README.md

---
language:
- en
license: other
extra_gated_prompt: "To get access to this dataset, you must subscribe to Papers With Backtest. To subscribe, go to https://paperswithbacktest.com/ > Login > Choose Your Plan > Subscribe."
dataset_info:
  features:
  - name: symbol
    dtype: string
  - name: datetime
    dtype: string
  - name: source
    dtype: string
  - name: subsidiary
    dtype: string
  - name: sponsor
    dtype: string
  - name: class
    dtype: string
  - name: trial_id
    dtype: string
  - name: sponsor_study_id
    dtype: string
  - name: official_title
    dtype: string
  - name: brief_title
    dtype: string
  - name: lead_sponsor
    dtype: string
  - name: lead_sponsor_name
    dtype: string
  - name: sponsor_type
    dtype: string
  - name: lead_sponsor_type
    dtype: string
  - name: study_type
    dtype: string
  - name: phase_category
    dtype: string
  - name: enrollment_type
    dtype: string
  - name: enrollment_count
    dtype: int64
  - name: study_size_category
    dtype: string
  - name: healthy_volunteers
    dtype: bool
  - name: condition_keywords
    dtype: string
  - name: primary_condition
    dtype: string
  - name: intervention_type
    dtype: string
  - name: primary_intervention
    dtype: string
  - name: intervention_name
    dtype: string
  - name: intervention_arm_group_labels
    dtype: string
  - name: intervention_description
    dtype: string
  - name: has_data_monitoring_committee
    dtype: bool
  - name: responsible_party_investigator_affiliation
    dtype: string
  - name: responsible_party_investigator_title
    dtype: string
  - name: responsible_party_investigator_name
    dtype: string
  - name: first_posted_date
    dtype: string
  - name: last_update_posted_date
    dtype: string
  - name: start_date
    dtype: string
  - name: primary_completion_date
    dtype: string
  - name: study_completion_date
    dtype: string
  - name: study_locations_city
    dtype: string
  - name: study_locations_state
    dtype: string
  - name: study_locations_country
    dtype: string
  - name: study_locations_zip
    dtype: string
  - name: study_locations_facility
    dtype: string
  - name: study_locations_geopoint
    dtype: string
  - name: standard_age_groups
    dtype: string
  - name: sex
    dtype: string
  - name: minimum_age
    dtype: int64
  - name: maximum_age
    dtype: int64
  - name: overall_status
    dtype: string
  - name: status_category
    dtype: string
  - name: status_verified_date
    dtype: string
  - name: primary_outcomes_measures
    dtype: string
  - name: primary_outcomes_timeframes
    dtype: string
  - name: primary_outcomes_descriptions
    dtype: string
  - name: secondary_outcomes_measures
    dtype: string
  - name: secondary_outcomes_timeframes
    dtype: string
  - name: secondary_outcomes_descriptions
    dtype: string
  - name: has_results
    dtype: bool
  - name: conditions
    dtype: string
  - name: brief_summary
    dtype: string
  - name: detailed_description
    dtype: string
  - name: masking
    dtype: string
  - name: allocation
    dtype: string
  - name: intervention_model
    dtype: string
  - name: primary_purpose
    dtype: string
  - name: has_expanded_access
    dtype: bool
  - name: study_duration_days
    dtype: int64
  - name: trial_duration
    dtype: float64
  - name: references_type
    dtype: string
  - name: references_citation
    dtype: string
  - name: references_pmid
    dtype: string
  - name: collaborators_name
    dtype: string
  - name: collaborators_class
    dtype: string
  - name: ipd_sharing
    dtype: string
  - name: success_prediction
    dtype: float64
  - name: economic_effect
    dtype: float64
  - name: duration_prediction
    dtype: float64
  - name: success_composite
    dtype: float64

---
# Dataset Information

Weekly **clinical trial intelligence** covering regulatory progress, sponsor metadata, and modelled outcomes, sourced from **SOV.AI**. Each row describes a single trial snapshot with rich metadata plus predictive scores for success, duration, and economic impact.

- **Coverage**: Global clinical trials mapped to corporate tickers or sponsor labels
- **Update cadence**: Weekly, with an average of **1,052 new trials** added per week
- **Performance**: Predictive models achieving **87% ROC-AUC** on held-out validation data
- **Contents**: Trial identifiers, sponsor information, design details, eligibility, outcomes, site geography, and SOV.AI predictive scores

## Data Source
Data provided by SOV.AI. Access programmatically via:

```python
import sovai as sov
df_clinical = sov.data("clinical/trials", full_history=True)
```

## Notes
- Date fields are normalised to ISO-8601 (`YYYY-MM-DD`).
- Boolean columns use nullable boolean semantics to preserve missing values.
- `success_prediction`, `success_composite`, and `economic_effect` are probabilistic or index scores between 0–1.
- `duration_prediction` and `trial_duration` express durations in days.