Dicing with design

Dicing with design

Dicing with design 558 344 Exploristics

By Aiden Flynn, Exploristics CEO

Recently, I visited the historic city of Stockholm in Sweden. While there I had the chance to explore the Vasa Museum, where the wooden galleon the Vasa is displayed. Costing the equivalent of 5% of Swedish Gross National Product to build, this remarkably preserved ship sank to bottom of Stockholm harbour on its maiden voyage in 1628. Since being raised in 1961, engineers and historians have investigated the causes of its premature sinking. The story of its build, demise and salvage are presented at the museum. Here, I was struck by the common mistakes that can be made when designing a project whether it’s a warship or a clinical trial.


The 17th century galleon ‘The Vasa’ at the Vasa Museum in Stockholm, Sweden.

What went wrong?

Multiple investigations have found that four main factors probably contributed to its failure:

  1. Changing specifications

The ship was originally ordered by King Gustav II Adolph of Sweden in January 1625. Sweden was involved in the Thirty Years War and needed more ships for its Navy. So, the King asked for four to be built over the next four years, including the Vasa. In the ensuing months, he would change his order several times.

  1. Innovating under unreasonable time pressure

In September 1625, the Swedish Navy lost ten ships in battle, so the ship building schedule was accelerated to compensate. The King now also wished to incorporate a second gun deck. The Vasa was originally designed to have one gun-deck, but the Danish were building ships with two. More guns could be decisive in battle but incorporating them involved design trade-offs. Double-gun decks were cutting-edge warship design so there were no standard plans to consult for this complex modification. Under time-pressure, the ship’s dimensions were simply scaled up to accommodate their size. This resulted in the vessel becoming much wider at the top than the bottom than originally planned, raising its centre of gravity.

  1. Lack of documentation or a coherent project plan

A year before the new completion date for the Vasa, the senior ship designer died. He had not documented his plans for the amended design and there were no detailed project specifications. At the time, five teams of shipwrights were building different segments of the hull. There was no intercommunication between them.

  1. Lack of scientific methods and reasoning

On its completion, the ship was tested in Stockholm harbour and found to be highly unstable. In 1628 there were no known scientific methods for measuring stability, centre of gravity and toppling. It was common for sea captains to use trial and error to optimise these factors. No one knew how to stabilise the Vasa given its innovative design and so under time pressure the ship was launched.

Outcome and recovery

In August 1628, the Vasa sailed from Stockholm harbour with its gun portals open. It had travelled less than two nautical miles when a gust of wind caught its sails, causing it to topple. Water poured in through the open portals, and the ship sank resulting in 53 deaths. During the inquest details of the poor stability were revealed. Recovery attempts were made but without success.

In 1956 the wreck was rediscovered initiating a complex and costly recovery mission lasting five years. Rescuing the wreck involved one year evaluating and planning potential recovery methods, two years digging tunnels under the ship for steel cables to attach to lifting pontoons, and two years lifting. The salvage operation took more than 1,300 dives.

Designing better

The Vasa disaster demonstrates how projects can fail spectacularly without a rigorous design approach. In many sectors opportunities remain to learn from historical mistakes and improve planning. As a statistician, I find similar problems to those that affected the Vasa can still afflict clinical study design today.

1.Changing specifications compromise the integrity of the study and result in complex and confusing protocols. Specifications need to be clarified at the outset so that all stakeholders understand the requirements. Simply scaling up studies or using previous studies as a template is rarely appropriate. A larger study does not always mean a more valuable one. A smaller study designed to collect the right data, in the right patients, in the right way can often provide more tangible insights. Whether a study is large or small, the stated objectives should be simple, and the design and study procedures should be completely aligned with them. Too many studies try to answer too many questions and end up collecting messy data that provide no useful information.

2.Unreasonable time pressure to start a study also impacts its likelihood of success. There is often a focus on achieving key milestones, such as first patient first visit, which prevents the evaluation of how robust the study is in different scenarios. Taking a little more time to understand the risks and design trade-offs involved ensures the right design features are selected and the objectives are more likely to be met. Investing time and money in design can identify cost-savings and operational efficiencies which more than cover the cost. It is ultimately a more efficient way to generate clinical evidence as it avoids the cost and delays associated with a failed study.

3.A coherent project plan is also vital, so that all aspects of the study can be reviewed, audited and understood by a multi-disciplinary group. Effective communication between all stakeholders at the planning stage means that a study is more likely to be designed that will work operationally, clinically and statistically, saving time, money and patient burden. Good documentation is the source of truth and allows new stakeholders to quickly understand the purpose, rationale and approach.

4.Statistical methods and tools are now widely available to those working in clinical trials that can help identify study risks upfront and mitigate for these in the design. Methods like simulation can investigate design trade-offs to pinpoint the key factors impacting success, so design can be optimised prospectively. Growing awareness and uptake of these as well as greater investment in planning can help build more success into development projects.

Bucking the trend

Visiting the Vasa highlighted the importance of a holistic design approach in delivering a successful project or study. Without one, a study might succeed but the trend in development attrition rates suggest this is unlikely. More likely, a study will fail or need to be salvaged costing time, money and resources. Why risk costly and unnecessary failures? With a little more patience, investment in better planning and use of technology, clinical projects can buck the trend and avoid the Vasa’s fate.


Read more:

Plague, problem and profiting from failure

Shifting the development paradigm with innovative trial design

Designing clinical studies for success and not failure

Statistical Consulting Services

Hear more:

The power of simulations for designing clinical studies and beyond

Simulations your most powerful study design tool

Watch more:

Why do correlations matter in clinical trial design?

Common causes of clinical trial failure

Can pharma learn from development in other industries?