US Clinical Trial Investigator Performance Data

## Summary

SunshineMD's database of US clinical investigators provides a definitive profile of every active US clinical investigator, the studies they have worked on, and how well they have performed. The data uses payments as a surrogate for enrollment performance. The SunshineMD adjusted payments correlate very strongly (.9) to actual enrollments. This has been proven with several sponsor companies' phase 2 and phase 3 studies. In addition to providing accurate payment records for investigators and studies, SunshineMD has two summary performance measures for each investigator: SunshineMD Rank and SunshineMD Startup Rank. These measures are percentile ranks of an investigators overall and startup (respectively) enrollment performance. Subscribers get access to custom (specialty, indication) versions of these measures and investigator contact information (phone and email).

## Data Updates

Data is updated daily from clinicaltrials.gov, and twice a year from Open Payments, corresponding with their release cycles.

## Grouping Payments Accurately

This data source is a result of operationalizing Open Payments research data and ClinicalTrials.gov. As is, Open Payments data is rife with issues, which prevent a simple download of the files to produce strong actionable results. Namely, there is no required study ID in Open Payments. Each payment is reported with a Study Title, which is an open text field. That field is subject to many errors related to humans or machines. Each year a sponsor company reports its payments, it can change the study title and hence make it much harder to combine the payments correctly. Common reporting issues with the study title include punctuation, capitalization, whitespace, spelling, use of abbreviations and acronyms, roman numerals vs arabic numbers, use of internal study IDs, and arbitrary string truncation (e.g. 255 chars).

Some of these issues are overcome with straightforward regex string transformations. However, in many studies, this is not enough to have an accurate enough payment record to have strong correlations with actual enrollments. SunshineMD has developed AI processes to overcome the shortcomings of string matching algorithms (regex, fuzzy algos) to produce accurate payment records.

## The Power of NCTs

With an NCT ID, a wealth of information can be pulled from clinicaltrials.gov including but not limited to conditions/diseases, official and brief study titles, start and end dates, phase, enrollment, and overall status. There is an optional NCT ID field in Open Payments, but it is only populated 15% of the time. For example, Gilead has never reported NCT IDs with its payments. SunshineMD has added over 6,700 NCT IDs to the database, covering 81% of all total payments. These data power SunshineMD’s webapp investigator and study search, which subscribers have access to. Select clinicaltrials.gov data is available in this Snowflake offering. These data points can also be used in constructing data models.

## Snowflake Trial Access

The trial access has study and investigator data from 2014.

搜集汇总

背景与挑战

背景概述

该数据集整合了美国临床研究者的支付数据与试验表现，通过AI技术关联Open Payments和ClinicalTrials.gov信息，提供研究者排名指标及试验详情。数据每日更新，包含2014年以来的研究者表现记录及补充的6700余条NCT ID信息。