Unveiling Shadows: Cognitive and Unconscious Bias in Retinoblastoma Research

Monday October 9, 2023

The human mind can subtly influence scientific research, with potentially serious consequences for patient care and outcomes. Rb survivor and WE C Hope CEO, Abby White, explores the nuanced world of cognitive and unconscious biases in retinoblastoma research, and strategies that can minimize their impact to ensure objective research and the best care possible for all.

The silhouette of a medical caduceus symbol features two snakes winding around a winged staff. A weighing dish hangs from each wing. The scales are in balance.

In the quest for knowledge and healing, we often view science and medicine as impartial, unswayed by the human mind. But even the most dedicated professionals have personal perceptions and tendencies that can subtly influence the course of research and patient care.

Many biases impact all aspects of retinoblastoma. They may influence every stage of research; diagnosis, treatment, and ongoing care decisions; and interactions between healthcare providers, patients and families, and within the wider retinoblastoma community. We cannot address all these concerns in a single article.

Here, we delve into the nuanced world of bias in research. We look at common types of bias, how they influence thinking and decision-making, how they can weave their way unchecked into the fabric of scientific investigation, and from there into medical practice. We use fictional examples to illustrate each bias.

Finally, we share practical strategies that medical professionals, scientists, parents, and survivors can use to reduce research bias. Individually and as a community, we can do much to combat these hidden influences, producing more objective research and compassionate care that focuses on the very best outcome possible.

Understanding Bias and its Origins

A bias is a prejudiced view or inclination that influences judgment and behaviour. It is an unbalanced weighing of factors or information, often stemming from personal experiences, beliefs, social norms, or stereotypes. Bias can consciously shape our decisions, or function outside our awareness.

Researchers have documented hundreds of different biases. Each is described as either a cognitive or unconscious bias.

Cognitive bias (also called explicit bias) refers to systematic errors in thinking that affect our decisions and judgments. Mental shortcuts can sometimes be useful to help us process information and make complex decisions, but they can also lead to faulty reasoning and conclusions.

With these biases, the individual is aware of their attitudes, opinions, thoughts, and feelings, and their related decisions and actions are deliberate. But how they perceive situations, people, and potential risks may be significantly distorted from reality, severely clouding their judgement.

Unconscious bias (also called implicit bias) refers to the attitudes, stereotypes, and opinions a person holds without being aware of them. They may contradict the person’s stated beliefs and values, and automatically influence their thoughts and feelings, perceptions, understanding, decisions, and actions.

These biases grow from influences like upbringing, media exposure, cultural and societal norms. They shape our attitudes and response to people based on characteristics like disability, ethnicity, gender, sexual orientation, age, or socio-economic status.

Biases are not inherently malicious. They are an outcome of our brain’s attempt to simplify information processing. However, when they enter sensitive arenas like medical research and care, the results can be serious. Let’s take a look at what can happen.

A woman wearing a lab coat, cap, gloves, and face mask looks intently at a test tube filled with liquid.

Cognitive Bias in Retinoblastoma Research

We are all likely to make systematic errors in thinking that affect our decisions and judgments. Here are some of the most common biases that can affect decision-making in retinoblastoma research, and the ways they affect patient care.

Confirmation Bias

The tendency to interpret new evidence as a confirmation of one’s existing beliefs or theories.

For example, a researcher might be more inclined to publish results that confirm their hypotheses while ignoring or discrediting contradictory findings.

Impact: validity of the research is threatened, limiting our understanding and progress in care.

Anchoring Bias

The individual depends too heavily on an initial piece of information they receive, impairing their ability to make subsequent judgments during decision making.

For example, a researcher is considering the direction of their future work. They read a study report that suggests a new drug has 70% success rate in shrinking tumours within the eye. They may anchor to this finding and pursue study of the treatment, neglecting more comprehensive studies that report lower success rates or concerning side effects.

Impact: Can result in over- or underestimating the effectiveness of treatment options, and affect the direction of further research or treatment plans.

Optimism Bias

The tendency to believe that we are less likely to experience a negative event.

For example, researchers might underestimate the challenges of access for blind and visually impaired survivors, and fail to design accessibility into the study. Researchers studying a new therapy might overestimate its potential for success, failing to recognise and acknowledge its potential risks.

Impact: Over-optimistic expectations can lead to poorly designed research, risky and unsuccessful treatments, and false hope in patients and families.

Framing Effect

The same person might draw different conclusions from the same information depending on how the information is presented.

For example, imagine two researchers, both seeking funding to study treatment for poor-response tumours. They reference their previous work that led to this proposal.

Researcher A describes how 80% of tumours shrank with initial treatment. Researcher B highlights the 20% of tumours that did not respond to treatment.

Both are presenting the same result, but researcher A might attract more funding because they present the positive, successful outcome – 80% of tumours shrank. This more appealing frame implies the new research has a higher likelihood of success, even though the focus will be different. Researcher B is more focused, and relevant, but highlighting the negative aspect of the previous study suggests a higher chance of failure in this new research – making it less attractive to grant funders and community donors.

Impact: May cause researchers, their funders, clinicians, and parents or patients to present or perceive information inappropriately for the context, potentially affecting decisions, and trends in research and patient management.

Representativeness Bias

The tendency to generalize a specific instance to a larger population.

For example, a researcher might study a small group of retinoblastoma patients who responded well to the therapy under investigation, then assume this therapy would work as effectively in all patients. However, the small sample may not represent the wider patient population – different stages of disease, genetic diversity, and other factors that may affect outcomes.

Impact: Can result in assumptions and misleading conclusions that can affect care decisions and future research.

Recency Bias

The tendency to emphasise more recent events over earlier ones.

For example, in research exploring the psychological impacts of Rb, the responses of a teenage or adult survivor or their parent are more likely to reflect on their recent experiences – those they recall most clearly, rather than those they relate directly to the childhood cancer experience.

Similarly, if recent studies report a shift in successful treatment, researchers and clinicians may emphasize this trend, even if the bulk of previous research suggests an alternative management approach remains more appropriate for some patients.

Impact: Can influence research findings and direction, and patient management trends, potentially overlooking established, effective treatments.

A multicultural group of young people walk together in a loose V shape through a forest, a little distance between each person. They are wearing back packs and looking at one another as they walk away from the camera.

Social Desirability Bias

A research participant gives responses they think will be received favourably by others. Most people want belonging and approval, which can lead them to present information in a way they believe will be socially acceptable.

For example, in a study exploring the psychological impacts of retinoblastoma, parents and survivors might underreport feelings like stress, anxiety, grief, guilt, fear of blindness, or experience of depression or trauma, to present themselves as stronger or more optimistic. This fits the pop-culture story of the cancer hero.

Impact: Both Recency Bias and Social Desirability Bias can lead to misleading data, inaccurate conclusions, and a lack of relevant retinoblastoma-specific supports and mental health care.

Hindsight Bias

The tendency to believe, after an outcome is already known, that one would have predicted or expected the outcome. This can distort the way people evaluate their decision-making processes.

For instance, a researcher may retrospectively think they “knew” a study would fail due to specific variables, causing them to reject similar research directions in the future, and potentially overlook important areas of investigation.

Impact: May lead to over-confidence in predicting a future course or outcome, potentially overlooking the unique factors of each individual situation.

Status Quo Bias

A resistance to change and preference for the current state of affairs.

For example, a research group has made steady progress on retinoblastoma over many years. They have refined approaches to planning and delivering research that they are familiar and comfortable with.

Now, a young researcher on the team is keen to advance patient-led research. Their proposal is backed by evidence showing that involving patients in research planning, design, and delivery improves study outcomes and patient care. But the scientists are hesitant. Adopting this new approach would involve learning new planning techniques, changing procedures they have used for years, and complex work with many more people. Despite the significant potential benefits to research and patient families, they decide to stick with what they know.

Impact: Can discourage everyone from updating or trying new approaches if they believe current methods and trends are acceptable, even if alternatives and additions are shown to enhance research, patient care, and outcomes.

Self-Serving Bias

The tendency to attribute success to one’s own efforts and skills, while blaming failures on external factors.

For example: a researcher might credit successful findings to their investigative skills or their expertise as a clinician-scientist, while attributing poor results to external factors like equipment failure, low quality samples, complex patient cases, or human error elsewhere in the team.

Impact: Can hinder self-improvement and learning from mistakes.

The Dunning-Kruger Effect

The tendency of people to overestimate their knowledge or skills when they lack proficiency in that subject. They don’t know enough to realize how little they know, or that what they know is incorrect.

For example, an ophthalmologist ventures into retinoblastoma research. After treating a small number of children for several years, with some success, they feel confident they understand the disease fully. They start making bold claims about a treatment method they have adapted, encouraging others to adopt it. But they don’t understand the complexity of retinoblastoma and how much they still have to learn.

Impact: Overconfidence can lead to misleading claims, wasting resources on ineffective treatments, giving false hope to patients and families, and potentially putting lives and eyes at risk.

A blue question mark hangs in the forefront of the image. In the background, a man in a lab coat holds the end of his stethoscope towards the camera to form the dot of the blue question mark.

Unconscious Bias in Retinoblastoma Research

Unconscious biases affect everyone, whether we want them to or not. Here are some of the most common that appear in retinoblastoma research, and how they impact patient care.

Affinity Bias

The tendency to seek out and get along with others who are like us.

For example, a researcher may collaborate primarily with people from the same discipline, area of expertise, institution, country, or cultural background, limiting the range of perspectives and knowledge that can benefit their work.

Impact: Hinders research innovation and global partnerships needed to solve complex challenges and advance patient care.

Gender Bias

The tendency to prefer or favour one gender over another.

For example, institutions, grant funders, research collaborators, and forums reporting on cancer research might favour studies led by one gender, or consider them more likely to produce positive outcomes. Traditionally, medical research has favoured men. Women are leaders and innovators in science, but this bias remains rampant.

Gender bias also occurs in research that does not ensure a balanced study population, or fails to separate sex and gender data to reveal important insights about each population.

Male lab mice are still favoured in pre-clinical research due to the prevailing assumption that female hormones cause variation that confuse the results – instead of thinking the variation might be relevant to study.

Impact: Unequal opportunities in research funding and authorship prevent important studies from taking place. Gender Imbalance among research subjects misses important insights that could influence patient care and outcomes.

Systematic Bias

Persistent errors that lead to inaccurate results. This may be due to a method that is not appropriate in the study’s context, or a broader issue with the research approach.

For example, most research on new and evolving eye-salvage therapies takes place in high-income countries. This may systematically exclude genetic, environmental, socioeconomic and other factors present in lower-income countries, potentially leading to treatments that may less appropriate, effective, or accessible worldwide.

Failure to design accessible research materials can also lead to systematic bias. For example, Blind/VI parents who cannot read a clinical trial protocol document and other key information may miss critical information and not consent to their child’s enrolment. Survivors who cannot independently complete research activities may give inaccurate or incomplete responses, or refuse to participate.

Impact: Results may not represent reality, compromising scientific progress, patient care, and outcomes.

Selection Bias

The study uses inappropriate criteria for recruiting participants, and small participant numbers. The cohort may not be representative of the wider retinoblastoma population, and the findings will be less valid.

For example, if researchers don’t make study advertisements accessible, they lose vision impaired participants. If study procedures take place only during work hours, they lose working parents. If they don’t offer study compensation or reimburse travel costs (including public transport, taxi fare, and fuel/parking fees), they lose participants with restricted finances or independent mobility.

A selection bias may be very specific to the study’s focus. For example, a study investigating the psychological effect of inherited retinoblastoma on parents and diagnosed children recruits only individuals with an identified RB1 pathogenic allele. In the process, it excludes the other parent, who is also deeply affected by retinoblastoma, and may be the child’s primary caregiver. If acknowledged, a separate study could explore this parent perspective.

Impact: When flawed research doesn’t account for diversity or experiential differences, we are less able to trust the results, generalize and apply them.

Survivorship Bias

The logical error of focusing on the people who made it past a selection process, while dismissing those who did not, making them less visible.

For example, researchers might concentrate more intensely on successful treatment strategies while ignoring the cases in which patients did not respond well, had adverse events, or died.

Impact: May create unrealistic expectations of vision-saving treatment outcomes, hamper timely life-saving care, and limit the development of strategies to manage unsuccessful treatment or palliative care.

Procedural Bias

The procedure or method used in a study systematically distorts the results.

For example, in a clinical study, patients receive the same new experimental treatment approach, but with differing combinations of drugs, doses, frequencies, and concurrent therapies. Any improvement in the patients’ cancer could be attributed to the new treatment approach, without being able to accurately define the most effective drugs, doses or frequency. The role of the concurrent treatment may also be overlooked. This could lead to overestimation of the new treatment’s efficacy, and too-quickly downgrade proven approaches to care.

Impact: May distort research results, leading to inaccurate conclusions and potentially misguiding clinical practice.

A young boy looks out over a stormy sea. Waves crash in and dark clouds gather in the distance. The photo is taken from behind the child.

Measurement Bias

Data is inaccurately gathered, either through faults in the measuring instrument or the individual collecting the data.

For example, a systematic measurement bias may arise when participating centres use different data management systems, and when those management tools cannot handle the unique complexity of retinoblastoma (two eyes, potentially multiple tumours, and different treatments for each). Important information may be lost in the process of collecting and aligning patient data, and researchers are less able to reconcile the data in the single study or in meta-analysis (comparison of multiple studies), potentially generating misleading results.

Bias may also occur when measuring only clinical outcomes using clinical tools, while ignoring Patient Reported Outcome Measures. PROMs are increasingly recognized as providing useful clinical insights that conventional tools don’t capture.

For example, researchers studying a new eye salvage therapy report that 90% of patients show complete tumour regression. The treatment is hailed as a breakthrough. However, PROMs are not collected as part of this study. Collecting PROMs during the initial study and longitudinal follow up could highlight that 75% of children experienced significant physical discomfort and/or psychological distress, and 60% experienced substantial or complete loss of visual function (how they use vision in daily life). These findings become even more significant when considering that the majority of children in the study had cancer in only one eye.

Impact: inconsistency in how data is collected or interpreted across different clinical settings or studies may distort our understanding of retinoblastoma, its treatments and outcomes, impeding the development of effective care. By only measuring clinical data with clinical tools, research overlooks significant quality of life factors that could inform decisions about care.

Analysis Bias

Processing data to confirm the researcher’s hypothesis.

For example, before launching their study of a new experimental treatment, a research team strongly believes this treatment will be more effective than existing eye-salvage therapies. As they analyse the data, they unintentionally focus more on the patients who showed a positive response, and overlook or downplay the patients who didn’t show any improvement or had adverse reactions. They interpret the data in a way that supports their preconceived ideas, instead of impartially considering all the results together. This could lead them to conclude the treatment is safer and more effective than it is.

Impact: May lead to inaccurate conclusions, which could negatively influence the direction of future research and treatment approaches. Patients may also be harmed if treatments being studied are not as beneficial as believed.

Confounding Variables

A factor in a study that gets mixed up with the things the researchers are trying to measure. This may be related to both what is being studied, and the tools being used to carry out the study. It can affect the results, suggesting a connection where there is none, or hiding a real connection.

For example, if a researcher doesn’t account for factors like patients’ specific age, stage of disease, and overall health while studying the efficacy of a new treatment, they may attribute variations in treatment outcomes solely to the new treatment, leading to false conclusions about its effectiveness.

Impact: Failing to account for confounding variables can generate misleading results and conclusions, potentially misguiding clinical decisions and future research focus.

Publication Bias

The outcome of a study influences the decision to publish its findings. Studies with positive outcomes are more likely to be published than those with negative or neutral results. Publication bias can be fuelled by multiple factors, meaning we may not get the full story from published research or the resulting media coverage.

Publication Criteria: Scientific journals love publishing studies that have a big impact or surprising results. Studies that don’t show much progress or report negative findings may be left out.

Researchers from Low and Middle Income Countries (LMICs) more commonly struggle to move beyond editorial review as too many editors erroneously believe that good science can’t come from LMICs.

Competition: Research institutions are often racing each other to make new discoveries, and competing for small pots of funding, and patients. If a study doesn’t show promising results, the team might choose not to publish so it doesn’t reflect negatively on their continuing research efforts and patient marketing campaigns.

Funding Sources: Sometimes, the people who pay for the research have a stake in the results. For instance, a company that makes a device or drug might fund a study to test it, or indirectly fund the institution carrying out the research. If the study shows the treatment isn’t very effective or has unexpected side effects or risks, the funder might not want those results to be published.

Impact: When only selected and successful studies are published, we have a false impression that exciting progress is being made all the time. Both the medical community and decision-making parents and survivors may overlook treatments that didn’t work or had severe side effects, leading to an incomplete understanding of treatment effectiveness and safety. xcluding or minimizing hinders important progress in global health knowledge and patient care.

We miss opportunities to learn from one another, from unsuccessful trials and patients’ adverse events, slowing progress towards the best outcomes for the child, family, and survivor.

An olive-green bird with distinctive white eyes perches on a branch just to the lower left of centre in the picture, almost hidden among the many large leaves and myriad shades of green.

Hidden biases are much like well-camouflaged birds. Even when you look carefully, blink for a moment, and you could easily miss this fast moving Japanese White Eye.

We Can All Reduce the Power of Bias

Reducing bias in any form is crucial to achieve valid and reliable research results. Below are some of the best ways to mitigate both cognitive and unconscious biases in retinoblastoma research.

Training and Awareness

Identifying bias is like birdwatching in a deep forest. If you’ve never done it before, you might not know where to look and what to listen for; you may not realize that a quiet rustling could be a bird. Training helps birdwatchers become more aware and adept at spotting birds of all kinds – even the most elusive.

Training helps everyone become more aware of the mental shortcuts that affect thinking and decisions. With awareness, they are more able to assess each step of the research process, identify potential bias, and manage it before it has an impact.

Patient-Led Research

Parents and survivors can advocate throughout the research process, ensuring the community’s experiences and needs are accurately represented.

Research Focus: When collaborating with researchers to identify priority questions, parents and survivors help ensure the resulting studies have maximum relevance and impact to patients and families in the real situations they face throughout life. This helps reduce biases that can arise when scientists and clinicians have a particular area of research interest, which may be out of step with patient, survivor and family needs and experience.

Design and Delivery: By working with parents and survivors on all stages of the study’s evolution and implementation, the investigators develop research methods and materials that are appropriate, accessible and inclusive for the retinoblastoma community. This can help reduce the barriers to study participation that become sources of bias in data collection and analysis.

Rigorous Study Design and Methodology

Minimizing bias at this stage is critical, especially as controlled clinical trials in retinoblastoma care are rarely feasible. Researchers must identify robust sampling techniques appropriate for the study population. They also need to plan how they will maintain uniformity in data collection, to ensure results aren’t skewed by varying methods.

Using prospective trials rather than case series also reduce bias as patient outcome is not yet known. Having peers review the plan before the study begins can be an additional safety net, helping to spot potential biases or errors. These combined strategies can produce rigorous research that drives care forward.

Pre-Registration

Pre-registration is possible for clinical trials and systematic reviews, but not practical for all study types. Researchers design all aspects of the study, from hypothesis to data collection and analysis methods, and register them publicly before starting the actual research.

Without a solid plan, scientists can be tempted to change their direction, methods, or analysis to fit what they find or hope to find in the data. Pre-registration reduces the risk of these post-hoc changes. Researchers are more likely to follow their pre-registration “treasure map”, regardless of what they find along the way. So we can be more confident that their findings result from following their hypothesis and plan, rather than random digging.

Transparent Reporting and Replication

Similar to pre-registration, transparent scientists fully describe how they carried out their studies – covering how they selected participants, gathered and analysed data, and especially what didn’t work out or was unexpected.

When researchers don’t tell us everything about how they did their study, we can’t fully understand or trust the results. By sharing every detail of their “recipe”, scientists allow others to better understand their results, repeat the study to confirm those findings, or identify any hidden biases or errors.

Scientists are increasingly asked to report their results against relevant standard checklists so that study strengths and limitations can be seen at a glance. Reporting guidelines can be found at the Equator Network.

Collaborative Meetings

Regular meetings and feedback loops between all stakeholders can help identify and mitigate any potential biases in the research process. Interdisciplinary collaboration can reduce bias by integrating diverse perspectives and expertise. Encouraging all stakeholders to continually learn and reflect on their own biases and how they might impact research and care.

Through open dialogue between parents, survivors, researchers and clinicians, each gain a better understanding of the other’s concerns and priorities. Parents and survivors gain deeper, more nuanced understanding of retinoblastoma research and management, and the professional learn more about the complex lived experience. This knowledge exchange can help reduce misunderstandings or misconceptions that can be a source of bias, and identify new opportunities for less biased collaborative research.

A black and silver compass tipped on its side facing the camera. The compass is sitting on top of an open paper map that extends out behind it out of focus.

Final Words

Cognitive and unconscious biases may influence retinoblastoma research focus, design, methodology, results, and conclusions. Biased outcomes can shape clinical care decisions, treatment trends, and the provision of psychosocial supports and lifelong care. Ultimately, this can lead to suboptimal medical care, preventable suffering, and even death.

We all have a responsibility to understand and address the most common types of bias impacting science. Doing so fosters a respectful, inclusive, patient-centred approach, while promoting the accurate evidence-based research we need to guide specialist medical care.

All biases, once unmasked, begin to lose their power. Through awareness and intentional action of the above strategies, we can minimize their impact on patient care and outcomes.

Most important is creating an open, understanding environment where everyone feels comfortable exploring the broad context of bias and discussing their personal biases. Progress will be greatest when bias is addressed in a safe, supportive community with the shared priority of best outcomes for the child, family and survivor.

By illuminating these shadowy influences, we hope to stimulate conversation and positive change for our global retinoblastoma community. Please join us as we venture on this enlightening journey of reflection and growth.

Please Join Us at One Rb World 2024!

If you would like to join a global community of clinicians, researchers, parents and survivors sharing knowledge and experience, and exploring priority subjects, we welcome you to One Retinoblastoma World, October 15-17 2024 in Honolulu, Hawaii. Community, conversation and collaboration for optimal care.

Visit our One Rb World 2024 microsite to read more, get your tickets, and book your group rate hotel room. We do hope very much to see you in Hawaii!

Sponsorship and partner opportunities are available for businesses, foundations and individuals. If you are interested in partnering with us to help host this important forum for global collaboration to improve patient care and cure, please email Marissa Gonzalez at marissa(at)wechope.org.

About the Author

Abby’s father was diagnosed with bilateral retinoblastoma in Kenya in 1946. Abby was also born with cancer in both eyes. She has an artificial eye and limited vision in her left eye that is now failing due to late effects of radiotherapy in infancy.

Abby studied geography at university, with emphasis on development in sub-Saharan Africa. She co-founded WE C Hope with Brenda Gallie, responding to the needs of one child and the desire to help many in developing countries. After receiving many requests for help from American families and adult survivors, she co-founded the US chapter to bring hope and encourage action across the country.

Abby enjoys listening to audio books, creative writing, open water swimming and long country walks.

Abby is wearing a blue and white patterned top, dark pink collared cardigan, and light blue patterned skirt, and a heart necklace. She is pictured with her black german sheperd - golden retriever guide dog, who is wearing a pink collar with guide dor tag showing. Both are smiling broadly at the camera. The background is a bed of bright orange marigolds, a low hedge and tall trees.

Read the story of how WE C Hope began.

0 replies

Want to join the discussion?
Feel free to contribute!