towards an integrative approach that can guarantee the suitability of information that can be shared through a digital biobank. Our analysis of the state of the art shows the presence and use of reference standards in biobanks and, generally, digital repositories for each specific domain. Despite this, standardization to guarantee the integration and reproducibility of the numerical descriptors generated by each domain, e.g. radiomic, pathomic and -omic features, is still an open challenge. Based on specific use cases and scenarios, an integration model, based on the JSON format, is proposed that can help address this problem. Ultimately, this work shows how, with specific standardization and promotion efforts, the digital biobank model can become an enabling technology for the comprehensive study of diseases and the effective development of data-driven technologies at the service of precision medicine.

Relationship Between Diet Quality and Statin Use Among Adults With Metabolic Syndrome From the CARTaGENE Cohort

Background

In metabolic syndrome (MetS), cardiovascular disease (CVD) risk reduction relies on the complementary use of diet and lipid-lowering medication. Evidence suggests that initiating such medication may impede diet quality. The objective of this study was to evaluate the relationship between diet quality and statin use among adults with MetS and free of CVD from the Province of Québec.

Methods

This cross-sectional study included 2481 adults with MetS (40-69 years of age) from the CARTaGENE Québec population-based cohort, of whom 463 self-reported using statin monotherapy. Diet was assessed using the Canadian Dietary History Questionnaire II, a food- frequency questionnaire, and diet quality was assessed using the Alternative Healthy Eating Index (AHEI).

Results

In multivariable-adjusted linear regression models, statin users had lower AHEI (%) compared with nonusers (users: 40.0; 95% confidence interval [CI], 38.9, 41.2 vs nonusers: 41.2; 95% CI, 40.4, 42.0; P = 0.03] because of a lower consumption of vegetables and whole grains. Stratified interaction analyses showed that the lower diet quality among statin users was mostly prevalent among men aged ≥ 50 years and women aged ≥ 60 years, among individuals with annual household incomes of < $50,000 and persons who self-reported history of high blood pressure.

Conclusions

In this cohort of adults with MetS from Quebéc, the use of statin monotherapy in primary prevention of CVD was associated with a slightly lower diet quality. These data suggest suboptimal complementarity between diet quality and use of cholesterol-lowering medication in primary prevention of CVD in MetS.

Relationship Between Diet Quality and Glucose-Lowering Medication Intensity Among Adults With Type 2 Diabetes: Results From the CARTaGENE Cohort

Background

In real-world settings, whether diet and medication are used as complements for glycemic management in type 2 diabetes (T2D) remains unclear. This study assessed the relationship between diet quality and intensity of glucose-lowering medication among adults with T2D.

Methods

This cross-sectional study included 352 adults with T2D from the CARTaGENE Québec population-based cohort. Diet quality was assessed using the Healthful Plant-Based Diet Index (hPDI). Glucose-lowering medication intensity was graded according to self-reported information on the type and number of drugs: no medication; oral monotherapy; oral polytherapy; and insulin with and without oral medication. In the subsample of 239 individuals who reported the medication dosages, intensity was also graded using the Medication Effect Score (MES).

Results

In multivariable-adjusted models, we found no evidence of a relationship between the hPDI and medication intensity, assessed using the categorical approach (Pbetween-group = 0.25) or the MES (P = 0.43). However, the hPDI was inversely associated with the MES among men < 50 years of age and women < 60 years (β1-point MES = -2.24 [95% confidence interval, -4.46, -0.02] hPDI points), but not among older individuals (β = -0.03 [-1.28, 1.21] hPDI points). Evidence of a nonsignificant inverse relationship between the hPDI and HbA1c was observed (β10-point hPDI = -0.23% [-0.63, 0.17]), whereas a positive and significant association between the MES and hemoglobin (Hb)A1c was found (β1-point MES = 0.30% [0.10, 0.51]).

Conclusions

In this cohort of adults with T2D, there was an overall lack of complementarity between diet quality and intensity of glucose-lowering medication. The issue was particularly important among younger adults for whom diet quality was inversely associated with intensity of medication.

Relationship Between Diet Quality and Antihypertensive Medication Intensity Among Adults With Metabolic Syndrome-Associated High Blood Pressure

Background

Management of high blood pressure (BP), a key feature of the metabolic syndrome (MetS), relies on diet and medication. Whether these modalities are used as complements has never been evaluated in real-world settings. This study assessed the relationship between diet quality and antihypertensive medication intensity among adults with MetS-associated high BP.

Methods

This cross-sectional study included 915 adults with MetS-associated high BP from the CARTaGENE cohort (Québec, Canada), of whom 677 reported using BP-lowering medication. Antihypertensive medication intensity was graded per the number of BP-lowering classes used simultaneously. Diet quality was assessed using the Dietary Approach to Stop Hypertension (DASH) score.

Results

No evidence of a relationship between antihypertensive medication intensity and diet quality was found (β for each additional antihypertensive = -0.05; 95% CI, -0.35; 0.26 DASH score points). However, among men aged < 50 years and women aged < 60 years, the DASH score was inversely associated with medication intensity (β = -0.72; 95% CI, -1.24, -0.19), whereas this relationship tended to be positive among older participants (β = 0.32; 95% CI, -0.05, 0.69). Among participants with low Framingham risk score, the DASH score was inversely associated with medication intensity (β = -0.70; 95% CI, -1.31, -0.09), but no evidence of an association was found among individuals at moderate (β = 0.00; 95% CI, -0.45, 0.45) or high (β = 0.30, 95% CI, -0.24, 0.84) risk.

Conclusions

In this cohort of adults with MetS-associated high BP, there was an overall lack of complementarity between diet quality and BP-lowering medication, especially among younger individuals and those with a lower risk for cardiovascular disease for whom diet quality was inversely associated with intensity of medication.

Sharing Medical Big Data While Preserving Patient Confidentiality in Innovative Medicines Initiative: A Summary and Case Report from BigData@Heart

Sharing individual patient data (IPD) is a simple concept but complex to achieve due to data privacy and data security concerns, underdeveloped guidelines, and legal barriers. Sharing IPD is additionally difficult in big data-driven collaborations such as Bigdata@Heart in the Innovative Medicines Initiative, due to competing interests between diverse consortium members. One project within BigData@Heart, case study 1, needed to pool data from seven heterogeneous data sets: five randomized controlled trials from three different industry partners, and two disease registries. Sharing IPD was not considered feasible due to legal requirements and the sensitive medical nature of these data. In addition, harmonizing the data sets for a federated data analysis was difficult due to capacity constraints and the heterogeneity of the data sets. An alternative option was to share summary statistics through contingency tables. Here it is demonstrated that this method along with anonymization methods to ensure patient anonymity had minimal loss of information. Although sharing IPD should continue to be encouraged and strived for, our approach achieved a good balance between data transparency while protecting patient privacy. It also allowed a successful collaboration between industry and academia.

Large-Scale Data Harmonization Across Prospective Studies

The Preconception Period Analysis of Risks and Exposures Influencing Health and Development (PrePARED) Consortium creates a novel resource for addressing preconception health by merging data from numerous cohort studies. In this paper, we describe our data harmonization methods and results. Individual-level data from 12 prospective studies were pooled. The crosswalk-cataloging-harmonization procedure was used. The index pregnancy was defined as the first postbaseline pregnancy lasting more than 20 weeks. We assessed heterogeneity across studies by comparing preconception characteristics in different types of studies. The pooled data set included 114,762 women, and 25,531 (22%) reported at least 1 pregnancy of more than 20 weeks' gestation during the study period. Babies from the index pregnancies were delivered between 1976 and 2021 (median, 2008), at a mean maternal age of 29.7 (standard deviation, 4.6) years. Before the index pregnancy, 60% of women were nulligravid, 58% had a college degree or more, and 37% were overweight or obese. Other harmonized variables included race/ethnicity, household income, substance use, chronic conditions, and perinatal outcomes. Participants from pregnancy-planning studies had more education and were healthier. The prevalence of preexisting medical conditions did not vary substantially based on whether studies relied on self-reported data. Use of harmonized data presents opportunities to study uncommon preconception risk factors and pregnancy-related events. This harmonization effort laid the groundwork for future analyses and additional data harmonization.

Facilitating Harmonization of Variables in Framingham, MESA, ARIC, and REGARDS Studies Through a Metadata Repository

BACKGROUND

High-quality research in cardiovascular prevention, as in other fields, requires inclusion of a broad range of data sets from different sources. Integrating and harmonizing different data sources are essential to increase generalizability, sample size, and representation of understudied populations-strengthening the evidence for the scientific questions being addressed.

METHODS

Here, we describe an effort to build an open-access repository and interactive online portal for researchers to access the metadata and code harmonizing data from 4 well-known cohort studies-the REGARDS (Reasons for Geographic and Racial Differences in Stroke) study, FHS (Framingham Heart Study), MESA (Multi-Ethnic Study of Atherosclerosis), and ARIC (Atherosclerosis Risk in Communities) study. We introduce a methodology and a framework used for preprocessing and harmonizing variables from multiple studies.

RESULTS

We provide a real-case study and step-by-step guidance to demonstrate the practical utility of our repository and interactive web page. In addition to our successful development of such an open-access repository and interactive web page, this exercise in harmonizing data from multiple cohort studies has revealed several key themes. These themes include the importance of careful preprocessing and harmonization of variables, the value of creating an open-access repository to facilitate collaboration and reproducibility, and the potential for using harmonized data to address important scientific questions and disparities in cardiovascular disease research.

CONCLUSIONS

By integrating and harmonizing these large-scale cohort studies, such a repository may improve the statistical power and representation of understudied cohorts, enabling development and validation of risk prediction models, identification and investigation of risk factors, and creating a platform for racial disparities research.

REGISTRATION

URL: https://precision.heart.org/duke-ninds.

Relationship between lifestyle habits and cardiovascular risk factors in familial hypercholesterolemia

BACKGROUND AND AIM

Little is known about the cardioprotective potential of a healthy lifestyle in familial hypercholesterolemia (FH). The objective of this study was to evaluate the relationship between lifestyle and cardiovascular risk factors in adults with FH.

METHODS AND RESULTS

This cross-sectional study leveraged data from the CARTaGENE Quebec population-based cohort (Canada). Participants with FH were identified using the validated Simplified Canadian Definition for FH. A healthy lifestyle score (HLS), ranging from 0 to 5, was calculated per adherence to 5 lifestyle habits: 1) not smoking; 2) being physically active (≥150 min/week of moderate or vigorous physical activity); 3) eating a healthy diet (Alternate Healthy Eating Index ≥50%); 4) having a light to moderate alcohol consumption (men: 1-30 g/day; women: 1-15 g/day); and 5) sleeping 7-8 h/day. Among the 122 included individuals (women, n = 78; men, n = 44; mean age ± SD: 57.3 ± 6.7 years), 92 (75.4%) had a HLS ≤3/5, while only 5 (4.1%) had a HLS of 5/5. After adjustments for sex, age, body mass index, and lipid-lowering medication use, we found no evidence of an association between the HLS and concentrations of LDL-cholesterol (β = 0.04, 95% CI = -0.08, 0.15 mmol/L; P = 0.54). However, the HLS was favorably associated with HbA1c levels (β = -0.07, 95% CI = -0.13, -0.01%; P = 0.02), and statistical trends suggested favorable associations with HDL-cholesterol (β = 0.06, 95% CI = -0.02, 0.14 mmol/L; P = 0.06) and waist circumference (β = -2.22, 95% CI = -4.62, 0.17 cm; P = 0.07).

CONCLUSION

This study suggests that a healthy lifestyle is favorably associated with CVD risk factors in adults with FH.

Developing Policy for the Healthy Life Trajectories Initiative: Going from National to International

Background: Scientific research is becoming an increasingly collaborative and global venture. The Healthy Life Trajectories Initiative (HeLTI), for instance, is an international Developmental Origins of Health and Disease research collaboration developed to address the increasing burden of noncommunicable diseases around the world. It comprises four separate but harmonized cohort trials in Canada, China, India, and South Africa. These cohorts will generate rich data and biosample sets that can be shared both within the HeLTI Consortium and with other researchers from around the world. Methods: To ensure the coordination and operation of these types of collaborative research initiatives, a standardized and harmonized governance model is required to regulate the processes and interactions between all involved actors. To develop the governance models, frameworks and related policies from other longitudinal cohort studies and biobanks were used, as were guidance documents on biobank and database governance and relevant literature on data and biobank governance. Results: This article outlines the key components of the governance model for the HeLTI Consortium, including management of the cohorts' respective databases and biobanks, access to data and biosamples, and considerations related to intellectual property and publications. Conclusion: Governance within international collaborative research ventures is critical to ensure the operations and benefits of these types of research apparatuses. Although this article focuses on the HeLTI Consortium as a model, it may nonetheless serve as a model for both current and future collaborative consortium-based research initiatives. Clinical Trial Registration Numbers: Canada, ISRCTN13308752; China, ChiCTR1800017773; India, ISRCTN20161479; South Africa, PACTR201903750173871.

Prepregnancy BMI, gestational weight gain, and susceptibility to autism-related traits: the EARLI and HOME studies

OBJECTIVE

Excessive gestational weight gain (GWG) has been associated with autism spectrum disorder (ASD). This study sought to examine whether familial susceptibility for autism, intensity of ASD-related behaviors, or prepregnancy BMI influences the association of GWG with ASD-related behaviors.

METHODS

Using data from the Early Autism Risk Longitudinal Investigation (EARLI) study (n = 136), a familial enriched cohort of mothers who had a previous child with ASD, and the Health Outcomes and Measures of the Environment (HOME) study (n = 253), a general population cohort, gestational age and prepregnancy BMI category-specific GWG z scores were calculated. Caregivers completed the Social Responsiveness Scale (SRS) to assess the presence and severity of ASD-related traits in children aged 3 to 8 years. Using quantile regression, the association between GWG z scores and ASD-related behaviors in children was estimated.

RESULTS

In HOME, among mothers who had overweight or obesity prepregnancy BMI values, GWG z scores and SRS scores were positively associated in children with more ASD-related traits (higher SRS scores), but not in children with fewer ASD-related traits. Similar patterns were observed in EARLI among mothers with prepregnancy obesity.

CONCLUSIONS

GWG may be associated with autism-related behaviors among children who have a greater predisposition to these behaviors and who have mothers with prepregnancy overweight or obesity.

How does arthritis affect employment? Longitudinal evidence on 18,000 British adults with arthritis compared to matched controls

INTRODUCTION

One in ten working age people in the UK live with arthritis or a similar condition affecting their joints. This impacts their quality of life, including through their work. But little is known about how arthritis affects labour market outcomes and the types of people most likely to be affected.

METHODS

Data from three population-representative household panel surveys (BHPS, ELSA, UKHLS) collected in 2001-2019 was harmonised. Propensity score matching was used to match 18,014 UK adults aged 18-80 who have arthritis with comparable adults without arthritis. The relationship between arthritis and employment, and earnings and work hours conditional on employment, were assessed using multilevel regression modelling. Heterogeneity in these relationships were assessed by age, gender, degree-level education status, NS-SEC job classification and employer type.

RESULTS

On average, arthritis was associated with a 3 percentage point reduction in the probability of employment. The effect size varied over people's life course and was larger amongst females, people without a degree, and those in routine or intermediate occupations (when compared to those in professional occupations) or working for small private companies (when compared to large private companies and non-private employers). Our models predict, for instance, that arthritis is associated with an 11 percentage point reduction in the probability of employment among 50-year-old women without a degree. This contrasts with a 5 percentage point reduction among 50-year-old men without a degree. If employed, men with a degree earned less if they had arthritis, whereas others (including women with a degree and men without a degree) had similar earnings regardless of their arthritis status. Those in professional occupations with arthritis also earnt less, especially if they were women aged over 40, with indications that this was driven by reduced work hours.

CONCLUSION

Policy interventions to support people with arthritis who wish to remain in work might be designed with people in routine work in mind, and targeted at those working in smaller private firms. More research on the cost-effectiveness of those interventions is needed.

Harmonizing Ethno-Regionally Diverse Datasets to Advance the Global Epidemiology of Dementia

Understanding dementia and cognitive impairment is a global effort needing data from multiple sources across diverse ethno-regional groups. Methodological heterogeneity means that these data often require harmonization to make them comparable before analysis. We discuss the benefits and challenges of harmonization, both retrospective and prospective, broadly and with a focus on data types that require particular sorts of approaches, including neuropsychological test scores and neuroimaging data. Throughout our discussion, we illustrate general principles and give examples of specific approaches in the context of contemporary research in dementia and cognitive impairment from around the world.

A Tale of Two Centers: Visual Exploration of Health Disparities in Cancer Care

The annual incidence of head and neck cancers (HNC) worldwide is more than 550,000 cases, with around 300,000 deaths each year. However, the incidence rates and disease-characteristics of HNC differ between treatment centers and different populations, due to undetermined reasons, which may or not include socioeconomic factors. The multi-faceted and multi-variate nature of the data in the context of the emerging field of health disparities research makes automated analysis impractical. Hence, we present a visual analysis approach to explore the health disparities in the data of HNC patients from two different cohorts at two cancer care centers. Our approach integrates data from multiple sources, including census data and city data, with custom visual encodings and with a nearest neighbor approach. Our design, created in collaboration with oncology experts, makes it possible to analyze the patients' demographic, disease characteristics, treatments and outcomes, and to make significant comparisons of these two cohorts and of individual patients. We evaluate this approach through two case studies performed with domain experts. The results demonstrate that this visual analysis approach successfully accomplishes the goal of comparing two cohorts in terms of different significant factors, and can provide insights into the main source of health disparities between the two centers.

A review of harmonization methods for studying dietary patterns

Data harmonization is the process by which each of the variables from different research studies are standardized to similar units resulting in comparable datasets. These data may be integrated for more powerful and accurate examination and prediction of outcomes for use in the intelligent and smart electronic health software programs and systems. Prospective harmonization is performed when researchers create guidelines for gathering and managing the data before data collection begins. In contrast, retrospective harmonization is performed by pooling previously collected data from various studies using expert domain knowledge to identify and translate variables. In nutritional epidemiology, dietary data harmonization is often necessary to construct the nutrient and food databases necessary to answer complex research questions and develop effective public health policy. In this paper, we review methods for effective data harmonization, including developing a harmonization plan, which common standards already exist for harmonization, and defining variables needed to harmonize datasets. Currently, several large-scale studies maintain harmonized nutrient databases, especially in Europe, and steps have been proposed to inform the retrospective harmonization process. As an example, data harmonization methods are applied to several U.S longitudinal diet datasets. Based on our review, considerations for future dietary data harmonization include user agreements for sharing private data among participating studies, defining variables and data dictionaries that accurately map variables among studies, and the use of secure data storage servers to maintain privacy. These considerations establish necessary components of harmonized data for smart health applications which can promote healthier eating and provide greater insights into the effect of dietary patterns on health.

Considerations for an integrated population health databank in Africa: lessons from global best practices

Background: The rising digitisation and proliferation of data sources and repositories cannot be ignored. This trend expands opportunities to integrate and share population health data. Such platforms have many benefits, including the potential to efficiently translate information arising from such data to evidence needed to address complex global health challenges. There are pockets of quality data on the continent that may benefit from greater integration. Integration of data sources is however under-explored in Africa. The aim of this article is to identify the requirements and provide practical recommendations for developing a multi-consortia public and population health data-sharing framework for Africa.

Methods: We conducted a narrative review of global best practices and policies on data sharing and its optimisation. We searched eight databases for publications and undertook an iterative snowballing search of articles cited in the identified publications. The Leximancer software © enabled content analysis and selection of a sample of the most relevant articles for detailed review. Themes were developed through immersion in the extracts of selected articles using inductive thematic analysis. We also performed interviews with public and population health stakeholders in Africa to gather their experiences, perceptions, and expectations of data sharing.

Results: Our findings described global stakeholder experiences on research data sharing. We identified some challenges and measures to harness available resources and incentivise data sharing.  We further highlight progress made by the different groups in Africa and identified the infrastructural requirements and considerations when implementing data sharing platforms. Furthermore, the review suggests key reforms required, particularly in the areas of consenting, privacy protection, data ownership, governance, and data access.

Conclusions: The findings underscore the critical role of inclusion, social justice, public good, data security, accountability, legislation, reciprocity, and mutual respect in developing a responsive, ethical, durable, and integrated research data sharing ecosystem.

Harmonized Phenotypes for Anxiety, Depression, and Attention-Deficit Hyperactivity Disorder (ADHD)

In multi-cohort consortia, the problem often arises that a phenotype is measured using different questionnaires. This study aimed to harmonize scores based on the Child Behaviour Check List (CBCL) and the Strength and Difficulties Questionnaire (SDQ) for anxiety/depression and ADHD. To link the scales, we used parent reports on 1330 children aged 10–11.5 years from the Raine study on both SDQ and CBCL. Harmonization was done based on Item Response Theory. We started from existing CBCL and SDQ scales related to anxiety/depression and ADHD (theoretical approach). Next, we conducted a data-driven approach using factor analysis to validate the theoretical approach. Both approaches yielded similar scales, validating the combination of existing scales. In addition, we studied the impact of harmonized (IRT-based) scores on the statistical power of the results in meta-analytic gene-finding studies. The results showed that the IRT-based harmonized scores increased the statistical power of the results compared to sum scores, even with an equal sample size. These findings can help future researchers to harmonize data from different samples and/or different questionnaires that measure anxiety, depression, and ADHD, in order to obtain the larger sample sizes, to compare research results across subpopulations or to increase generalizability, the validity or statistical power of research results. We recommend using our item parameters to estimate harmonized scores that represent commensurate phenotypes across cohorts, and we explained in detail how other researchers can use our results to harmonize data in their studies.

Development of a data classification system for preterm birth cohort studies: the RECAP Preterm project

Background

The small sample sizes available within many very preterm (VPT) longitudinal birth cohort studies mean that it is often necessary to combine and harmonise data from individual studies to increase statistical power, especially for studying rare outcomes. Curating and mapping data is a vital first step in the process of data harmonisation. To facilitate data mapping and harmonisation across VPT birth cohort studies, we developed a custom classification system as part of the Research on European Children and Adults born Preterm (RECAP Preterm) project in order to increase the scope and generalisability of research and the evaluation of outcomes across the lifespan for individuals born VPT.

Methods

The multidisciplinary consortium of expert clinicians and researchers who made up the RECAP Preterm project participated in a four-phase consultation process via email questionnaire to develop a topic-specific classification system. Descriptive analyses were calculated after each questionnaire round to provide pre- and post- ratings to assess levels of agreement with the classification system as it developed. Amendments and refinements were made to the classification system after each round.

Results

Expert input from 23 clinicians and researchers from the RECAP Preterm project aided development of the classification system’s topic content, refining it from 10 modules, 48 themes and 197 domains to 14 modules, 93 themes and 345 domains. Supplementary classifications for target, source, mode and instrument were also developed to capture additional variable-level information. Over 22,000 individual data variables relating to VPT birth outcomes have been mapped to the classification system to date to facilitate data harmonisation. This will continue to increase as retrospective data items are mapped and harmonised variables are created.

Conclusions

This bespoke preterm birth classification system is a fundamental component of the RECAP Preterm project’s web-based interactive platform. It is freely available for use worldwide by those interested in research into the long term impact of VPT birth. It can also be used to inform the development of future cohort studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01494-5.

A Secure and Reusable Software Architecture for Supporting Online Data Harmonization

Retrospective data harmonization across multiple research cohorts and studies is frequently done to increase statistical power, provide comparison analysis, and create a richer data source for data mining. However, when combining disparate data sources, harmonization projects face data management and analysis challenges. These include differences in the data dictionaries and variable definitions, privacy concerns surrounding health data representing sensitive populations, and lack of properly defined data models. With the availability of mature open-source web-based database technologies, developing a complete software architecture to overcome the challenges associated with the harmonization process can alleviate many roadblocks. By leveraging state-of-the-art software engineering and database principles, we can ensure data quality and enable cross-center online access and collaboration. This paper outlines a complete software architecture developed and customized using the Django web framework, leveraged to harmonize sensitive data collected from three NIH-support birth cohorts. We describe our framework and show how we successfully overcame challenges faced when harmonizing data from these cohorts. We discuss our efforts in data cleaning, data sharing, data transformation, data visualization, and analytics, while reflecting on what we have learned to date from these harmonized datasets.

Asselbergs F, Bernier A, Costas R, Friedrich MG, Gelpi JL, Jarvelin MR, Kuulasmaa K, Lekadir K, Mayrhofer MT, Papez V, Pasterkamp G, Petersen SE, Schmidt CO, Schulz-Menger J, Söderberg S, Shabani M, Veronesi G, Viezzer DS, Borry P. An agenda-setting paper on data sharing platforms: euCanSHare workshop

Various data sharing platforms are being developed to enhance the sharing of cohort data by addressing the fragmented state of data storage and access systems. However, policy challenges in several domains remain unresolved. The euCanSHare workshop was organized to identify and discuss these challenges and to set the future research agenda. Concerns over the multiplicity and long-term sustainability of platforms, lack of resources, access of commercial parties to medical data, credit and recognition mechanisms in academia and the organization of data access committees are outlined. Within these areas, solutions need to be devised to ensure an optimal functioning of platforms.

Cohort description: Measures of early-life behaviour and later psychopathology in the LifeCycle Project - EU Child Cohort Network

BACKGROUND

The EU LifeCycle Project was launched in 2017 to combine, harmonise, and analyse data from more than 250,000 participants across Europe and Australia, involving cohorts participating in the EU-funded LifeCycle Project. The purpose of this cohort description is to provide a detailed overview over the major measures within mental health domains that are available in 17 European and Australian cohorts participating in the LifeCycle Project.

METHODS

Data on cognitive, behavioural and psychological development has been collected on participants from birth until adulthood through questionnaire and medical data. We developed an inventory of the available data by mapping individual instruments, domain types, and age groups, providing the basis for statistical harmonization across mental health measures.

RESULTS

The mental health data in LifeCycle contain longitudinal and cross-sectional data for ages 0-18+ years, covering domains across a wide range of behavioural and psychopathology indicators and outcomes (including executive function, depression, ADHD and cognition). These data span a unique combination of qualitative data collected through behavioural/cognitive/mental health questionnaires and examination, as well as data from biological samples and indices in the form of brain imaging (MRI, foetal ultrasound) and DNA methylation data. Harmonized variables on a subset of mental health domains have been developed, providing statistical equivalence of measures required for longitudinal meta-analyses across instruments and cohorts.

CONCLUSION

Mental health data harmonized through the LifeCycle project can be used to study life course trajectories and exposure-outcome models that examine early life risk factors for mental illness and develop predictive markers for later-life disease.

The PhenX Toolkit: Establishing Standard Measures for COVID-19 Research

The PhenX (consensus measures for Phenotypes and eXposures) Toolkit (https://www.phenxtoolkit.org/) is a publicly available, web‐based catalog of recommended, well‐established measurement protocols of phenotypes and exposures. The goal of PhenX is to facilitate the use of standard measures, enhance data interoperability, and promote collaborative and translational research. PhenX is driven by the scientific community and historically has depended on working groups of experts to recommend measures for release in the PhenX Toolkit. The urgent need for recommended, standard measures for COVID‐19 research triggered the development of a “rapid release” process for releasing new content in the PhenX Toolkit. Initially, PhenX collaborated with the National Institutes of Health (NIH) Office of Behavioral and Social Sciences Research, the National Human Genome Research Institute, and the NIH Disaster Research Response (DR2) program to create a library of COVID‐19 measurement protocols. With additional support from NIH, PhenX adapted crowdsourcing techniques to accelerate prioritization and recommendation of protocols for release in the PhenX Toolkit. Prioritized COVID‐19‐specific protocols were used to anchor and define specialty collections of protocols that were subject to review and approval by the PhenX Steering Committee. In addition to the COVID‐19‐specific protocols, the specialty collections include existing, well‐established PhenX protocols, use of which will further enhance data interoperability and cross‐study analysis. The COVID‐19 specialty collections are Behaviors and Risks; Ethnicity, Race and Demographics; History, Treatment and Outcomes; Information Resources; Psychosocial and Mental Health; and Socioeconomic. The development and usage of PhenX COVID‐19 specialty collections are described in this article. © 2021 The Authors.

Basic Protocol: Selecting COVID‐19 protocols

Facilitating harmonized data quality assessments. A data quality framework for observational health research data collections with software implementations in R

Background

No standards exist for the handling and reporting of data quality in health research. This work introduces a data quality framework for observational health research data collections with supporting software implementations to facilitate harmonized data quality assessments.

Methods

Developments were guided by the evaluation of an existing data quality framework and literature reviews. Functions for the computation of data quality indicators were written in R. The concept and implementations are illustrated based on data from the population-based Study of Health in Pomerania (SHIP).

Results

The data quality framework comprises 34 data quality indicators. These target four aspects of data quality: compliance with pre-specified structural and technical requirements (integrity); presence of data values (completeness); inadmissible or uncertain data values and contradictions (consistency); unexpected distributions and associations (accuracy). R functions calculate data quality metrics based on the provided study data and metadata and R Markdown reports are generated. Guidance on the concept and tools is available through a dedicated website.

Conclusions

The presented data quality framework is the first of its kind for observational health research data collections that links a formal concept to implementations in R. The framework and tools facilitate harmonized data quality assessments in pursue of transparent and reproducible research. Application scenarios comprise data quality monitoring while a study is carried out as well as performing an initial data analysis before starting substantive scientific analyses but the developments are also of relevance beyond research.

A systematic review of asthma case definitions in 67 birth cohort studies

BACKGROUND

Birth cohort studies are a valuable source of information about potential risk factors for childhood asthma. To better understand similarities and variations in findings between birth cohort studies, the methodologies used to measure asthma require consideration.

OBJECTIVE

To review and appraise the definitions of "asthma" used in birth cohort studies.

METHODS

A literature search, conducted in December 2017 in the MEDLINE database and birth cohort repositories, identified 1721 citations published since 1990. Information extracted included: study name, year of publication, sample size, sample age, prevalence of asthma (%), study region, source of information about asthma, measured outcome, and asthma case definition. A meta-analysis evaluated whether asthma prevalence in cohorts from Europe and North America varied by the studies' definition of asthma and by their data sources.

RESULTS

The final review included 67 birth cohorts, of which 48 (72%) were from Europe, 14 (21%) from North America, 3 (5%) from Oceania, 1 (1%) from Asia and 1 (1%) from South America. We identified three measured outcomes: "asthma ever", "current asthma", and "asthma" without further specification. Definitions of "asthma ever" were primarily based upon an affirmative parental response to the question whether the child had ever been diagnosed with asthma by a physician. The most frequently used definition of "current asthma" was "asthma ever" and either asthma symptoms or asthma medications in the last 12 months. This definition of "current asthma" was used in 16 cohorts. There was no statistically significant difference in the pooled asthma prevalence in European and North American cohorts that used questionnaire alone versus other data sources to classify asthma.

CONCLUSION

There is substantial heterogeneity in childhood asthma definitions in birth cohort studies. Standardisation of asthma case definitions will improve the comparability and utility of future cohort studies and enable meta-analyses.

Gestational Exposure to Phthalates and Social Responsiveness Scores in Children Using Quantile Regression: The EARLI and HOME Studies

Linear regression is often used to estimate associations between chemical exposures and neurodevelopment at the mean of the outcome. However, the potential effect of chemicals may be greater among individuals at the 'tails' of outcome distributions. Here, we investigated distributional effects on the associations between gestational phthalate exposure and child Autism Spectrum Disorder (ASD)-related behaviors using quantile regression. We harmonized data from the Early Autism Risk Longitudinal Investigation (EARLI) (n = 140) Study, an enriched-risk cohort of mothers who had a child with ASD, and the Health Outcomes and Measures of the Environment (HOME) Study (n = 276), a general population cohort. We measured concentrations of 9 phthalate metabolites in urine samples collected twice during pregnancy. Caregivers reported children's ASD-related behaviors using the Social Responsiveness Scale (SRS) at age 3-8 years; higher scores indicate more ASD-related behaviors. In EARLI, associations between phthalate concentrations and SRS scores were predominately inverse or null across SRS score quantiles. In HOME, positive associations of mono-n-butyl phthalate, monobenzyl phthalate, mono-isobutyl phthalate, and di-2-ethylhexyl phthalate concentrations with SRS scores increased in strength from the median to 95th percentile of SRS scores. These results suggest associations between phthalate concentrations and SRS scores may be stronger in individuals with higher SRS scores.

Association between self-reported caffeine intake during pregnancy and social responsiveness scores in childhood: The EARLI and HOME studies

Maternal nutrition during gestation has been investigated for its role in child neurodevelopment. However, little is known about the potential impact of gestational caffeine exposure on child autistic behaviors. Here, we assess the relation between maternal caffeine intake during pregnancy and children's behavioral traits related to Autism Spectrum Disorder (ASD). We harmonized data from two pregnancy cohorts, Early Autism Risk Longitudinal Investigation (EARLI) (n = 120), an enriched-risk cohort of mothers who previously had a child with ASD, from Pennsylvania, Maryland, and Northern California (2009-2012), and the Health Outcomes and Measures of the Environment (HOME) Study (n = 269), a general population cohort from Cincinnati, Ohio (2003-2006). Mothers self-reported caffeine intake twice during pregnancy. Caregivers reported child behavioral traits related to ASD using the Social Responsiveness Scale (SRS) when children were aged 3-8 years. Higher scores indicate more ASD-related behaviors. We estimated covariate-adjusted differences in continuous SRS T-scores per interquartile range increase in caffeine intake. Self-reported caffeine intake during pregnancy was positively associated with SRS T-scores among children in EARLI (β: 2.0; 95% CI -0.1, 4.0), but to a lesser extent in HOME (β: 0.6; 95% CI -0.5, 1.6). In HOME, pre-pregnancy body mass index (BMI) modified the association between caffeine intake and SRS T-scores, where more positive associations were observed among women with higher BMIs. Our findings suggest gestational caffeine intake may represent a marker of vulnerability to childhood ASD-related behaviors. Additional studies are warranted to extend these findings.

Overview of retrospective data harmonisation in the MINDMAP project: process and results

Background

The MINDMAP project implemented a multinational data infrastructure to investigate the direct and interactive effects of urban environments and individual determinants of mental well-being and cognitive function in ageing populations. Using a rigorous process involving multiple teams of experts, longitudinal data from six cohort studies were harmonised to serve MINDMAP objectives. This article documents the retrospective data harmonisation process achieved based on the Maelstrom Research approach and provides a descriptive analysis of the harmonised data generated.

Methods

A list of core variables (the DataSchema) to be generated across cohorts was first defined, and the potential for cohort-specific data sets to generate the DataSchema variables was assessed. Where relevant, algorithms were developed to process cohort-specific data into DataSchema format, and information to be provided to data users was documented. Procedures and harmonisation decisions were thoroughly documented.

Results

The MINDMAP DataSchema (v2.0, April 2020) comprised a total of 2841 variables (993 on individual determinants and outcomes, 1848 on environmental exposures) distributed across up to seven data collection events. The harmonised data set included 220 621 participants from six cohorts (10 subpopulations). Harmonisation potential, participant distributions and missing values varied across data sets and variable domains.

Conclusion

The MINDMAP project implemented a collaborative and transparent process to generate a rich integrated data set for research in ageing, mental well-being and the urban environment. The harmonised data set supports a range of research activities and will continue to be updated to serve ongoing and future MINDMAP research needs.

A framework for assessing the impact of chemical exposures on neurodevelopment in ECHO: Opportunities and challenges

The Environmental influences on Child Health Outcomes (ECHO) Program is a research initiative funded by the National Institutes of Health that capitalizes on existing cohort studies to investigate the impact of early life environmental factors on child health and development from infancy through adolescence. In the initial stage of the program, extant data from 70 existing cohort studies are being uploaded to a database that will be publicly available to researchers. This new database will represent an unprecedented opportunity for researchers to combine data across existing cohorts to address associations between prenatal chemical exposures and child neurodevelopment. Data elements collected by ECHO cohorts were determined via a series of surveys administered by the ECHO Data Analysis Center. The most common chemical classes quantified in multiple cohorts include organophosphate pesticides, polychlorinated biphenyls, polybrominated diphenyl ethers, environmental phenols (including bisphenol A), phthalates, and metals. For each of these chemicals, at least four ECHO cohorts also collected behavioral data during infancy/early childhood using the Child Behavior Checklist. For these chemicals and this neurodevelopmental assessment (as an example), existing data from multiple ECHO cohorts could be pooled to address research questions requiring larger sample sizes than previously available. In addition to summarizing the data that will be available, the article also describes some of the challenges inherent in combining existing data across cohorts, as well as the gaps that could be filled by the additional data collection in the ECHO Program going forward.

The challenges in data integration - heterogeneity and complexity in clinical trials and patient registries of Systemic Lupus Erythematosus

Background

Individual clinical trials and cohort studies are a useful source of data, often under-utilised once a study has ended. Pooling data from multiple sources could increase sample sizes and allow for further investigation of treatment effects; even if the original trial did not meet its primary goals. Through the MASTERPLANS (MAximizing Sle ThERapeutic PotentiaL by Application of Novel and Stratified approaches) national consortium, focused on Systemic Lupus Erythematosus (SLE), we have gained valuable real-world experiences in aligning, harmonising and combining data from multiple studies and trials, specifically where standards for data capture, representation and documentation, were not used or were unavailable. This was not without challenges arising both from the inherent complexity of the disease and from differences in the way data were captured and represented across different studies.

Main body

Data were, unavoidably, aligned by hand, matching up equivalent or similar patient variables across the different studies. Heterogeneity-related issues were tackled and data were cleaned, organised and combined, resulting in a single large dataset ready for analysis. Overcoming these hurdles, often seen in large-scale data harmonization and integration endeavours of legacy datasets, was made possible within a realistic timescale and limited resource by focusing on specific research questions driven by the aims of MASTERPLANS. Here we describe our experiences tackling the complexities in the integration of large, diverse datasets, and the lessons learned.

Conclusions

Harmonising data across studies can be complex, and time and resource consuming. The work carried out here highlights the importance of using standards for data capture, recording, and representation, to facilitate both the integration of large datasets and comparison between studies. Where standards are not implemented at the source harmonisation is still possible by taking a flexible approach, with systematic preparation, and a focus on specific research questions.

Extending the Minimum Information About BIobank Data Sharing Terminology to Describe Samples, Sample Donors, and Events

Introduction: The Minimum Information About BIobank data Sharing (MIABIS) was initiated in 2012. MIABIS aims to create a common biobank terminology to facilitate data sharing in biobanks and sample collections. The MIABIS Core terminology consists of three components describing biobanks, sample collections, and studies, in which information on samples and sample donors is provided at aggregated form. However, there is also a need to describe samples and sample donors at an individual level to allow more elaborate queries on available biobank samples and data. Therefore the MIABIS terminology has now been extended with components describing samples and sample donors at an individual level. Materials and Methods: The components were defined according to specific scope and use cases by a large group of experts, and through several cycles of reviews, according to the new MIABIS governance model of BBMRI-ERIC (Biobanking and Biomolecular Resources Research Infrastructure-European Research Infrastructure Consortium). The guiding principles applied in developing these components included the following terms: model should consider only samples of human origin, model should be applicable to all types of samples and all sample donors, and model should describe the current status of samples stored in a given biobank. Results: A minimal set of standard attributes for defining samples and sample donors is presented here. We added an "event" component to describe attributes that are not directly describing samples or sample donors but are tightly related to them. To better utilize the generic data model, we suggest a procedure by which interoperability can be promoted, using specific MIABIS profiles. Discussion: The MIABIS sample and donor component extensions and the new generic data model complement the existing MIABIS Core 2.0 components, and substantially increase the potential usability of this terminology for better describing biobank samples and sample donors. They also support the use of individual level data about samples and sample donors to obtain accurate and detailed biobank availability queries.

Li-Fraumeni Exploration Consortium Data Coordinating Center: Building an Interactive Web-Based Resource for Collaborative International Cancer Epidemiology Research for a Rare Condition

BACKGROUND

The success of multisite collaborative research relies on effective data collection, harmonization, and aggregation strategies. Data Coordination Centers (DCC) serve to facilitate the implementation of these strategies. The utility of a DCC can be particularly relevant for research on rare diseases where collaboration from multiple sites to amass large aggregate datasets is essential. However, approaches to building a DCC have been scarcely documented.

METHODS

The Li-Fraumeni Exploration (LiFE) Consortium's DCC was created using multiple open source packages, including LAM/G Application (Linux, Apache, MySQL, Grails), Extraction-Transformation-Loading (ETL) Pentaho Data Integration Tool, and the Saiku-Mondrian client. This document serves as a resource for building a rare disease DCC for multi-institutional collaborative research.

RESULTS

The primary scientific and technological objective to create an online central repository into which data from all participating sites could be deposited, harmonized, aggregated, disseminated, and analyzed was completed. The cohort now include 2,193 participants from six contributing sites, including 1,354 individuals from families with a pathogenic or likely variant in TP53. Data on cancer diagnoses are also available. Challenges and lessons learned are summarized.

CONCLUSIONS

The methods leveraged mitigate challenges associated with successfully developing a DCC's technical infrastructure, data harmonization efforts, communications, and software development and applications.

IMPACT

These methods can serve as a framework in establishing other collaborative research efforts. Data from the consortium will serve as a great resource for collaborative research to improve knowledge on, and the ability to care for, individuals and families with Li-Fraumeni syndrome.

Overcoming the Barriers That Obscure the Interlinking and Analysis of Clinical Data Through Harmonization and Incremental Learning

Goal: To present a framework for data sharing, curation, harmonization and federated data analytics to solve open issues in healthcare, such as, the development of robust disease prediction models. Methods: Data curation is applied to remove data inconsistencies. Lexical and semantic matching methods are used to align the structure of the heterogeneous, curated cohort data along with incremental learning algorithms including class imbalance handling and hyperparameter optimization to enable the development of disease prediction models. Results: The applicability of the framework is demonstrated in a case study of primary Sjögren's Syndrome, yielding harmonized data with increased quality and more than 85% agreement, along with lymphoma prediction models with more than 80% sensitivity and specificity. Conclusions: The framework provides data quality, harmonization and analytics workflows that can enhance the statistical power of heterogeneous clinical data and enables the development of robust models for disease prediction.

Latent variable models for harmonization of test scores: A case study on memory

Combining data from different studies has a long tradition within the scientific community. It requires that the same information is collected from each study to be able to pool individual data. When studies have implemented different methods or used different instruments (e.g., questionnaires) for measuring the same characteristics or constructs, the observed variables need to be harmonized in some way to obtain equivalent content information across studies. This paper formulates the main concepts for harmonizing test scores from different observational studies in terms of latent variable models. The concepts are formulated in terms of calibration, invariance, and exchangeability. Although similar ideas are present in measurement reliability and test equating, harmonization is different from measurement invariance and generalizes test equating. In addition, if a test score needs to be transformed to another test score, harmonization of variables is only possible under specific conditions. Observed test scores that connect all of the different studies, are necessary to be able to test the underlying assumptions of harmonization. The concepts of harmonization are illustrated on multiple memory test scores from three different Canadian studies.

KETOS: Clinical decision support and machine learning as a service - A training and deployment platform based on Docker, OMOP-CDM, and FHIR Web Services

Background and objective

To take full advantage of decision support, machine learning, and patient-level prediction models, it is important that models are not only created, but also deployed in a clinical setting. The KETOS platform demonstrated in this work implements a tool for researchers allowing them to perform statistical analyses and deploy resulting models in a secure environment.

Methods

The proposed system uses Docker virtualization to provide researchers with reproducible data analysis and development environments, accessible via Jupyter Notebook, to perform statistical analysis and develop, train and deploy models based on standardized input data. The platform is built in a modular fashion and interfaces with web services using the Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR) standard to access patient data. In our prototypical implementation we use an OMOP common data model (OMOP-CDM) database. The architecture supports the entire research lifecycle from creating a data analysis environment, retrieving data, and training to final deployment in a hospital setting.

Results

We evaluated the platform by establishing and deploying an analysis and end user application for hemoglobin reference intervals within the University Hospital Erlangen. To demonstrate the potential of the system to deploy arbitrary models, we loaded a colorectal cancer dataset into an OMOP database and built machine learning models to predict patient outcomes and made them available via a web service. We demonstrated both the integration with FHIR as well as an example end user application. Finally, we integrated the platform with the open source DataSHIELD architecture to allow for distributed privacy preserving data analysis and training across networks of hospitals.

Conclusion

The KETOS platform takes a novel approach to data analysis, training and deploying decision support models in a hospital or healthcare setting. It does so in a secure and privacy-preserving manner, combining the flexibility of Docker virtualization with the advantages of standardized vocabularies, a widely applied database schema (OMOP-CDM), and a standardized way to exchange medical data (FHIR).

Sex-specific incidence of asthma, rhinitis and respiratory multimorbidity before and after puberty onset: individual participant meta-analysis of five birth cohorts collaborating in MeDALL

Introduction

To understand the puberty-related sex shift in the prevalence of asthma and rhinitis as single entities and as respiratory multimorbidities, we investigated if there is also a sex-specific and puberty-related pattern of their incidences.

Methods

We used harmonised questionnaire data from 18 451 participants in five prospective observational European birth cohorts within the collaborative MeDALL (Mechanisms of the Development of Allergy) project. Outcome definitions for IgE-associated and non-IgE-associated asthma, rhinitis and respiratory multimorbidity (first occurrence of coexisting asthma and rhinitis) were based on questionnaires and the presence of specific antibodies (IgE) against common allergens in serum. For each outcome, we used proportional hazard models with sex–puberty interaction terms and conducted a one-stage individual participant data meta-analysis.

Results

Girls had a lower risk of incident asthma (adjusted HR 0.67, 95% CI 0.61 to 0.74), rhinitis (0.73, 0.69 to 0.78) and respiratory multimorbidity (0.58, 0.51 to 0.66) before puberty compared with boys. After puberty onset, these incidences became more balanced across the sexes (asthma 0.84, 0.64 to 1.10; rhinitis 0.90, 0.80 to 1.02; respiratory multimorbidity 0.84, 0.63 to 1.13). The incidence sex shift was slightly more distinct for non-IgE-associated respiratory diseases (asthma 0.74, 0.63 to 0.87 before vs 1.23, 0.75 to 2.00 after puberty onset; rhinitis 0.88, 0.79 to 0.98 vs 1.20, 0.98 to 1.47; respiratory multimorbidity 0.66, 0.49 to 0.88 vs 0.96, 0.54 to 1.71) than for IgE-associated respiratory diseases.

Discussion

We found an incidence ‘sex shift’ in chronic respiratory diseases from a male predominance before puberty to a more sex-balanced incidence after puberty onset, which may partly explain the previously reported sex shift in prevalence. These differences need to be considered in public health to enable effective diagnoses and timely treatment in adolescent girls.

Harmonization of the Health and Risk Factor Questionnaire data of the Canadian Partnership for Tomorrow Project: a descriptive analysis

BACKGROUND

The Canadian Partnership for Tomorrow Project is a multistudy platform integrating the British Columbia Generations Project, Alberta's Tomorrow Project, the Ontario Health Study, CARTaGENE (Quebec) and the Atlantic Partnership for Tomorrow's Health. This paper describes the process used to harmonize the Health and Risk Factor Questionnaire data and provides an overview of the key information required to properly use the core data set generated.

METHODS

This is a descriptive analysis of the harmonization process that was developed on the basis of the Maelstrom Research guidelines for retrospective harmonization. Core variables (DataSchema) to be generated across cohorts were defined and the potential for cohort-specific data sets to generate the DataSchema variables was assessed. Where relevant, algorithms were developed and applied to process cohort-specific data into the DataSchema format, and information to be provided to data users was documented.

RESULTS

The Health and Risk Factor Questionnaire DataSchema (version 2.0, October 2017) comprised 694 variables. The assessment of harmonization potential for the variables over 12 cohort-specific data sets resulted in 6799 (81.6%) of the variables being considered as harmonizable. A total of 307 017 participants were included in the harmonized data set. Through the cohort data portal, researchers can find information about the definitions of variables, harmonization potential, algorithms applied to generate harmonized variables and participant distributions.

INTERPRETATION

The harmonization process enabled the creation of a unique data set including data on health and risk factors from over 307 000 Canadians. These data, in combination with complementary data sets, can be used to investigate the impact of biological, environmental and behavioural factors on cancer and chronic diseases.

Collaborative, pooled and harmonized study designs for epidemiologic research: challenges and opportunities

Collaborative study designs (CSDs) that combine individual-level data from multiple independent contributing studies (ICSs) are becoming much more common due to their many advantages: increased statistical power through large sample sizes; increased ability to investigate effect heterogeneity due to diversity of participants; cost-efficiency through capitalizing on existing data; and ability to foster cooperative research and training of junior investigators. CSDs also present surmountable political, logistical and methodological challenges. Data harmonization may result in a reduced set of common data elements, but opportunities exist to leverage heterogeneous data across ICSs to investigate measurement error and residual confounding. Combining data from different study designs is an art, which motivates methods development. Diverse study samples, both across and within ICSs, prompt questions about the generalizability of results from CSDs. However, CSDs present unique opportunities to describe population health across person, place and time in a consistent fashion, and to explicitly generalize results to target populations of public health interest. Additional analytic challenges exist when analysing CSD data, because mechanisms by which systematic biases (e.g. information bias, confounding bias) arise may vary across ICSs, but multidisciplinary research teams are ready to tackle these challenges. CSDs are a powerful tool that, when properly harnessed, permits research that was not previously possible.

Integrating Clinical and Epidemiologic Data on Allergic Diseases Across Birth Cohorts: A Harmonization Study in the Mechanisms of the Development of Allergy Project

The numbers of international collaborations among birth cohort studies designed to better understand asthma and allergies have increased in the last several years. However, differences in definitions and methods preclude direct pooling of original data on individual participants. As part of the Mechanisms of the Development of Allergy (MeDALL) Project, we harmonized data from 14 birth cohort studies (each with 3-20 follow-up periods) carried out in 9 European countries during 1990-1998 or 2003-2009. The harmonization process followed 6 steps: 1) organization of the harmonization panel; 2) identification of variables relevant to MeDALL objectives (candidate variables); 3) proposal of a definition for each candidate variable (reference definition); 4) assessment of the compatibility of each cohort variable with its reference definition (inferential equivalence) and classification of this inferential equivalence as complete, partial, or impossible; 5) convocation of a workshop to agree on the reference definitions and classifications of inferential equivalence; and 6) preparation and delivery of data through a knowledge management portal. We agreed on 137 reference definitions. The inferential equivalence of 3,551 cohort variables to their corresponding reference definitions was classified as complete, partial, and impossible for 70%, 15%, and 15% of the variables, respectively. A harmonized database was delivered to MeDALL investigators. In asthma and allergy birth cohorts, the harmonization of data for pooled analyses is feasible, and high inferential comparability may be achieved. The MeDALL harmonization approach can be used in other collaborative projects.

Prenatal determinants of childhood obesity: a review of risk factors 1

Childhood obesity is a predictor of adult obesity and has its roots in the pre-pregnancy or pregnancy period. This review presents an overview of the prenatal risk factors for childhood obesity, which were categorized into 2 groups: biological risk factors (maternal pre-pregnancy body mass index, gestational weight gain, diabetes in pregnancy, and caesarean section), and environmental and behavioural risk factors (maternal smoking and exposure to obesogens, maternal dietary patterns, maternal intestinal microbiome and antibiotics exposure, and maternal psychosocial stress). Identifying modifiable predisposing prenatal factors for obesity will inform further development of inventions to prevent obesity over the life course, and future directions for research and intervention are discussed.

The International Childhood Cancer Cohort Consortium (I4C): A research platform of prospective cohorts for studying the aetiology of childhood cancers

BACKGROUND

Childhood cancer is a rare but leading cause of morbidity and mortality. Established risk factors, accounting for <10% of incidence, have been identified primarily from case-control studies. However, recall, selection and other potential biases impact interpretations particularly, for modest associations. A consortium of pregnancy and birth cohorts (I4C) was established to utilise prospective, pre-diagnostic exposure assessments and biological samples.

METHODS

Eligibility criteria, follow-up methods and identification of paediatric cancer cases are described for cohorts currently participating or planning future participation. Also described are exposure assessments, harmonisation methods, biological samples potentially available for I4C research, the role of the I4C data and biospecimen coordinating centres and statistical approaches used in the pooled analyses.

RESULTS

Currently, six cohorts recruited over six decades (1950s-2000s) contribute data on 388 120 mother-child pairs. Nine new cohorts from seven countries are anticipated to contribute data on 627 500 additional projected mother-child pairs within 5 years. Harmonised data currently includes over 20 "core" variables, with notable variability in mother/child characteristics within and across cohorts, reflecting in part, secular changes in pregnancy and birth characteristics over the decades.

CONCLUSIONS

The I4C is the first cohort consortium to have published findings on paediatric cancer using harmonised variables across six pregnancy/birth cohorts. Projected increases in sample size, expanding sources of exposure data (eg, linkages to environmental and administrative databases), incorporation of biological measures to clarify exposures and underlying molecular mechanisms and forthcoming joint efforts to complement case-control studies offer the potential for breakthroughs in paediatric cancer aetiologic research.

Harmonization, data management, and statistical issues related to prospective multicenter studies in Ankylosing spondylitis (AS): Experience from the Prospective Study Of Ankylosing Spondylitis (PSOAS) cohort

Ankylosing spondylitis (AS) is characterized by inflammation of the spine and sacroiliac joints causing pain and stiffness and, in some patients, ultimately new bone formation, and progressive joint ankyloses. The classical definition of AS is based on the modified New York (mNY) criteria. Limited data have been reported regarding data quality assurance procedure for multicenter or multisite prospective cohort of patients with AS. Since 2002, 1272 qualified AS patients have been enrolled from five sites (4 US sites and 1 Australian site) in the Prospective Study Of Ankylosing Spondylitis (PSOAS). In 2012, a Data Management and Statistical Core (DMSC) was added to the PSOAS team to assist in study design, establish a systematic approach to data management and data quality, and develop and apply appropriate statistical analysis of data. With assistance from the PSOAS investigators, DMSC modified Case Report Forms and developed database in Research Electronic Data Capture (REDCap). DMSC also developed additional data quality assurance procedure to assure data quality. The error rate for various forms in PSOAS databases ranged from 0.07% for medications data to 1.1% for arthritis activity questionnaire-Global pain. Furthermore, based on data from a sub study of 48 patients with AS, we showed a strong level (90.0%) of agreement between the two readers of X-rays with respect to modified Stoke Ankylosing Spondylitis Spine Score (mSASSS). This paper not only could serve as reference for future publications from PSOAS cohort but also could serve as a basic guide to ensuring data quality for multicenter clinical studies.

Rationale and design of a placebo controlled randomized trial to assess short term, high-dose oral cholecalciferol on select laboratory and genomic responses in African Americans with hypovitaminosis D

Cardiovascular Disease (CVD) and related disorders remain a leading cause of health disparities and premature death for African Americans. Hypovitaminosis D is disproportionately prevalent in African Americans and has been linked to CVD and CVD risk factors including hypertension, diabetes and obesity. Thus, hypovitaminosis D may represent a common pathway influencing CV risk factors in a select subgroup of persons. The purpose of this paper is to report the study design of a prospective eight week prospective double-blind randomized, placebo-controlled trial (n = 330 allocated 2:1 to intervention vs. control) to assess the effect of placebo vs. high-dose oral cholecalciferol (100,000 IU vitamin D3 at baseline and week 2) on 6-week change of select biologic cardiometabolic risk factors (including parathyroid hormone to assess biologic activity, pro-inflammatory/pro-thrombotic/fibrotic markers, insulin sensitivity and vitamin D metabolites) and their relationship to vitamin D administration and modification by vitamin D receptor polymorphisms in overweight, hypertensive African Americans with hypovitaminosis D. Findings from this trial will present insights into potential causal links between vitamin D repletion and mechanistic pathways of CV disease, including established and novel genomic markers.

MIRACUM: Medical Informatics in Research and Care in University Medicine

INTRODUCTION

This article is part of the Focus Theme of Methods of Information in Medicine on the German Medical Informatics Initiative. Similar to other large international data sharing networks (e.g. OHDSI, PCORnet, eMerge, RD-Connect) MIRACUM is a consortium of academic and hospital partners as well as one industrial partner in eight German cities which have joined forces to create interoperable data integration centres (DIC) and make data within those DIC available for innovative new IT solutions in patient care and medical research.

OBJECTIVES

Sharing data shall be supported by common interoperable tools and services, in order to leverage the power of such data for biomedical discovery and moving towards a learning health system. This paper aims at illustrating the major building blocks and concepts which MIRACUM will apply to achieve this goal.

GOVERNANCE AND POLICIES

Besides establishing an efficient governance structure within the MIRACUM consortium (based on the steering board, a central administrative office, the general MIRACUM assembly, six working groups and the international scientific advisory board), defining DIC governance rules and data sharing policies, as well as establishing (at each MIRACUM DIC site, but also for MIRACUM in total) use and access committees are major building blocks for the success of such an endeavor.

ARCHITECTURAL FRAMEWORK AND METHODOLOGY

The MIRACUM DIC architecture builds on a comprehensive ecosystem of reusable open source tools (MIRACOLIX), which are linkable and interoperable amongst each other, but also with the existing software environment of the MIRACUM hospitals. Efficient data protection measures, considering patient consent, data harmonization and a MIRACUM metadata repository as well as a common data model are major pillars of this framework. The methodological approach for shared data usage relies on a federated querying and analysis concept.

USE CASES

MIRACUM aims at proving the value of their DIC with three use cases: IT support for patient recruitment into clinical trials, the development and routine care implementation of a clinico-molecular predictive knowledge tool, and molecular-guided therapy recommendations in molecular tumor boards.

RESULTS

Based on the MIRACUM DIC release in the nine months conceptual phase first large scale analysis for stroke and colorectal cancer cohorts have been pursued.

DISCUSSION

Beyond all technological challenges successfully applying the MIRACUM tools for the enrichment of our knowledge about diagnostic and therapeutic concepts, thus supporting the concept of a Learning Health System will be crucial for the acceptance and sustainability in the medical community and the MIRACUM university hospitals.

A framework for measurement and harmonization of pediatric multiple sclerosis etiologic research studies: The Pediatric MS Tool-Kit

Background:

While studying the etiology of multiple sclerosis (MS) in children has several methodological advantages over studying etiology in adults, studies are limited by small sample sizes.

Objective:

Using a rigorous methodological process, we developed the Pediatric MS Tool-Kit, a measurement framework that includes a minimal set of core variables to assess etiological risk factors.

Methods:

We solicited input from the International Pediatric MS Study Group to select three risk factors: environmental tobacco smoke (ETS) exposure, sun exposure, and vitamin D intake. To develop the Tool-Kit, we used a Delphi study involving a working group of epidemiologists, neurologists, and content experts from North America and Europe.

Results:

The Tool-Kit includes six core variables to measure ETS, six to measure sun exposure, and six to measure vitamin D intake. The Tool-Kit can be accessed online (www.maelstrom-research.org/mica/network/tool-kit).

Conclusion:

The goals of the Tool-Kit are to enhance exposure measurement in newly designed pediatric MS studies and comparability of results across studies, and in the longer term to facilitate harmonization of studies, a methodological approach that can be used to circumvent issues of small sample sizes. We believe the Tool-Kit will prove to be a valuable resource to guide pediatric MS researchers in developing study-specific questionnaire

Launch of an Infrastructure for Health Research: BBMRI-ERIC

Biobanking and BioMolecular resources Research Infrastructure (BBMRI)- European Research Infrastructure Consortium (ERIC) is the largest infrastructure launched in Europe in health research. By nature it is a distributed infrastructure, in which biological samples and data are hosted by the European Member States biobanks. As of today, BBMRI-ERIC consists of 19 European Member States and 1 international organization, the International Agency for Research on Cancer. This means that BBMRI-ERIC has a population of >500 million individuals in Europe. BBMRI-ERIC is a truly Pan-European Research Infrastructure for health research. Given that BBMRI-ERIC is set up to become a key source for users in both academic and scientific institutions as well as in the pharmaceutical and life science industries, it contributes directly to the Innovation Union concept. It is pan-European because BBMRI-ERIC already shows an excellent geographic and regional coverage all over Europe involving countries from South, East, West, North, and Central Europe. BBMRI-ERIC is a service-driven infrastructure for the European Member States, driven by science. The BBMRI-ERIC Directory consists of 100 million samples and a roadmap for better-defined quality in European biobanks for improving reproducibility and reliability of the biological sample and data.