Data sharing in clinical trials: An experience with two large cancer screening trials

Status, use and impact of sharing individual participant data from clinical trials: a scoping review

OBJECTIVES

To explore the impact of data-sharing initiatives on the intent to share data, on actual data sharing, on the use of shared data and on research output and impact of shared data.

ELIGIBILITY CRITERIA

All studies investigating data-sharing practices for individual participant data (IPD) from clinical trials.

SOURCES OF EVIDENCE

We searched the Medline database, the Cochrane Library, the Science Citation Index Expanded and the Social Sciences Citation Index via Web of Science, and preprints and proceedings of the International Congress on Peer Review and Scientific Publication. In addition, we inspected major clinical trial data-sharing platforms, contacted major journals/publishers, editorial groups and some funders.

CHARTING METHODS

Two reviewers independently extracted information on methods and results from resources identified using a standardised questionnaire. A map of the extracted data was constructed and accompanied by a narrative summary for each outcome domain.

RESULTS

93 studies identified in the literature search (published between 2001 and 2020, median: 2018) and 5 from additional information sources were included in the scoping review. Most studies were descriptive and focused on early phases of the data-sharing process. While the willingness to share IPD from clinical trials is extremely high, actual data-sharing rates are suboptimal. A survey of journal data suggests poor to moderate enforcement of the policies by publishers. Metrics provided by platforms suggest that a large majority of data remains unrequested. When requested, the purpose of the reuse is more often secondary analyses and meta-analyses, rarely re-analyses. Finally, studies focused on the real impact of data-sharing were rare and used surrogates such as citation metrics.

CONCLUSIONS

There is currently a gap in the evidence base for the impact of IPD sharing, which entails uncertainties in the implementation of current data-sharing policies. High level evidence is needed to assess whether the value of medical research increases with data-sharing practices.

Obtaining and managing data sets for individual participant data meta-analysis: scoping review and practical guide

BACKGROUND

Shifts in data sharing policy have increased researchers' access to individual participant data (IPD) from clinical studies. Simultaneously the number of IPD meta-analyses (IPDMAs) is increasing. However, rates of data retrieval have not improved. Our goal was to describe the challenges of retrieving IPD for an IPDMA and provide practical guidance on obtaining and managing datasets based on a review of the literature and practical examples and observations.

METHODS

We systematically searched MEDLINE, Embase, and the Cochrane Library, until January 2019, to identify publications focused on strategies to obtain IPD. In addition, we searched pharmaceutical websites and contacted industry organizations for supplemental information pertaining to recent advances in industry policy and practice. Finally, we documented setbacks and solutions encountered while completing a comprehensive IPDMA and drew on previous experiences related to seeking and using IPD.

RESULTS

Our scoping review identified 16 articles directly relevant for the conduct of IPDMAs. We present short descriptions of these articles alongside overviews of IPD sharing policies and procedures of pharmaceutical companies which display certification of Principles for Responsible Clinical Trial Data Sharing via Pharmaceutical Research and Manufacturers of America or European Federation of Pharmaceutical Industries and Associations websites. Advances in data sharing policy and practice affected the way in which data is requested, obtained, stored and analyzed. For our IPDMA it took 6.5 years to collect and analyze relevant IPD and navigate additional administrative barriers. Delays in obtaining data were largely due to challenges in communication with study sponsors, frequent changes in data sharing policies of study sponsors, and the requirement for a diverse skillset related to research, administrative, statistical and legal issues.

CONCLUSIONS

Knowledge of current data sharing practices and platforms as well as anticipation of necessary tasks and potential obstacles may reduce time and resources required for obtaining and managing data for an IPDMA. Sufficient project funding and timeline flexibility are pre-requisites for successful collection and analysis of IPD. IPDMA researchers must acknowledge the additional and unexpected responsibility they are placing on corresponding study authors or data sharing administrators and should offer assistance in readying data for sharing.

Deep Learning to Assess Long-term Mortality From Chest Radiographs

Importance

Chest radiography is the most common diagnostic imaging test in medicine and may also provide information about longevity and prognosis.

Objective

To develop and test a convolutional neural network (CNN) (named CXR-risk) to predict long-term mortality, including noncancer death, from chest radiographs.

Design, Setting, and Participants

In this prognostic study, CXR-risk CNN development (n = 41 856) and testing (n = 10 464) used data from the screening radiography arm of the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) (n = 52 320), a community cohort of asymptomatic nonsmokers and smokers (aged 55-74 years) enrolled at 10 US sites from November 8, 1993, through July 2, 2001. External testing used data from the screening radiography arm of the National Lung Screening Trial (NLST) (n = 5493), a community cohort of heavy smokers (aged 55-74 years) enrolled at 21 US sites from August 2002, through April 2004. Data analysis was performed from January 1, 2018, to May 23, 2019.

Exposure

Deep learning CXR-risk score (very low, low, moderate, high, and very high) based on CNN analysis of the enrollment radiograph.

Main Outcomes and Measures

All-cause mortality. Prognostic value was assessed in the context of radiologists' diagnostic findings (eg, lung nodule) and standard risk factors (eg, age, sex, and diabetes) and for cause-specific mortality.

Results

Among 10 464 PLCO participants (mean [SD] age, 62.4 [5.4] years; 5405 men [51.6%]; median follow-up, 12.2 years [interquartile range, 10.5-12.9 years]) and 5493 NLST test participants (mean [SD] age, 61.7 [5.0] years; 3037 men [55.3%]; median follow-up, 6.3 years [interquartile range, 6.0-6.7 years]), there was a graded association between CXR-risk score and mortality. The very high-risk group had mortality of 53.0% (PLCO) and 33.9% (NLST), which was higher compared with the very low-risk group (PLCO: unadjusted hazard ratio [HR], 18.3 [95% CI, 14.5-23.2]; NLST: unadjusted HR, 15.2 [95% CI, 9.2-25.3]; both P < .001). This association was robust to adjustment for radiologists' findings and risk factors (PLCO: adjusted HR [aHR], 4.8 [95% CI, 3.6-6.4]; NLST: aHR, 7.0 [95% CI, 4.0-12.1]; both P < .001). Comparable results were seen for lung cancer death (PLCO: aHR, 11.1 [95% CI, 4.4-27.8]; NLST: aHR, 8.4 [95% CI, 2.5-28.0]; both P ≤ .001) and for noncancer cardiovascular death (PLCO: aHR, 3.6 [95% CI, 2.1-6.2]; NLST: aHR, 47.8 [95% CI, 6.1-374.9]; both P < .001) and respiratory death (PLCO: aHR, 27.5 [95% CI, 7.7-97.8]; NLST: aHR, 31.9 [95% CI, 3.9-263.5]; both P ≤ .001).

Conclusions and Relevance

In this study, the deep learning CXR-risk score stratified the risk of long-term mortality based on a single chest radiograph. Individuals at high risk of mortality may benefit from prevention, screening, and lifestyle interventions.

Prostate Cancer Incidence and Mortality in Relationship to Family History of Prostate Cancer; Findings From The PLCO Trial

BACKGROUND

The purpose of the study was to determine the relationship between family history of prostate cancer in a first-degree relative (FDR) and prostate cancer incidence and mortality.

PATIENTS AND METHODS

Deidentified data sets of men recruited in the Prostate, Lung, Colorectal, and Ovary (PLCO) trial were accessed. Men with complete information about family history of prostate cancer in an FDR were included. The effect of family history on prostate cancer incidence and mortality was assessed in a multivariate Cox regression model. Likewise, the effect of the number of FDRs with prostate cancer and the effect of youngest diagnosis age of an FDR with prostate cancer were assessed.

RESULTS

A total of 74,781 participants were included in the current analysis, including 5281 participants with family history of prostate cancer in an FDR and 69,500 participants without family history of prostate cancer in an FDR. Among participants without family history of prostate cancer in an FDR, a total of 7450 patients (10.5%) were subsequently diagnosed with prostate cancer; whereas among patients with family history of prostate cancer in an FDR, a total of 889 patients (16.5%) were subsequently diagnosed with prostate cancer. In an adjusted multivariate Cox regression model, family history of prostate cancer was associated with a higher probability of prostate cancer diagnosis (hazard ratio [HR], 1.590; 95% confidence interval [CI], 1.482-1.705; P < .001). The number of FDRs with prostate cancer proportionally correlated with higher prostate cancer incidence (HR, 1.529; 95% confidence interval [CI], 1.439-1.624; P < .001). Family history of prostate cancer in an FDR was not predictive of higher prostate cancer mortality in the PLCO screening (intervention) arm (HR, 0.829; 95% CI, 0.422-1.629; P = .587) whereas it was predictive of a higher prostate cancer mortality in the PLCO nonscreening (control) arm (HR, 1.894; 95% CI, 1.154-3.109; P = .012). Number of FDRs with prostate cancer was not associated with higher prostate cancer mortality in the PLCO screening (intervention) arm (HR, 0.956; 95% CI, 0.541-1.691; P = .878), whereas it was associated with higher prostate cancer mortality in the PLCO nonscreening (control) arm (HR, 1.643; 95% CI, 1.083-2.493; P = .020).

CONCLUSION

Family history of prostate cancer is associated with an increased risk of prostate cancer diagnosis in the overall cohort of patients as well as a higher risk of prostate cancer mortality in the nonscreened subcohort. Further prospective assessment of the role of screening among selected high-risk populations (including those with strong family history) is warranted.

Availability and Use of Shared Data From Cardiometabolic Clinical Trials

BACKGROUND

Sharing of patient-level clinical trial data has been widely endorsed. Little is known about how extensively these data have been used for cardiometabolic diseases. We sought to evaluate the availability and use of shared data from cardiometabolic clinical trials.

METHODS

We extracted data from ClinicalStudyDataRequest.com, a large, multisponsor data-sharing platform hosting individual patient-level data from completed studies sponsored by 13 pharmaceutical companies.

RESULTS

From January 2013 to May 2017, the platform had data from 3374 clinical trials, of which 537 (16%) evaluated cardiometabolic therapeutics (phase 1, 36%; phase 2, 17%; phase 2/3, 1%; phase 3, 42%; phase 4, 4%). They covered 74 therapies and 398 925 patients. Diabetes mellitus (60%) and hypertension (15%) were the most common study topics. Median time from study completion to data availability was 79 months. As of May 2017, ClinicalStudyDataRequest.com had received 318 submitted proposals, of which 163 had signed data-sharing agreements. Thirty of these proposals were related to cardiometabolic therapies and requested data from 79 unique studies (15% of all trials, 29% of phase 3/4 trials). Most (96%) data requesters of cardiometabolic clinical trial data were from academic centers in North America and Western Europe, and half the proposals were unfunded. Most proposals were for secondary hypothesis-generating questions, with only 1 proposed reanalysis of the original study primary hypothesis. To date, 3 peer-reviewed articles have been published after a median of 19 months (9-32 months) from the data-sharing agreement.

CONCLUSIONS

Despite availability of data from >500 cardiometabolic trials in a multisponsor data-sharing platform, only 15% of these trials and 29% of phase 3/4 trials have been accessed by investigators thus far, and a negligible minority of analyses have reached publication.