Machine Learning and Natural Language Processing for Cost Analysis - International Cost Estimating & Analysis (ICEAA) Workshop May 2019

How can machine learning and natural language processing enhance the traditional methods of cost analysis? By making it easier for humans to analyze large datasets, and by automating many of the manual, time-consuming tasks needed to cleanse and prepare the data for analysis.

This presentation examines several applications in which the authors integrated Natural Language Processing using python libraries, with various Machine Learning methods, in particular trees, randomization, and boosting, to improve prediction accuracy.

Quick Learning Resource: Random Forest - Fun and Easy Machine Learning

Here’s a great video tutorial by Augmented Startups. Please feel free to support them on Patreon.

Decision tree is a type of supervised learning algorithm (having a pre-defined target variable) that is mostly used in classification problems. A tree has many analogies in real life, and turns out that it has influenced a wide area of machine learning, covering both classification and regression (CART). So a decision tree is a flow-chart-like structure, where each internal node denotes a test on an attribute, each branch represents the outcome of a test, and each leaf (or terminal) node holds a class label. The topmost node in a tree is the root node.

TrialValue A.I.+ Budget Prediction Model Workshop: Presentation Demo is Out!

Thanks to everyone who attended last week’s public online workshop where we presented use cases and demos on how our budget prediction and benchmarking model work.

We showed that you only need a few key study assumptions to predict an optimal budget with value assured confidence in less than 10 seconds!

Pharma and biotech folks can use the model for long term financial planning, benchmark proposal estimates, calculate savings, budget oversight/audit, cost control, project valuation, financial modeling and more.

The recorded demo is available on request or please contact us to arrange a training session for your team:

Screen Shot 2019-02-26 at 2.17.25 PM.png

TrialValue Analyst

We are looking for a financial analyst to join us on a new project focused on economic evaluations in clinical research. Do you have hands-on experience in this area? Can you play chords with budget numbers? Then we can’t wait to hear you play in our “add-value” jam sessions?


fancy being our instrumental analyst for a new gig in 2019?

Main tasks

Financial Modeling

develop financial and pricing models based on clinical research protocols, economics evaluation objectives/strategies, study plans, trial design scenarios using TrialValueapp, A.I.+ models, external data sources and industry applications

Cost Estimation and Budget Development

run simulations, review cost estimates, develop budgets and financial plans for clinical research projects and programs

Financial Benchmarking

use TrialValue tools to validate and benchmark project staffing estimate (FTE allocations), tasks/deliverables pricing, total budgets, client’s internal ballpark estimates and CROs’ proposals


  • Location: UK, Netherlands, Spain, US or Remote

  • Duration: 3 to 6 months (Part-Time/Flexible schedule)

  • Start date: January 2019

  • Experience and Skills:
    Biotech/Pharmaceutical industry, Clinical research, Clinical trials, Budgeting, Financial/Cost modeling, Health economics, Statistical data analysis and Creative problem solver

    Interested? Please drop us a note to arrange a chat:

TrialValue A.I.+ Launch Workshop

The inaugural workshop for #TrialValue A.I.+ is scheduled for Thursday 8 November 2018 (Utrecht inc, Netherlands). All TrialValue community partners and PRO members are welcome to attend. Please contact us to inquire about passes: Further details on times and agenda will be posted in the coming days.

Screen Shot 2018-10-08 at 5.02.40 PM.png

TrialValueapp: Updated version released.

Happy to announce that the new cloud version of TrialValueapp is released. 

Main changes:
- Updates to user interface and navigation improvements
- Minor bug fixes

Many thanks to the team, our user community, colleagues and friends for all the support and helpful feedback.  The result is a better user experience and intuitive navigation.

Outsourcing and Budget Management Course: TrialValueapp partners with CILIQUE

April 2018: TrialValueapp and CILIQUE, have teamed up as training partners for CILIQUE's on-line course on Outsourcing and Budget Management in Clinical Trials: This course is the first of its kind in the industry and offers a comprehensive learning opportunity for colleagues involved in the business of clinical research. A short introduction module on how to use TrialValueapp to estimate and develop a benchmark budgets is included in the course. Check it out. (Disclosure: we will receive a commission if you sign-up). 

2017 Budget Benchmark Challenge Results

Thank you to everyone that participated in our first Budget Benchmark challenge. The challenge closed in April 2018. However, the turn out was low as we received only 9 submissions of planned costs. In view of this, the data collected wasn't sufficient to draw any meaningful conclusions or comparisons.

TrialValue's benchmark algorithm generated median per-patient-budget (all costs) for phase III studies ranged from $34,500 to $68,000 factoring various complexity scenarios and assumptions. (Data: >100K study budgets, compiled from published data, meta-analysis, simulations and expert knowledge).

Why the low rate of participation? This could be due to a number of reasons: including low publicity (challenge was only promoted on social media) and perhaps it didn't reach the right audience. Also, based on the benchmarking roundtable discussions  during this year's CBI clinical trial budget conference it appears that many pharma companies don't typically conduct analysis of final study budgets (actuals) against historical costs or compare with industry available data. Benchmarking is mainly done to obtain per-patient-cost and procedure costs at a study and site level for contract negotiations and to support FMV compliance process.

It was also mentioned that benchmarking is difficult to implement because of lack of standardization of budget components, time/resource constraints, variable study complexity and continuous changes to R&D portfolio assumptions.

Now the good news: To help overcome some of these hurdles, we're fine tuning our machine learning enabled algorithms TrialValue A.I.+ we expect to start roll out in a few weeks.