Random Acts of Medicine, written by physician and policy researcher Christopher Worsham and physician-economist Anupam (Bapu) Jena, explores a range of fascinating case studies detailing the authors’ natural experiments revealing the hidden forces that sway doctors, impact patients, and shape our health. In natural experiments, allocation to conditions is not controlled by researchers – but is expected to be unrelated to other factors that cause the outcome of interest. This means allocation can be assumed to resemble random assignment and causal inference can be inferred. Natural experiments are useful because they enable the evaluation of events or interventions that are difficult or impossible to manipulate experimentally – for example, exposure to a health event which may have detrimental effects. In this blog, I’ll outline a few examples from the book which reveal real world examples of behavioral heuristics and biases influencing decision making and health outcomes. Why are kids with summer birthdays more likely to get flu? Hypothesis: Children with birthdays in summer months, will be less likely to get flu vaccinations than those with fall birthdays. As young children tend to have annual checkups around their birthdays, those with birthdays in the latter months of the year are more likely to be vaccinated, as they are likely to be routinely offered the vaccine at their annual check-ups. Parents of summer babies on the other hand would have to actively schedule another appointment later in the year to receive a flu shot. Experimental Factor: Birth Month There is no inherent reason why a child born in March, for example, would be more or less likely to receive a flu shot than a child born in any other month. Evidence: 2-3 year olds born in October had vaccination rates of 55% compared to 40% in those with May birthdays. Using large amounts of data from insurance claims, the experiment confirmed that that most 2-3-year-olds do undergo annual checkups in the month surrounding their birthday – and that the children born in fall months did have significantly higher vaccination rates compared to those born in other months. These kids (and their families) were also less likely to get flu than those with summer birthdays. The Bias: Logistical Friction The authors discuss the 3 C’s of vaccine hesitancy. These include Complacency, perceiving the risk to be low, and Confidence, trusting in the efficacy of vaccines. This study shows the importance of the third C, convenience. As humans, we prefer to take the simplest path; the more steps something has, the less likely we are to do it. These hassle factors are commonly referred to as logistical friction. In this case, the additional logistical friction of scheduling and attending appointments significantly reduced vaccination and subsequent flu infection rates in young children with summer birthdays. The solution: Make it easy! Logistical friction can be overcome by simplifying the vaccination process – for example, simplifying the process of booking and scheduling an appointment or bringing mobile vaccination centers to workplaces or day-cares. Another way to make an option appear easier is to reframe it as the default option. For example, in a different experiment, one group of adults had vaccination appointments made for them, and had to cancel if they didn’t want to go (opt-out group), while the other received information on how to make an appointment and had to book it themselves (opt-in group). The vaccination rate in the easier ‘opt out’ group was 45%, compared to just 33% in the opt-in group. Hypothesis: Mortality rates following heart attacks will be higher during the dates of cardiology conferences. During conferences, the authors hypothesised that the leading cardiologists would be away from the hospital, and anyone presenting with heart failure may receive poorer or slower care. Experimental Factor: Cardiology conferences and heart attacks There is no inherent reason why individuals having a heart attack during a cardiology conference will be significantly different from those attending on non-conference days. Evidence: The thirty-day mortality rate actually lowered during cardiology conferences – Using data from around 6,000 patients experiencing heart failure or cardiac arrest during or just before/after cardiology conferences, the authors found that, contrary to the hypothesis, mortality rates were actually lower during cardiology conferences (17%) than on non-conference days (25%). The Bias: Action Bias/ Commision Bias The study identified a crucial factor—the increased use of cardiac stenting. On conference days, the utilization of cardiac stenting was lower (20.8%) compared to non-conference days (28.2%), When the leading cardiologists were in town, they were occasionally performing unnecessary procedures, which could actually be contributing to increased mortality. This is reflective of an action bias – our innate tendency to prefer action (e.g. perform a surgery or prescribe a medication), over inaction (doing nothing). They note various other studies have found similar results, for example decreased mortality rates during physician labour strikes was also attributed to lower rates of non-urgent surgical procedures taking place. Does a child’s birthday influence their likelihood of ADHD diagnosis? Hypothesis: Children who are younger for their school year, are more likely to receive an ADHD diagnosis. Children who are the youngest in a school year, can be up to 364 days younger than those who are the oldest. Traits associated with ADHD, such as inattention, hyperactivity, and impulsivity, often align with behaviors exhibited by younger children (e.g. not being able to sit still or follow orders). Therefore the researchers hypothesised that teachers may mistake these characteristics as ADHD – and that children who are young for their year group may be over-diagnosed. Experimental Factor: Birth Month Researchers have uncovered a phenomenon known as the “birthday effect.” This effect manifests when the cutoff for school year enrolment creates age differences among students, leading to substantial disparities in abilities and outcomes in fields such as academia and sports. For example, in one analysis of the national hockey league players, 63% were born between January and June versus 38% born between July and December. Evidence: In states with a September 1st cutoff for kindergarten enrollment, children born in August had a 34% higher rate of ADHD diagnosis compared to their September-born peers Additional analyses revealed there to be no significant differences were observed between other months (e.g. children born in July versus August) proving the cut-off was important. The Bias: Representativeness heuristic This relative age effect is an example of the representativeness heuristic. This heuristic leads individuals to apply expectations to a category (e.g., kindergartners) without considering significant age gaps among them. So what have we learnt – heuristics and biases can influence our decision making and affect our health outcomes (sometimes for the worse!) But what can we do about it? Education: Simply raising awareness of biases and heuristics can help reduce their influence on decision making Debiasing techniques and cognitive forcing strategies: Forcing individuals to slow down and make more considered, deliberate choices – for example, asking someone to consider how a certain situation might look from another person’s perspective, rehearsing potential cases ahead of time to make decisions in a less ‘hot state’ or introducing a friction point to encourage the decision maker to stop, reevaluate thinking and actively consider and reject all options before making a final choice Tools and Guidelines: For example, using objective risk calculation tools, or checklists outlining decision steps to avoid rash choices and add objectivity and structure to decision making. Technology and AI can further enhance these tools – for example automating ‘nudges’ (e.g. reminder) or using algorithms to support decision making For more fascinating examples, and deeper exploration of solutions, I highly recommend giving this book a go! At HRW Shift, books like this help us continue to build our expertise of behavioural science applications in the healthcare space and ensures we are applying the latest evidence within our tailored recommendations. Apply Now!