DataSHIELD: taking the analysis to the data, not the data to the analysis

Background: Research in modern biomedicine and social science requires sample sizes so large that they can often only be achieved through a pooled co-analysis of data from several studies. But the pooling of information from individuals in a central database that may be queried by researchers raises important ethico-legal questions and can be controversial. In the UK this has been highlighted by recent debate and controversy relating to the UK’s proposed ‘care.data’ initiative, and these issues reflect important societal and professional concerns about privacy, confidentiality and intellectual property. DataSHIELD provides a novel technological solution that can circumvent some of the most basic challenges in facilitating the access of researchers and other healthcare professionals to individual-level data.

Methods: Commands are sent from a central analysis computer (AC) to several data computers (DCs) storing the data to be co-analysed. The data sets are analysed simultaneously but in parallel. The separate parallelized analyses are linked by non-disclosive summary statistics and commands transmitted back and forth between the DCs and the AC. This paper describes the technical implementation of DataSHIELD using a modified R statistical environment linked to an Opal database deployed behind the computer firewall of each DC. Analysis is controlled through a standard R environment at the AC.

Results: Based on this Opal/R implementation, DataSHIELD is currently used by the Healthy Obese Project and the Environmental Core Project (BioSHaRE-EU) for the federated analysis of 10 data sets across eight European countries, and this illustrates the opportunities and challenges presented by the DataSHIELD approach.

Conclusions: DataSHIELD facilitates important research in settings where: (i) a co-analysis of individual-level data from several studies is scientifically necessary but governance restrictions prohibit the release or sharing of some of the required data, and/or render data access unacceptably slow; (ii) a research group (e.g. in a developing nation) is particularly vulnerable to loss of intellectual property—the researchers want to fully share the information held in their data with national and international collaborators, but do not wish to hand over the physical data themselves; and (iii) a data set is to be included in an individual-level co-analysis but the physical size of the data precludes direct transfer to a new site for analysis.

3D Variation in delineation of head and neck organs at risk

Background

Consistent delineation of patient anatomy becomes increasingly important with the growing use of highly conformal and adaptive radiotherapy techniques. This study investigates the magnitude and 3D localization of interobserver variability of organs at risk (OARs) in the head and neck area with application of delineation guidelines, to establish measures to reduce current redundant variability in delineation practice.

Methods

Interobserver variability among five experienced radiation oncologists was studied in a set of 12 head and neck patient CT scans for the spinal cord, parotid and submandibular glands, thyroid cartilage, and glottic larynx. For all OARs, three endpoints were calculated: the Intraclass Correlation Coefficient (ICC), the Concordance Index (CI) and a 3D measure of variation (3D SD).

Results

All endpoints showed largest interobserver variability for the glottic larynx (ICC = 0.27, mean CI = 0.37 and 3D SD = 3.9 mm). Better agreement in delineations was observed for the other OARs (range, ICC = 0.32-0.83, mean CI = 0.64-0.71 and 3D SD = 0.9-2.6 mm). Cranial, caudal, and medial regions of the OARs showed largest variations. All endpoints provided support for improvement of delineation practice.

Conclusions

Variation in delineation is traced to several regional causes. Measures to reduce this variation can be: (1) guideline development, (2) joint delineation review sessions and (3) application of multimodality imaging. Improvement of delineation practice is needed to standardize patient treatments.

Adhesion forces and coaggregation between vaginal staphylococci and lactobacilli

Urogenital infections are the most common ailments afflicting women. They are treated with dated antimicrobials whose efficacy is diminishing. The process of infection involves pathogen adhesion and displacement of indigenous Lactobacillus crispatus and Lactobacillus jensenii. An alternative therapeutic approach to antimicrobial therapy is to reestablish lactobacilli in this microbiome through probiotic administration. We hypothesized that lactobacilli displaying strong adhesion forces with pathogens would facilitate coaggregation between the two strains, ultimately explaining the elimination of pathogens seen in vivo. Using atomic force microscopy, we found that adhesion forces between lactobacilli and three virulent toxic shock syndrome toxin 1-producing Staphylococcus aureus strains, were significantly stronger (2.2–6.4 nN) than between staphylococcal pairs (2.2–3.4 nN), especially for the probiotic Lactobacillus reuteri RC-14 (4.0–6.4 nN) after 120 s of bond-strengthening. Moreover, stronger adhesion forces resulted in significantly larger coaggregates. Adhesion between the bacteria occurred instantly upon contact and matured within one to two minutes, demonstrating the potential for rapid anti-pathogen effects using a probiotic. Coaggregation is one of the recognized mechanisms through which lactobacilli can exert their probiotic effects to create a hostile micro-environment around a pathogen. With antimicrobial options fading, it therewith becomes increasingly important to identify lactobacilli that bind strongly with pathogens.

Multimorbidity Frameworks Impact Prevalence and Relationships with Patient-Important Outcomes

OBJECTIVES

To explore how different frameworks and categories of chronic conditions impact multimorbidity (defined as two or more chronic conditions) prevalence estimates and associations with patient-important functional outcomes.

DESIGN

Baseline data from a population-based cohort study.

SETTING

National sample of Canadians.

PARTICIPANTS

A total of 51 338 community-living adults, aged 45 to 85 years.

MAIN OUTCOME MEASURES

Chronic conditions from three commonly recognized frameworks were categorized as: (1) diseases, (2) risk factors, or (3) symptoms. Estimates of multimorbidity prevalence were compared among frameworks by age and sex. Separate weighted logistic regression models were used to explore the impact of the different frameworks and categories of chronic conditions on odds ratios (ORs) for multimorbidity for four patient-important functional outcomes: disability, social participation restriction, and self-rated physical and mental health.

RESULTS

One framework included diseases and risk factors, and two frameworks included diseases, risk factors, and symptoms. The prevalence of multimorbidity differed among the frameworks, ranging from 33.5% to 60.6% having two or more chronic conditions. Including risk factors in frameworks increased prevalence estimates, while including symptoms increased prevalence estimates and associations with most patient-important outcomes. The two frameworks that included symptoms had the largest ORs for associations with disability, social participation restriction, and self-rated physical health but not self-rated mental health. Similar results were found when we compared ORs for patient-important outcome for multimorbidity based on three subframeworks: one including diseases only, one including diseases and risk factors, and one including diseases, risk factors, and symptoms.

CONCLUSIONS

Including risk factors appeared to increase only the prevalence of multimorbidity without significantly altering relationships to outcomes. The inclusion of symptoms increased prevalence and associations with patient-important outcomes. These findings underscore the importance of considering not only the number, but also the category, of conditions included in multimorbidity frameworks, as simply counting the number of diagnoses may reduce sensitivity to outcomes that are important to individuals. J Am Geriatr Soc 67:1632-1640, 2019.

Influence of baseline severity on antidepressant efficacy for anxiety disorders: meta-analysis and meta-regression

BACKGROUND

Antidepressants are established first-line treatments for anxiety disorders, but it is not clear whether they are equally effective across the severity range.

AIMS

To examine the influence of baseline severity of anxiety on antidepressant efficacy for generalised anxiety disorder (GAD), social anxiety disorder (SAD), obsessive-compulsive disorder (OCD), post-traumatic stress disorder (PTSD) and panic disorder.

METHOD

Fifty-six trials of second-generation antidepressants for the short-term treatment of an anxiety disorder were included. Baseline and change scores were extracted for placebo and treatment groups in each trial. Mixed effects meta-regression was used to investigate the effects of treatment group, baseline severity and their interaction.

RESULTS

Increased baseline severity did not predict greater improvement in drug groups compared with placebo groups. Standardised regression coefficients of the interaction term between baseline severity and treatment group were 0.04 (95% CI -0.13 to 0.20, P = 0.65) for GAD, -0.06 (95% CI -0.20 to 0.09, P = 0.43) for SAD, 0.04 (95% CI -0.07 to 0.16, P = 0.46) for OCD, 0.16 (95% CI -0.22 to 0.53, P = 0.37) for PTSD and 0.002 (95% CI -0.10 to 0.10, P = 0.96) for panic disorder. For OCD, baseline severity did predict improvement in both placebo and drug groups equally (β = 0.11, 95% CI 0.05 to 0.17, P = 0.001).

CONCLUSIONS

No relationship between baseline severity and drug-placebo difference was found for anxiety disorders. These results suggest that if the efficacy of antidepressants is considered clinically relevant, they may be prescribed to patients with anxiety regardless of symptom severity.

Statistical approaches to harmonize data on cognitive measures in systematic reviews are rarely reported

OBJECTIVES

To identify statistical methods for harmonization, the procedures aimed at achieving the comparability of previously collected data, which could be used in the context of summary data and individual participant data meta-analysis of cognitive measures.

STUDY DESIGN AND SETTING

Environmental scan methods were used to conduct two reviews to identify (1) studies that quantitatively combined data on cognition and (2) general literature on statistical methods for data harmonization. Search results were rapidly screened to identify articles of relevance.

RESULTS

All 33 meta-analyses combining cognition measures either restricted their analyses to a subset of studies using a common measure or combined standardized effect sizes across studies; none reported their harmonization steps before producing summary effects. In the second scan, three general classes of statistical harmonization models were identified (1) standardization methods, (2) latent variable models, and (3) multiple imputation models; few publications compared methods.

CONCLUSION

Although it is an implicit part of conducting a meta-analysis or pooled analysis, the methods used to assess inferential equivalence of complex constructs are rarely reported or discussed. Progress in this area will be supported by guidelines for the conduct and reporting of the data harmonization and integration and by evaluating and developing statistical approaches to harmonization.

Biomechanics of slow running and walking with a rocker shoe

Evidence suggests a link between the loading of the Achilles tendon and the magnitude of the ankle internal plantar flexion moment during late stance of gait, which is clinically relevant in the management of Achilles tendinopathy. Some studies showed that rocker shoes can reduce the ankle internal plantar flexion moment. However, the existing evidence is not conclusive and focused on walking and scarce in running. Sixteen healthy runners participated in this study. Lower extremity kinetics, kinematics and electromyographic (EMG) signals of triceps surae and tibialis anterior were quantified for two types of shoes during running and walking. The peak ankle plantar flexion moment was reduced significantly in late stance of running (0.27 Nm/kg; p<0.001) and walking (0.24 Nm/kg; p<0.001) with the rocker shoe compared to standard shoe. The ankle power generation and plantar flexion moment impulse were also reduced significantly when running and walking with the rocker shoe (p<0.001). No significant changes in the knee and hip moments were found in running and walking. A significant delay of the EMG peak, approximately 2% (p<0.001), was present in the triceps surae when walking with rocker shoes. There were no significant changes in the EMG peak amplitude of triceps surae in running and walking. The peak amplitude of tibialis anterior was significantly increased (64.7 μV, p<0.001) when walking with rocker shoes. The findings show that rocker shoes reduce the ankle plantar flexion moment during the late stance phase of running and walking in healthy people.

Differences in estimates for 10-year risk of cardiovascular disease in Black versus White individuals with identical risk factor profiles using pooled cohort equations: an in silico cohort study

BACKGROUND

Sex-specific and race-specific pooled cohort equations (PCEs) are recommended for estimating the 10-year risk of cardiovascular disease, with an absolute risk of more than 7·5% indicating a clinical decision threshold. We compared differences between Black and White individuals in PCE-estimated absolute cardiovascular disease risk across various plausible risk factor combinations with the aim of evaluating if using the PCE might result in different clinical decisions in Black versus White individuals with identical risk profiles.

METHODS

We generated in silico patient risk profiles by combining numerical risk factors (age [5-year intervals], total cholesterol [20-mg/dl intervals], HDL cholesterol [5-mg/dl intervals], systolic blood pressure [10-mm Hg intervals]) and binary risk factors (smoking, diabetes, and antihypertensive treatment). We compared PCE-estimated 10-year cardiovascular disease risk in Black versus White individuals with identical risk profiles. We did similar comparisons using eligible participants in the Framingham Heart Study (FHS) third generation cohort and the National Health and Nutrition Examination Survey (NHANES) 2017-18.

FINDINGS

For our in silico analysis, we evaluated 29 515 risk factor combinations for women and 30 565 for men, after excluding profiles that generated 10-year cardiovascular disease risk estimates below 1% or above 30%. There were 6357 risk profiles associated with 10-year cardiovascular disease risk above 7·5% for Black men but not for White men (median risk difference [RD] 6·25%, range 0·15-22·8; median relative risk [RR] 2·40, range 1·02-12·6). There were 391 profiles with 10-year cardiovascular disease risk above 7·5% for White men but not Black men (median RD 2·68%, range 0·07-16·9%; median RR 1·42, range 1·01-3·57). There were 6543 risk profiles associated with 10-year estimated cardiovascular disease risk above 7·5% for Black women but not for White women (median RD 6·14%, range 0·35-26·8%; median RR 2·29, range 1·05-12·6). There were 318 profiles with 10-year cardiovascular disease risk above 7·5% for White women but not Black women (median RD 3·71%, range 0·22-20·1%; median RR 1·66, range 1·03-5·46). For the population-based samples, we calculated the PCE-based 10-year cardiovascular disease risk for 1272 eligible participants (378 women; median age 48 years [IQR 44-52]; 100% White) in the FHS third generation cohort and 550 participants (223 women [36·8% Black] and 327 men [40·4% Black]; median age 61 years [IQR 52-67]) in the NHANES cohort. The population-based samples showed similar risk differences to that of the in silico analyses.

INTERPRETATION

The PCE might generate substantially divergent cardiovascular disease risk estimates for Black versus White individuals with identical risk profiles, which could introduce race-related variations in clinical recommendations for cardiovascular disease prevention.

FUNDING

US National Institutes of Health.

Rocker shoe, minimalist shoe, and standard running shoe: a comparison of running economy

OBJECTIVES

Running with rocker shoes is believed to prevent lower limb injuries. However, it is not clear how running in these shoes affects the energy expenditure. The purpose of this study was, therefore, to assess the effects of rocker shoes on running economy in comparison with standard and minimalist running shoes.

DESIGN

Cross-over design.

METHODS

Eighteen endurance female runners (age=23.6 ± 3 years), who were inexperienced in running with rocker shoes and with minimalist/barefoot running, participated in this study. Oxygen consumption, carbon dioxide production, heart rate and rate of perceived exertion were measured while participants completed a 6-min sub-maximal treadmill running test for each footwear condition. The data of the last 2 min of each shoe condition were averaged for analysis. A linear mixed model was used to compare differences among three footwear conditions.

RESULTS

Oxygen consumption during running with rocker shoes was on average 4.5% higher than with the standard shoes (p<0.001) and 5.6% higher than with the minimalist shoe (p<0.001). No significant differences were found in heart rate and rate of perceived exertion across three shoe conditions.

CONCLUSIONS

Female runners, who are not experienced in running with the rocker shoes and minimalist shoes, show more energy expenditure during running with the rocker shoes compared with the standard and minimalist shoes. As the studied shoes were of different masses, part of the effect of increased energy expenditure with the rocker shoe is likely to be due to its larger mass as compared with standard running shoes and minimalist shoes.

Effect of rocker shoes on plantar pressure pattern in healthy female runners

Rocker profile shoes (rocker shoes) are one of the treatment options of metatarsalgia and forefoot stress fractures. The efficacy of rocker shoes in unloading the forefoot pressure has been shown in walking. In running, however, the effect of rocker shoes on forefoot pressure is unknown. Eighteen healthy female runners participated in this study. In-shoe plantar pressures were recorded during running with the standard running shoes and rocker shoes. Shoe comfort was assessed after each shoe measurement. Peak pressure (PP), maximum mean pressure (MMP) and force-time integral (FTI) were determined for seven foot areas. The effects of shoes on the different outcome variables were statistically analyzed using a linear mixed model. Running with the rocker shoes caused a significant reduction (p<0.001) in all pressure parameters in the central and lateral forefoot. FTI and MMP were also reduced by 11% and 12% in the medial forefoot while running with rocker shoes. Running with rocker shoes resulted in a significant increase in all pressure parameters at the heel region (p<0.001). Running with rocker shoes received a significant (p<0.01) lower comfort rate than running with standard running shoes. Rocker shoes might be beneficial for runners who are recovering from metatarsalgia or stress fractures of the forefoot region, as it reduces plantar pressure in the forefoot region.

Comparison of Standardization Methods for the Harmonization of Phenotype Data: An Application to Cognitive Measures

Standardization procedures are commonly used to combine phenotype data that were measured using different instruments, but there is little information on how the choice of standardization method influences pooled estimates and heterogeneity. Heterogeneity is of key importance in meta-analyses of observational studies because it affects the statistical models used and the decision of whether or not it is appropriate to calculate a pooled estimate of effect. Using 2-stage individual participant data analyses, we compared 2 common methods of standardization, T-scores and category-centered scores, to create combinable memory scores using cross-sectional data from 3 Canadian population-based studies (the Canadian Study on Health and Aging (1991-1992), the Canadian Community Health Survey on Healthy Aging (2008-2009), and the Quebec Longitudinal Study on Nutrition and Aging (2004-2005)). A simulation was then conducted to assess the influence of varying the following items across population-based studies: 1) effect size, 2) distribution of confounders, and 3) the relationship between confounders and the outcome. We found that pooled estimates based on the unadjusted category-centered scores tended to be larger than those based on the T-scores, although the differences were negligible when adjusted scores were used, and that most individual participant data meta-analyses identified significant heterogeneity. The results of the simulation suggested that in terms of heterogeneity, the method of standardization played a smaller role than did different effect sizes across populations and differential confounding of the outcome measure across studies. Although there was general consistency between the 2 types of standardization methods, the simulations identified a number of sources of heterogeneity, some of which are not the usual sources considered by researchers.

Do individuals with and without depression value depression differently? And if so, why?

Purpose

Health state valuations, used to evaluate the effectiveness of healthcare interventions, can be obtained either by the patients or by the general population. The general population seems to value somatic conditions more negatively than patients, but little is known about valuations of psychological conditions. This study examined whether individuals with and without depression differ in their valuations of depression and whether perceptions regarding depression (empathy, perceived susceptibility, stigma, illness perceptions) and individual characteristics (mastery, self-compassion, dysfunctional attitudes) bias valuations of either individuals with or without depression.

Methods

In an online study, a general population sample used a time-trade-off task to value 30 vignettes describing depression states (four per participant) and completed questionnaires on perceptions regarding depression and individual characteristics. Participants were assigned to depression groups (with or without depression), based on the PHQ-9. A generalized linear mixed model was used to assess discrepancies in valuations and identify their determinants.

Results

The sample (N = 1268) was representative of the Dutch population on age, gender, education and residence. We found that for mild depression states, individuals with depression (N = 200) valued depression more negatively than individuals without depression (N = 1068) (p = .007). Variables related to perceptions of depression and individual characteristics were not found to affect valuations of either individuals with or individuals without depression.

Conclusion

Since the general population values depression less negatively, using their perspective might result in less effectiveness for interventions for mild depression. Perceptions of depression or to individual characteristics did not seem to differentially affect valuations made by either individuals with or without depression.

The analysis of terminal endpoint events in stepped wedge designs

The stepped wedge design is a unique clinical trial design that allows for a sequential introduction of an intervention. However, the statistical analysis is unclear when this design is applied in survival data. The time-dependent introduction of the intervention in combination with terminal endpoints and interval censoring makes the analysis more complicated. In this paper, a time-on-study scale discrete survival model was constructed. Simulations were conducted primarily to study the performance of our model for different settings of the stepped wedge design. Secondary, we compared our approach to continuous Cox proportional hazard model. The results show that the discrete survival model estimates the intervention effects unbiasedly. If the length of the censoring interval is increased, the precision of the estimates is decreased. Without left truncation and late entry, the number of steps improves the precision of the estimates, whereas in combination of left truncation and late entry, the number of steps decreases the precision. Given the same number of participants and clusters, a parallel group design has higher precision than a stepped wedge design. Copyright © 2016 John Wiley & Sons, Ltd.

Improving outcomes of AF ablation by integrated personalized lifestyle interventions: rationale and design of the prevention to improve outcomes of PVI (POP) trial

Progression of atrial fibrillation (AF) and outcomes of ablation therapy are strongly affected by modifiable risk factors. Although previous studies show beneficial effects of modifying single risk factors, there is lack of evidence from randomized controlled trials on the effects of integrated AF lifestyle programmes. The POP trial is designed to evaluate the clinical outcomes of a dedicated nurse-led AF lifestyle outpatient clinic in patients with symptomatic AF. This study is a prospective, 1:1 randomized, single centre, investigator-initiated clinical trial in 150 patients with paroxysmal or persistent AF referred for a first pulmonary vein isolation (PVI). Prior to the ablation, patients in the intervention group receive a personalized risk factor treatment programme in a specialized, protocolized, nurse-led outpatient clinic. Patient education and durable lifestyle management is promoted with an e-health platform. Patients in the control group receive standard care by cardiologists before ablation. The primary endpoint is the number of hospitalizations for re-ablation and cardioversion, with a follow-up of 12 months after ablation. Secondary endpoints include mortality, number of acute ischemic events, stroke or hospitalizations for heart failure, quality of life, number of ablations cancelled because of symptom reduction, and ablation success rate at 12 months. Determinants of patient and staff experience are explored and a cost-effectiveness analysis is included. The POP trial will help ascertain the efficacy and cost-effectiveness of an integrated technology-supported lifestyle therapy in patients with symptomatic AF. The trial is funded by the Netherlands Organisation for Health Research and Development [10070012010001]. Home sleep apnoea testing devices were provided by Itamar Medical, Ltd.ClinicalTrials.gov Identifier NCT05148338. AF atrial fibrillation, OSA obstructive sleep apnoea, PFA pulsed field ablation, PVI pulmonary vein isolation.

The Combined Effect of Cancer and Cardiometabolic Conditions on the Mortality Burden in Older Adults

BACKGROUND

The number of older people living with cancer and cardiometabolic conditions is increasing, but little is known about how specific combinations of these conditions impact mortality.

METHODS

A total of 22,692 participants aged 65 years and older from four international cohorts were followed-up for mortality for an average of 10 years (8,596 deaths). Data were harmonized across cohorts and mutually exclusive groups of disease combinations were created for cancer, myocardial infarction (MI), stroke, and diabetes at baseline. Cox proportional hazards models for all-cause mortality were used to estimate the age- and sex-adjusted hazard ratio and rate advancement period (RAP) (in years).

RESULTS

At baseline, 23.6% (n = 5,116) of participants reported having one condition and 4.2% (n = 955) had two or more conditions. Data from all studies combined showed that the RAP increased with each additional condition. Diabetes advanced the rate of dying by the most years (5.26 years; 95% confidence interval [CI], 4.53-6.00), but the effect of any single condition was smaller than the effect of disease combinations. Some combinations had a significantly greater impact on the period by which the rate of death was advanced than others with the same number of conditions, for example, 10.9 years (95% CI, 9.4-12.6) for MI and diabetes versus 6.4 years (95% CI, 4.3-8.5) for cancer and diabetes.

CONCLUSIONS

Combinations of cancer and cardiometabolic conditions accelerate mortality rates in older adults differently. Although most studies investigating mortality associated with multimorbidity used disease counts, these provide little guidance for managing complex patients as they age.

The electrical heart axis of the fetus between 18 and 24 weeks of gestation: A cohort study

Introduction

A fetal anomaly scan in mid-pregnancy is performed, to check for the presence of congenital anomalies, including congenital heart disease (CHD). Unfortunately, 40% of CHD is still missed. The combined use of ultrasound and electrocardiography might boost detection rates. The electrical heart axis is one of the characteristics which can be deduced from an electrocardiogram (ECG). The aim of this study was to determine reference values for the electrical heart axis in healthy fetuses around 20 weeks of gestation.

Material and methods

Non-invasive fetal electrocardiography was performed subsequent to the fetal anomaly scan in pregnant women carrying a healthy singleton fetus between 18 and 24 weeks of gestation. Eight adhesive electrodes were applied on the maternal abdomen including one ground and one reference electrode, yielding six channels of bipolar electrophysiological measurements. After removal of interferences, a fetal vectorcardiogram was calculated and then corrected for fetal orientation. The orientation of the electrical heart axis was determined from this normalized fetal vectorcardiogram. Descriptive statistics were used on normalized cartesian coordinates to determine the average electrical heart axis in the frontal plane. Furthermore, 90% prediction intervals (PI) for abnormality were calculated.

Results

Of the 328 fetal ECGs performed, 281 were included in the analysis. The average electrical heart axis in the frontal plane was determined at 122.7° (90% PI: -25.6°; 270.9°).

Discussion

The average electrical heart axis of healthy fetuses around mid-gestation is oriented to the right, which is, due to the unique fetal circulation, in line with muscle distribution in the fetal heart.