Boyd cycle

11. october 2015 at 9:40 | Veronika Valdova, ARETE-ZOE

To make any system oriented toward desired outcome, systems approach is necessary to make the feedback mechanism work. Boyd's OODA loop is an elegantly simple grand theory, which contains high quality insight into strategic essentials. OODA loop has extensive domain of applicability, including clinical research and post-market drug surveillance.

Boyd cycle, or OODA loop, explains fundamental principles of decision-making loop based on observation, orientation, decision, and action. Author of the concept, Col. John Boyd, was a fighter pilot and author of designs of F-15 and F-16, who challenged Air Force orthodoxy at the heart of the service's very identity. In his 15-hour briefing "Discourse of Winning and Losing" Col. Boyd challenged the theory how wars were to be fought and won in the era after Vietnam. In Vietnam, the 10 to 1 kill ratio from the Korean War came close to 1 to 1. Boyd's theory redefined tactical air operations after almost 20 years under bomber generals who grew on Curtis LeMay.

The scenario described above shows some analogies with the current situation in the clinical research and pharmaceutical industries. Only about 2 out of 10 marketed drugs return revenues that match or exceed R&D costs. Whilst about 2/3 of clinical trials are still conducted in the U.S., Western Europe, and Australia, globalization has led to outsourcing and off-shoring critical operations overseas to decrease costs and reduce liabilities. The industry found itself in a quagmire of ever increasing costs and low returns.

OBSERVATION: known and observable FACTS (systematic collection of information)
1) Outside information: Proper background research is essential to identify potential critical points which can trigger additional expenses necessary to overcome obstacles or lead to halt of a trial. Critically important information must be "observable" and "detectable" in real time to allow processing. Obscuring information in the collection phase disrupts the OODA loop at the very beginning.
2) Unfolding circumstances: Efficacy and safety profile of a new drug, changing regulatory environment, but also organizational culture affect information coming in.
èFeed observations forward for orientation.

ORIENTATION: Information processing
1) Cultural traditions: Outsourcing and off-shoring brings different cultural traditions. The same information in exactly the same context can be evaluated very differently depending on the person's educational, cultural and personal background. Risk-perception or communication of project issues may be influenced by cultural traditions.
2) Genetic heritage: Nature and nurture affects our ability to detect and communicate potentially critical safety problems before they become self-evident. Different types of intelligence play a role - recognition of patterns, risk-assessment, as well as ability to communicate the message to the leadership. Distribution of personality types in pharma industry shall copy distribution in other high-risk industries to achieve similar results. In pharma, however, teams are often built to avoid personality clashes and ensure smooth function of office environment. System thinkers, who have the capacity to identify systemic flaws, may find it difficult to thrive in such environment.
3) Previous experience: Education, training, personal values, and individual biases affect reasoning and perception of risk. In clinical research, evidence comes from very diverse sources: study findings, observations by investigators, literature screening, colleagues and conference participants, imaging technologies, medical records, and patient feedback. Each source has its unique strengths, weaknesses, potential or actual biases, and vulnerability to manipulation and deception. Systematic work with meaning of the information is essential to accurately judge whether the drug candidate can successfully enter the market.
4) New information: Unexpected surprises can emerge at any stage of a trial or after drug approval. Boyd's OODA loop was designed to present decision-making and behavior of a single individual, well-coordinated team, or an organization with defined structure which behaves in a coordinated manner. Multiplicity of stakeholders within pharma may negatively affect information flow due to their at times conflicting interests. The need for accurate judgment is the same for a fighter pilot and for a pharmaceutical company. The only difference is, that the first will face consequences of a bad decision very fast, whilst the latter with long delay; the principle, however, remains exactly the same. For sound and timely decision-making it is necessary to process information in a structured and coherent manner, and to pass it on along with correct analysis including assessment of confidence levels.
5) Analysis & synthesis: Awareness and conscious examination of own biases and motivations facilitate self-correction in scientific judgement. If some national regulators do not include certain parts of the drug development process in their scrutiny, i.e. design of a trial, delegate certain tasks to the ethics committees without providing appropriate oversight, or if they fail to search for signs of scientific fraud and only focus on speed of certain procedures of interest, they effectively resign on their role in the system.

The process of ORIENTATION provides feedback back to OBSERVATION to alter data collection methods or forms in order to provide the right information for processing. More information does not necessarily mean better and more accurate judgment. This applies to intelligence as well as medicine. In horse racing selection of the right criteria and omission of noise leads to accurate judgment and statistically more successful prediction than indiscriminate assessment of numerous factors without providing value to them. In drug safety a good example of inclusion of "noise" in assessments is the screening of social networks rather than reliance on hard data.

In clinical research there is no structured way of documenting correctness of decisions. At the highest levels the most important indicator of correctness of a decision is amount of money earned on a particular product. The time lapse between preclinical phase, clinical phase, market approval and launch and capitalization on the investments typically exceeds time an average CEO spends in the office. Because of the long gap between initial investment, and series of decisions that follow until the product enters the market, and capitalization, the moment of truth, there is very little incentive to make any unpopular decisions on a project if the feedback (evaluation whether the decision was correct or not) is too far in the future to have any impact on an individual. Hence it is very easy to imagine how non-existence of a real-time real-life feedback leads to fundamentally flawed decision-making process by design. In biology feedback mechanism matters only if the organism can react in real-time. Postprandial insulin meaningfully decreases glucose levels if released in bloodstream in minutes-hours, not if the feedback is delayed long enough to become irrelevant.

Correctness of decision depends on accuracy and timing of information obtained, on its amount, and its form. Decision may lead to test of a hypothesis or directly to action. Whilst it would make sense to halt a trial as early as possible should the study prove non-viable, it does not always happen, and more resources are thrown at a project before it finally reaches a point when it has to be stopped. These delays are caused by the fact that all critical activities of OODA loop are not performed by a single well-coordinated entity but by numerous independent business units and individuals with often conflicting interests. Because the model is dynamic, new information comes in all the time.

Decision-making centers need to adjust information which is being gathered and how this is done. In clinical research and drug safety decisions are made based on data from clinical trials and post-market surveillance. Before the drug enters the market, the total number of patients exposed to the drug is relatively small, the population is well defined, and the frequency of any adverse events can be relatively easily established. After approval efficacy, unlike safety is no longer actively monitored. Post-market surveillance conducted via active reporting cannot provide reliable information on trends, populations exposed, and indications including off-label use. Due to disruption in observation part of the Boyd cycle, some kinds of safety information become invisible after drug launch. This can be resolved by making links between exposure to a substance and clinical outcome observable. To do that it would be necessary to make analytical outputs from real-life use available in real-time in a form which our brains can process.

Action is relevant if it follows decision which has been made. If nobody acts on decision made in real time the whole process of surveillance is fruitless and only documents activity (due diligence) without having an actual impact. Functionality of a system must not fall victim to observation of its outer attributes which document "activities" without substantive action.

New comment

Log in
  Don't you have your own web yet? Create it for free on

Actual articles