Does CluePoints have a software product that can be used by sponsors or CROs instead of using the service approach?

Yes! CluePoints has developed a Software as a Service (SaaS) version of the solution which was launched in June 2014 to let the sponsor or CRO perform analyses themselves. The SaaS (web application) provides an analytics and visualization suite where customers can have an online view for interrogation of the results from the SMART engine.

On the portal, the sponsor can upload clinical data for assessment, visualize the findings and interact with the analysis. The web application is designed to make the understanding of results as intuitive as possible. The analyst can set-up the engine, construct and manage data dictionaries within the system, select the appropriate statistical tests and perform the complete CSM process. An action follow-up/ tracking system module is also available in the web application. The company has been beta-testing the portal with current customers while obtaining feedback on preferred data visualization options.

We have documented appropriate training material since we believe that the biggest challenge for any analyst is to be appropriately coached in how to process the data to get the best out of the CluePoints solution. We provide videos for all the different features in the web application, and we also organize online and face-to-face training sessions with our customers.

We are continuously improving our web application thanks to our customers’ feedback. They allow us to find new requirements and define priorities that are executed using an Agile product development methodology.

What is the recommended frequency for analyzing data with CluePoints?

Ideally, the frequency of analyses of operational data would match traditional schedule for on-site monitoring with CluePoints being deployed every 6-8 weeks. We usually run analyses of the clinical data every three months, but this is dependent on the size and shape of study and its enrolment rate.

Some of the most powerful work that we have done in analyzing both clinical and operational data on behalf of sponsors is to determine where issues exist on an ongoing basis.

This allows operational teams to take corrective action, re-address the monitoring plan and to improve Key Risk Indicators associated with that particular study.

If these kinds of analyses are undertaken early in the study, and especially during the ramp-up phase, then the more effective it will be to make sure issues are ironed out quickly, particularly regarding site training and coaching.

How is the FDR Bubble Plot Calculated?

SMART™ calculates a score between 0 and 1 for each center. The smaller the value of the score, the greater the difference in the data identified at that specific center compared to all other centers. The score can be interpreted as a P-value (technically, it uses an aggregate of the P-values of all tests comparing a center to all other centers).

bubble_plot_website_small

Is CluePoints using Bayesian type analysis to do real-time type analysis of the data in the process to be able to show sites drifting into risk territory?

There are some fairly sophisticated statistical modules but nothing Bayesian as of yet.  However, we do run CluePoints repeatedly in the same study and analyze the incremental data. A center that may have been normal at one run can become abnormal or outlying in the next run, and that is quite an important finding.

How many studies has this solution been used on?

CluePoints has been utilized in more than sixty studies since the company’s inception in October 2012. The SMART engine took over a decade to develop and has analysed many studies in the R&D phase. The first publication CluePoints released about the tecniques used in SMART date back to 1999 (Buyse M, George SL, Evans S, et al., for the ISCB Subcommittee on Fraud). The role of biostatistics in the prevention, detection and treatment of fraud in clinical trials. Stat Med 18: 3435-3452, 1999).