Saama’s Smart Applications are revolutionizing data management in clinical research. These AI-powered, independent software modules sit on top of your existing infrastructure to automate manual tasks and analyze data with greater accuracy and speed.
Already deployed by some of the largest pharmaceutical companies and available directly from Saama or through partners such as Oracle, Saama’s Smart Applications are freeing clinical data management professionals from tedious manual work so they can spend more time making critical decisions and contributing strategically to larger clinical research efforts.
The following three Smart Applications are currently available:
Smart Data Quality (SDQ)
Developed in collaboration with Pfizer and famous for helping speed data analysis in Pfizer’s COVID-19 vaccine trial, SDQ (formerly known as Smart Data Query) is a transformative AI engine that accelerates data cleaning and generates and solves queries on the fly.
The machine learning technology enabled Pfizer’s data management team to maintain an exceptional level of data quality throughout the trial: All of the data was ready for review a mere 22 hours after meeting the primary efficacy case counts.
SDQ reduced the cycle time from data entry to query generation by 95%, thanks to its ability to automatically identify data discrepancies with a high degree of accuracy. Not only does the tool improve operational efficiency by reducing human effort in discrepancy identification, it also helps to uncover data mismatches that are often missed by human review.
Here are just some of the reviews that have been automated by SDQ:
- Checking that concomitant medications are consistent with AE terms
- Checking if duplicate medications are given for the same condition
- Checking the relative toxicity of related AEs
- Checking if AE terms are related
Unresolved queries are referred to data managers for further investigation, and the resulting human feedback trains the machine learning model for even greater accuracy over time.
SDQ also offers a medical Coding module, which uses natural language processing (NLP) to auto-code adverse events and medications with accuracy rates up to 250% better than traditional coding methods. Here’s how it works:
- Terms from an EDC application, such as Oracle InForm or Medidata Rave, flow into SCT for auto-coding or querying
- Deep learning models predict coding decisions for each term
- A human user approves or rejects the proposed coding decisions
- Approved terms flow back to the coder via import APIs or flat file downloads, creating a single source of all final coding decisions
In recent tests using popular dictionaries, the Medical Coding Module improved MedDRA coding accuracy from 50%-75% and WHODrug coding accuracy from 30% to 78%.
Smart Auto Mapping (SAM)
Inspecting and mapping complex datasets to CDISC or internal business standards can be tedious and time-consuming work. Smart Auto Mapper (SAM) shifts the paradigm by using AI to overcome standard data onboarding and transformation challenges.
SAM gives data managers an easy way to clean and map non-standard (and often messy and incomplete) data to conform to sponsor submission standards, such as SDTM and SDTM+. About 70% of your work can be automated with codeless, out-of-the-box capabilities that transform disparate data into a submission and analytics-ready dataset. As with SDQ, accuracy is ensured with human-in-the-loop controls that enable data traceability and ongoing training of AI algorithms.
SAM’s flexible and scalable architecture enables seamless integration with source systems and pre-trained machine learning models, making new transformations at the study and therapeutic level more efficient and repeatable.
Biostatisticians and clinical teams also reap the rewards of fast access to the datasets they need with fine-grained permissions for each user. Blinding is customizable at the row/column and conditional levels, and bias can be prevented by blocking statisticians from seeing certain datasets before comparing them with experiment results.
The generated datasets can be used for further downstream analysis, such as ADaM and TLF, or for reporting and analytics use cases like medical and risk-based monitoring..
Smart Programming and Analysis Computing Environment (SPACE)
A statistical computing environment designed just for life sciences, SPACE is designed to make clinical study design, management, and regulatory review much easier.
Because this efficient, collaborative work environment is compatible with all popular IDEs, clinical data programmers and biostatisticians can now analyze data and publish study results with far fewer frustrations.
SPACE’s modern design and built-in flexibility eliminates complexity when it comes to critical activities, such as providing access to quality data, TLF generation, and CDISC dataset formatting. When integrated with Saama’s Clinical Data Hub and other Smart Applications, SPACE delivers a powerful, end-to-end solution for the critical data submission pathway:
- Access and ingest current and historical data
- Write, review, edit, and run programs in any IDE (SAS, Jupyter, Visual Studio, Rstudio, NONMEM, etc.) while maintaining strict version control
- Leverage industry best practice version control
- Conduct exploratory analytics (descriptive and predictive)
- Generate valid submission datasets in SDTM, ADaM, and TLG formats
Saama’s Smart Applications are designed to meet the complex data management and analytics challenges of clinical research sponsors with extensive study portfolios, vast data silos, and diverse end-user needs. These solutions are surgically inserted into existing platforms and application infrastructure, and risk is further mitigated with convenient usage-based pricing.
To learn more and see a demo of Saama’s Smart Applications in action, contact Saama today.