clinicaltrials.gov snapshot
To download only this data file: clinicaltrialsgov.rds (16 MB)
To download all BEAR datasets, click here.
Description
Trials which reported results at clinicaltrials.gov (National Library of Medicine 2025)
Snapshot date: 12 August 2025
Data source: This is a set of flat files of studies obtained from the AACT downloads page (https://aact.ctti-clinicaltrials.org/downloads/; over 2GBs of data for relevant tables), with relevant columns extracted and joined by us before BEAR processing.
Data availability: terms and conditions for ClinicalTrials.gov are at https://clinicaltrials.gov/about-site/terms-conditions. We make distributions of derived effect sizes available in accordance with ToC’s. We only use publicly available data and do not imply any endorsement of this work by National Library of Medicine, ClinicalTrials.gov or AACT.
Data processing: extensive. There are some differences between this full data table (BEAR_data/clinicaltrialsgov.rds) and what is retained for the main BEAR dataset. We briefly discuss them below. Because of rich data structure, please refer to data dictionary for the full data table.
Briefly, BEAR targets primary outcome results for completed studies using two data sources:
- Author-reported effect sizes
- Effect sizes derived from raw outcome data (group means, counts), alongside source data
Where both (1) and (2) are available, we prefer (1), but also make (2) available in the full data table. In case (2) we use standardised mean differences for continuous data and probit for count.
If working with full data set: variable author_raw_overlap_key can be used to link the two different calculations of the same effect; variable include_in_bear identifies rows retained after applying this author-preferred rule.
The main BEAR dataset keeps (1) randomized, (2) interventional studies with (3) fewer than 20 candidate effect rows (see below).
The full ClinicalTrials.gov table uses only completed studies but does not apply these three restrictions. It contains 151,313 rows from 27,446 studies; 46.3% are author-reported and 53.7% are raw-derived. The randomized interventional subset used in the main BEAR release contains 60,470 rows from 23,060 studies.
Note that a single ClinicalTrials.gov study can contribute several primary outcome IDs, and each primary outcome can contain many measurements. The full data table records this as n_effect_rows_per_study and n_outcomes_per_study. The median study contributes one outcome ID and two rows to the full data table. Studies with 20 or more candidate effect rows remain in the full data table but are excluded from the main BEAR build.
In the full data table, some raw-derived rows have direction_unknown == TRUE, meaning that the control/reference arm could not be reliably inferred, so the sign should be treated as random. Users interested in controlled trials and/or signed z-values should filter the results accordingly.
See Standard procedure for dealing with p-values and confidence intervals for how we created and chose the z-values based on p-values and confidence intervals. Briefly, we calculated z-values using two separate procedures (based on p-values and confidence intervals) and, if both were available, we made a judgement on which one was more appropriate using a pre-defined rule. ClinicalTrials.gov reports the number of sides for confidence intervals but not for p-values, so we treated p-values as two-sided. We dropped rows only when z was missing. Notably, infinite z-values can arise when p_value is recorded as exactly 0, and these are retained because they are not missing.
Model of z-values
BEAR fits a single model to ClinicalTrials.gov and EU CTR because the two clinical-trials registries have similar reporting structures and z-value derivations. The plot therefore shows the combined clinical-trials fit. The companion dataset page is EU Clinical Trials Register (EU CTR).
| Characteristic | Estimate |
|---|---|
| Probability of significance | 33% |
| Relative probability of publication for |z| < 1.96 | 0.56 |
| Successful replication for |z| > 1.96 | 73% |
| Correct sign for |z| > 1.96 | 97% |
What do these terms mean?
- Probability of significance
- The reported value is the assurance: the proportion of significant results adjusted for publication bias.
- Relative probability of publication
- The relative probability of observing a result below the |z| = 1.96 threshold rather than above it. Values below one indicate lower observation probability below the conventional two-sided significance threshold.
- Successful replication
- The probability that an exact replication has the same sign and |z| greater than 1.96, conditional on the original result having |z| greater than 1.96.
- Correct sign
- The probability that the observed effect has the same direction as the true effect, conditional on an original result with |z| greater than 1.96.
Data dictionary
The summary column is computed from clinicaltrialsgov.rds when this dictionary is regenerated. Counts are row-level effect counts, not unique trial counts.
Row and source identifiers
| Variable | Definition | Summary |
|---|---|---|
effect_id |
Unique row identifier for the merged ClinicalTrials.gov effect row. | |
author_effect_id |
Identifier for the author-reported outcome-analysis effect row. | |
raw_effect_id |
Identifier for the selected raw endpoint calculation row. | |
effect_source |
Source of the effect estimate used for b, se, z, and z_operator. | raw_derived 81,321 (53.7%); author_reported 69,992 (46.3%) |
include_in_bear |
Whether this row is used by the main BEAR ClinicalTrials.gov build. TRUE keeps author-reported rows and raw-derived rows only when no author-reported overlap is available. | TRUE 132,092 (87.3%) |
import_decision |
Import decision for the row after author-reported and raw-derived overlaps are compared. | include_author_preferred 69,992 (46.3%); include_raw_only 62,100 (41.0%); exclude_author_preferred_overlap 19,221 (12.7%) |
author_raw_overlap_key |
Shared key for linking author-reported rows to raw-derived rows for the same study, outcome, and result-group pair. | |
nct_id |
ClinicalTrials.gov study identifier. | |
outcome_id |
AACT outcome identifier linked to the effect row. | |
outcome_analysis_id |
AACT outcome-analysis identifier for source-reported analyses. |
Study characteristics
| Variable | Definition | Summary |
|---|---|---|
brief_title |
Brief trial title. | |
official_title |
Official trial title. | |
study_type |
ClinicalTrials.gov study type. | INTERVENTIONAL 145,055 (95.9%); OBSERVATIONAL 6,258 (4.1%) |
phase |
Trial phase as recorded by ClinicalTrials.gov. | PHASE2 42,295 (28.0%); missing 41,422 (27.4%); PHASE3 29,267 (19.3%); PHASE1 19,455 (12.9%); PHASE4 11,436 (7.6%); remaining 7,438 (4.9%) |
overall_status |
ClinicalTrials.gov recruitment or completion status. | completed 151,313 (100.0%) |
year |
Year used by BEAR for time summaries. | median 2,016 [2,012, 2,020]; range 1,992-2,025 |
enrollment |
Trial enrollment count. | median 122 [50, 400]; range 1-4,052,231; missing 2 |
enrollment_type |
Whether enrollment is actual or anticipated. | |
number_of_arms |
Number of intervention arms recorded in the ClinicalTrials.gov study design table. | median 2 [2, 4]; range 1-32; missing 6,760 |
number_of_groups |
Number of observational groups recorded in the ClinicalTrials.gov study design table; usually missing for interventional trials. | median 1 [1, 2]; range 1-49; missing 145,418 |
results_first_posted_year |
Calendar year when results were first posted. | |
primary_completion_year |
Calendar year of primary completion. | |
n_effect_rows_per_study |
Number of public candidate effect rows linked to the same ClinicalTrials.gov study. | median 2 [1, 4]; range 1-2,474 |
n_outcomes_per_study |
Number of distinct outcome IDs represented for the same ClinicalTrials.gov study in the public file. | median 1 [1, 2]; range 1-96 |
allocation |
Allocation design recorded for the study. The main BEAR ClinicalTrials.gov dataset keeps randomized rows; non-randomized and missing-allocation rows remain identifiable in the public file. | RANDOMIZED 134,860 (89.1%); missing 11,191 (7.4%); NON_RANDOMIZED 5,262 (3.5%) |
intervention_model |
Intervention model recorded for the study. The main BEAR ClinicalTrials.gov dataset does not filter on intervention model, although the raw-derived endpoint path is restricted to eligible randomized parallel interventional trials. | PARALLEL 126,468 (83.6%); CROSSOVER 9,371 (6.2%); SINGLE_GROUP 6,883 (4.5%); missing 6,460 (4.3%); FACTORIAL 1,422 (0.9%); remaining 709 (0.5%) |
primary_purpose |
Primary purpose recorded for the study. The main BEAR ClinicalTrials.gov dataset does not filter on primary purpose. | TREATMENT 91,849 (60.7%); PREVENTION 28,250 (18.7%); missing 8,001 (5.3%); BASIC_SCIENCE 7,730 (5.1%); OTHER 6,179 (4.1%); remaining 9,304 (6.1%) |
observational_model |
Observational model if applicable; this is usually present only for observational studies and is therefore missing for most interventional rows. | |
masking |
Masking design recorded for the study. | NONE 37,804 (25.0%); DOUBLE 36,861 (24.4%); QUADRUPLE 28,602 (18.9%); TRIPLE 25,968 (17.2%); SINGLE 15,381 (10.2%); remaining 6,697 (4.4%) |
is_factorial |
Whether the intervention model indicates a factorial design. | TRUE 1,422 (0.9%); missing 18 (0.0%) |
is_crossover |
Whether the intervention model indicates a crossover design. | TRUE 9,371 (6.2%); missing 18 (0.0%) |
n_design_groups |
Number of design-group records linked to the study. Unlike number_of_arms, this is counted from the design_groups table and is available for comparator summaries. | median 2 [2, 4]; range 0-49 |
has_placebo_comparator |
Whether any design-group record appears to be a placebo comparator. | TRUE 60,202 (39.8%) |
comparator_summary |
Compact study-level comparator summary derived from design-group labels and descriptions. | |
is_multi_arm |
Whether the study has more than two arms or groups. | TRUE 70,856 (46.8%) |
has_baseline_measurements |
Whether baseline measurement rows were available in AACT. | TRUE 151,293 (100.0%) |
Trial categories
| Variable | Definition | Summary |
|---|---|---|
domain_all |
All mapped clinical domains for the trial. | |
domain_n |
Number of mapped clinical domains for the trial. | |
domain_source |
Source used for domain mapping. | |
domain_oncology |
Whether the trial maps to oncology. | TRUE 10,498 (6.9%) |
domain_primary |
Primary mapped clinical domain. | infectious_disease 30,451 (20.1%); unknown 19,341 (12.8%); mental_health 16,643 (11.0%); endocrine_metabolic 13,647 (9.0%); cardiovascular 11,589 (7.7%); remaining 59,642 (39.4%) |
intervention_types_all |
All intervention types recorded for the study. | |
intervention_type_n |
Number of intervention types recorded for the study. | |
intervention_type_primary |
Primary intervention type used for summaries. | drug 91,364 (60.4%); biological 21,460 (14.2%); behavioral 14,415 (9.5%); device 11,064 (7.3%); other 7,017 (4.6%); remaining 5,993 (4.0%) |
is_drug_trial |
Whether any intervention is coded as drug. | TRUE 102,524 (67.8%) |
is_biological_trial |
Whether any intervention is coded as biological. | TRUE 31,637 (20.9%) |
is_device_trial |
Whether any intervention is coded as device. | TRUE 12,304 (8.1%) |
is_procedure_trial |
Whether any intervention is coded as procedure. | TRUE 3,584 (2.4%) |
is_behavioral_trial |
Whether any intervention is coded as behavioral. | TRUE 16,774 (11.1%) |
is_vaccine_trial |
Whether intervention text indicates a vaccine trial. | TRUE 17,880 (11.8%) |
is_cell_or_gene_therapy_trial |
Whether intervention text indicates cell or gene therapy. | TRUE 1,030 (0.7%) |
is_diagnostic_trial |
Whether any intervention is coded as diagnostic test. | TRUE 5,579 (3.7%) |
lead_sponsor_name |
Lead sponsor name. | |
lead_sponsor_class |
Lead sponsor class. | industry 98,817 (65.3%); other 44,853 (29.6%); other_gov 3,959 (2.6%); nih 3,027 (2.0%); network 657 (0.4%) |
has_industry_sponsor |
Whether the lead sponsor class is industry. | TRUE 105,287 (69.6%) |
Outcome descriptors
| Variable | Definition | Summary |
|---|---|---|
outcome_type |
Outcome type recorded in ClinicalTrials.gov. | PRIMARY 151,313 (100.0%) |
outcome_title |
Outcome title. | |
outcome_time_frame |
Outcome time frame. | |
measurement_stratum_id |
AACT measurement stratum identifier for raw-derived rows. | |
measurement_title |
Measurement title for raw-derived endpoint rows. | |
measurement_units |
Measurement units for raw-derived endpoint rows. |
BEAR effect fields
| Variable | Definition | Summary |
|---|---|---|
measure_class |
Effect-size family used by BEAR summaries. | Standardized Mean Difference 49,955 (33.0%); Probit Difference 31,366 (20.7%); missing 24,211 (16.0%); Mean Difference 17,588 (11.6%); Geometric Ratio 7,079 (4.7%); remaining 21,114 (14.0%) |
measure_detailed |
Clarifies the estimator where this is known. SMDs derived by us from raw outcome data are Hedges’ g; where authors supply outcome data, the measure remains SMD because the reported estimator is not generally identifiable. |
|
scale |
Scale of the stored effect size. | raw 53,099 (35.1%); smd 49,955 (33.0%); probit 31,366 (20.7%); log 16,893 (11.2%) |
effect |
Effect estimate used for BEAR-facing analysis. | |
b |
Effect estimate used for z-value calculations where available. | |
se |
Standard error of the effect estimate where available. | |
z |
Signed z-value used in BEAR analyses. | |
z_operator |
Operator describing whether z is exact or bounded. | = 143,593 (94.9%); > 6,915 (4.6%); < 805 (0.5%) |
Raw endpoint calculation inputs
| Variable | Definition | Summary |
|---|---|---|
raw_measure |
Raw endpoint effect measure selected for BEAR import. | missing 69,992 (46.3%); standardized_mean_difference 49,955 (33.0%); probit_difference 31,366 (20.7%) |
raw_effect_family |
Raw endpoint outcome family. | missing 69,992 (46.3%); continuous 49,955 (33.0%); binary 31,366 (20.7%) |
n_raw_overlap_matches |
Number of selected raw endpoint rows matching the same study, outcome, and result-group pair for an author-reported row. | median 0 [0, 0]; range 0-96; missing 81,321 |
raw_event_t |
Event count in the treatment or focal group for binary raw contrasts. | |
raw_event_c |
Event count in the comparator group for binary raw contrasts. | |
raw_n_t |
Analysis count in the treatment or focal group for raw contrasts. | |
raw_n_c |
Analysis count in the comparator group for raw contrasts. | |
raw_mean_t |
Mean in the treatment or focal group for continuous raw contrasts. | |
raw_mean_c |
Mean in the comparator group for continuous raw contrasts. | |
raw_sd_t |
Standard deviation in the treatment or focal group for continuous raw contrasts. | |
raw_sd_c |
Standard deviation in the comparator group for continuous raw contrasts. |
Raw result-group metadata
| Variable | Definition | Summary |
|---|---|---|
raw_group_t_id |
AACT result-group identifier for the treatment or focal group. | |
raw_group_c_id |
AACT result-group identifier for the comparator group. | |
raw_group_t_title |
Title of the treatment or focal result group. | |
raw_group_c_title |
Title of the comparator result group. | |
direction_unknown |
Whether treatment-comparator direction could not be inferred for the raw-derived contrast. | TRUE 22,174 (14.7%); missing 69,992 (46.3%) |
raw_multi_arm_trial |
Whether the raw-derived contrast comes from a multi-arm trial. | TRUE 47,183 (31.2%); missing 69,992 (46.3%) |