When making decisions in the professional context it is often important to ensure these are risk-based. This means using a proven method for deciding whether to do something or not to do something; and where we do something, how we prioritise the order in which tasks are to be completed. Here it is important to emphasise the risk component. This means understanding the potential negative outcomes (risk) is fundamental to making informed choices.
In a previous RSSL article 1 we have examined the complexities of risk management (including the issues created by uncertainty). In this article, we explore the uncertainty concept further in the risk-based decision making context, with the aim of aiding readers when they are reflecting on their own approaches to the management of risk.
Decision-making can be defined as the process of choosing between different options by identifying a decision, gathering information, and assessing alternatives to reach a solution. A common framework involves seven steps:
Step 1: Identify the decision. Clearly define the problem or decision you need to make.
Step 2: Gather relevant information. Collect necessary information from both internal self-assessment and external sources like books, the internet, or other people.
Step 3: Identify the alternatives. List all possible courses of action.
Step 4: Weigh the evidence. Evaluate the pros and cons of each alternative.
Step 5: Choose among alternatives. Select the best option based on your evaluation.
Step 6: Take action. Implement your chosen alternative.
Step 7: Review your decision. Reflect on the outcome to see if it was successful.
Such a systematic approach can lead to more thoughtful and effective choices.
This is a case of ‘so far, so good’. However, a combination of factors will influence the process. We can separate these into:
While the above can apply to everyday circumstances, in professional management within life sciences we are often called upon to make ‘risk based’ decisions. What does this mean?
Risk-based decision-making is also a systematic process. The difference is that it includes the prioritising of choices by considering their potential risks and benefits.
This involves identifying issues, examining options, assessing risks and uncertainties, and implementing the chosen solution. The objective is to make informed, strategic choices that optimise outcomes by understanding the risk-reward balance.
This type of uncertainty arises from a lack of knowledge about the system or process being assessed. This can often be mitigated by gathering more data or conducting further research. Examples are [4]:
To address conflicting opinions, approaches like fuzzy-logic can be beneficial [5].
This form of uncertainty refers to the natural randomness and variability within a system that cannot be reduced, even with more knowledge [6]. An example of this is natural variability: Processes that are naturally random, such as weather patterns or the behaviour of individuals, cannot be fully predicted or controlled. For example, the body weights of individuals in a population will vary naturally, and this variability cannot be eliminated.
Data and parameter estimates can have a degree of uncertainty due to random errors in measurement or sampling techniques (such as the use of imprecise monitor instruments or the choice of a less-precise technique) or systemic biases in measurements (such as estimates that are reported consistently without acknowledging the nature or extent of the estimation). Estimates can also include uncertainty due to use of surrogate data, misclassifications, or random sampling errors.
Examples of how we can end up with data uncertainty include:
Importantly, data uncertainty differs to data variation. The latter refers to the inherent heterogeneity or diversity of data in an assessment. This is often expressed through statistical metrics such as variance, standard deviation, and interquartile ranges that reflect the variability of the data. With data variability we cannot reduce it down any further, but we can often characterise it better. Whereas data uncertainty is a property that ca be improved or eliminated with better or more data [7].
Examples of where and how uncertainty can impact on our work include:
Hazard identification is essential for many risk assessment tools, such as FMEA and HACCP. Uncertainty can arise when it is unclear whether a particular agent or event is actually a hazard and what adverse effects it may cause 8.
The uncertainty can be quantified (such as “the sample contains 100 microorganisms of which 10 cells are of a Bacillus bacterium”) or characterised qualitatively, for example by verbal phrases such as “the substance is probably contaminated with Bacillus bacteria”. The former approach is generally more accurate since quantifying uncertainty, say on a percentage scale, is more effective because it reduces the room for ambiguity.
This can lead to uncertainty in risk characterisation (such as ‘how high is the risk?’) and uncertainty in risk management (for example, ‘is the proposed risk protection sufficient?’).
In clinical trials, for example, it can be difficult to determine the relationship between an exposure and a response, especially when extrapolating from high-dose animal studies to low-dose human exposures 9.
When developing a community wide medicine, such as a vaccine, then estimating the type, magnitude, and duration of exposure for a population is challenging, particularly due to variable human behaviours and environmental conditions.
The entire risk assessment process may be affected by uncertainties present in its constituent steps, resulting in an overall uncertainty in the final risk estimate. This means that the existence of a hazard is proven (hazard identification) but the magnitude of the associated risk cannot be unambiguously determined 10.
The most widespread tool for quantifying uncertainties is the mathematical concept of probability 11. This can be useful, but it is not without its issues (which can be discussed in another article). Probability can be thought of as a means of limiting frequency. In other words, reducing the range within a data set that creates uncertainty.
This article has touched on the concept of uncertainty when we conduct risk assessments. By acknowledging and seeking to reduce uncertainty we can produce better risk management processes. Hence, uncertainty is a crucial issue for any risk assessment. It is also important for risk communication. Here it is important that we report where we have uncertainties in our risk assessments. Such communication needs to be clear and we need to be mindful that our audience may not fully understand what our uncertainty disclosure actually means.
1. Sandle, T. (2024) Walking the tightrope: Uncertainty and quality risk management, RRSL Life Sciences https://www.rssl.com/insights/life-science-pharmaceuticals/walking-the-tightrope-uncertainty-and-quality-risk-management/
2. Wiedemann P, Boerner FU, Freudenstein F. Effects of communicating uncertainty descriptions in hazard identification, risk characterization, and risk protection. PLoS One. 2021 Jul 13;16(7):e0253762
3. Gerrard, B. (2022). The Road Less Travelled: Keynes and Knight on Probability and Uncertainty. Review of Political Economy, 36(3), 1253–1278. https://doi.org/10.1080/09538259.2022.2114291
4. Fox CR, Ülkümen G. Distinguishing Two Dimensions of Uncertainty. In: Brun W, Kirkebøen G, Montgomery H. editors. Essays in Judgment and Decision Making. Oslo: 2011
5. Sandle, T. (2020) Can 'Fuzzy Logic' Be Applied To Risk Management In Pharmaceuticals And Healthcare?, IVT Network: https://www.researchgate.net/publication/346972390_Can_%27Fuzzy_Logic%27_Be_Applied_To_Risk_Management_In_Pharmaceuticals_And_Healthcare
6. Ülkümen G, Fox CR, Malle B. Two dimensions of subjective uncertainty: Clues from natural language. Journal of Experimental Psychology: General. 2016;145(10):1280–1297
7. US NRC (National Research Council) (1994) Science and Judgment in Risk Assessment. Washington, DC: National Academy Press.
8. International Agency for Research on Cancer. Non-Ionizing Radiation, Part 2: Radiofrequency Electromagnetic Fields, IARC Monograph on the Evaluation of Carcinogenic Risks to Humans. Lyon: International Agency for Research on Cancer; 2013
9. Committee on Decision Making Under Uncertainty; Board on Population Health and Public Health Practice; Institute of Medicine. Environmental Decisions in the Face of Uncertainty. Washington (DC): National Academies Press (US); 2013 May 20. 2, Risk Assessment and Uncertainty
10. Han PKJ, Klein WMP, Lehman TC, Massett H, Lee SC, Freedman AN. Laypersons’ Responses to the Communication of Uncertainty Regarding Cancer Risk Estimates. Medical Decision Making. 2009;29(3):391–403
11. Parry, G.W., Uncertainty in PRA and its Implications for use in Risk-Informed Decision Making. Proceedings of the 4th International Conference on Probabilistic Safety Assessment and Management, PSAM 4, Edited by Mosleh, A. & Bari, R.A., New York, 1998
