Closing the Variance Gap: The Challenges with Clinical Trial Budget Management and Forecasting

May 04, 2015

As clinical trials continue to grow in complexity, tolerance for the growing variance between forecasted and actual clinical trial costs is shrinking.

To set the stage, let’s take a look at what is considered an acceptable variance in today’s clinical trial scenario:  A recent survey from Clinverse showed 38% of sponsors will accept a variance of <5% and 31% will accept a current variance of 5-10%. Yet, remarkably, the variance between forecasted and actual clinical trial costs for life science companies can be as high as 16%, according to an industry survey.1 With pressure on earnings, financial predictability and accurate reporting, a high variance is less tolerable today as it affects revenue and future R&D budgeting and cash flow for all involved. Inaccurate accruals can also have an effect on earnings in the next reporting period.

What are the top causes of high variances?

1.    Outsourcing. The estimated average cost of bringing a drug to market in the U.S. is $2.6 billion, according to a recent Tufts CDSS report. In a similar study published in 2003, Tufts CDSS estimated the cost to be $802 million, equal to $1.04 billion in 2013 dollars, which indicates a 145% increase to develop and win marketing approval for a new drug.2 That said, clinical trial
costs are typically one of the biggest expense categories for biopharmaceutical companies. Due to the dynamic nature of clinical trials today, sponsors have opted to move fixed expense to variable expense, which has been accomplished by outsourcing parts of trial management to multiple vendors.

A major influence on forecasting has been the increase in outsourcing clinical trials
from 20% in 2012 to 41% in 2014, according to a Nice Insight survey. Large pharmaceutical companies have the highest rate of outsourcing at 46%, and emerging pharma showed the lowest incidence at 36%.3 The trend in outsourcing more and more trial management activities to multiple vendors has not only resulted in the lack
of effective financial management systems, but it has also driven a wedge in having access to a complete financial picture in a single consolidate manner.  Multiple vendors and contracts end up in disparate systems, and the lack of integration doesn’t allow that data to be brought together and used for financial management in an easy or efficient way, making forecasting and budgeting extremely challenging.

2.    Delays. More than 80% of clinical trials experience delays range on average from one to six months, costing companies upwards of $35,000 per day per trial. A mere 10% of trials are completed on time.4 Time delays generate significant variability in clinical development budgets and can add substantial costs. Not surprisingly, only 14% of clinical financial planners at pharmaceutical companies are highly confident in their budget forecasts.

3.    Complexity of trials. Another cause of forecasting difficulty is the expansion in number, size, length and complexity of clinical trials. In fact, as of Dec. 22, 2014, the current number of registered studies on is 181,107 with locations in all 50 States and in 187 countries.5 Over the past two decades, the average length of a clinical trial increased 70%, the average number of routine procedures per trial has risen 65%, and the average clinical trial staff work burden increased 67%.5 Trials today often undergo protocol amendments, which can add new trial populations, extension arms, increased assessments, and other design modifications.

In addition, due to the rapidly changing nature of clinical trials, forecasters may not be able to use historical data to accurately predict expenses for future clinical trials. The trend toward adaptive trial design, where the trial can be modified during its progress based on interim results, makes forecasting difficult.

4.    Unexpected events. Forecasters are also challenged by unexpected events such as the impact of slower or faster site activation, lags in enrollment activity, and underperforming sites. When sites do not enroll enough subjects, it may be necessary to add more sites across multiple geographical regions and to close sites early due to non-performance. On top of that, many trials have significant funds tied up in procedural costs that may or may not occur in that particular trial, and it can be hard to predict since it isn’t known if patients will progress to a certain stage or if additional procedures will need to happen. And with forecasting often done for the worst-case scenario (i.e., every patient completes every procedure and visit), it may result in not having the ability to release funds early. This can be prohibitive to other R&D efforts since those funds can’t be allocated to other studies or efforts that may be underfunded. Since site start-up costs and protocol design drive the significant portion of study expense, these changes drastically impact the forecast and expense requirements.

5.    Globalization. In the past decade, there has been a major increase in the globalization of trials, with multiple countries using their own cobbled-together financial management systems and managing expense in their own local currencies.

Its also important to note that forecasting typically falls to the clinical operations team, which means the manual and labor-intensive task of forecasting to a team who is already managing trial execution, data collection, site relationships, and a myriad of other service responsibilities. Now they must also be responsible for tracking, evaluating, reconciling and more accurately reforecasting future expense needs based on study information, which likely is not real time or even near it. In addition, these tasks require pulling data from multiple systems and analyzing it in cumbersome excel worksheets, which is a laborious effort that takes away critical time from trial execution.

As a result, the life science industry is struggling to forecast and reforecast present and future trial expense accurately and efficiently to allow for real-time decision making and risk mitigation to an already expensive effort, especially with spreadsheets continuing to be a predominant method. In Clinverse’s survey, a whopping 70% reported the primary tool used for budgeting and forecasting at their company was Microsoft Excel. Spreadsheets are cumbersome to share and consolidate with other financial forecasting and budgeting data, do not help forecast subject and site activation, take too long to update, and are prone to error. Using this rigid, time-consuming method, it is difficult to get a clear, consolidated view across all sites and protocols.

Clearly, the industry needs a better approach to forecast expenses and easily track the financial progress of a trial as compared to the original budget. The necessary solution is a robust, purpose-built, operational-based tool designed around a core site and subject-forecasting engine to achieve timely, accurate forecasting.

The ideal system can be customized by both organization and trial with unique user-definable variables to forecast sites, subjects, using dynamic dates, and projected expense amounts. Such a system would enable financial managers to calculate expense and cash forecasts as well as have a view into operational performance and changes that drive this expense.   Further, the system can generate an expense forecast, let users see when cash will be needed, and designate a primary budget to compare and report against actual results.


1. Grygiel A. The Struggle With Clinical Study Budgeting. Contract Pharma. Oct. 11, 2011. Accessible at:

2. Cost to Develop and Win Marketing Approval for a New Drug Is $2.6 Billion. November 18, 2914. Accessible at:

3. Hammeke K. Can CROs Help Reduce the Expense of Clinical Trials? Clinical Leader. April 29, 2014. Accessible at: help-reduce-the-expense-of-clinical-trials-0001.

4. Clinical Trials Forecasting for Finance Professionals. July 2011. Accessible at:

5. Active trials as of November 2014. Accessible at: ls+2014&Search=Search resources/trends#RegisteredStudiesOverTime.

lorem ipsum