Introduction
The modern health market is a space of immense information asymmetry. The line between professional medicine and the commercial wellness sector is blurring, leaving patients vulnerable to manipulation and costly errors.
In this article, based on Deborah Cohen's book "Bad Influence," we analyze the concept of calibrating institutions. You will learn why individual patient vigilance is not enough and how systemic oversight mechanisms can protect our autonomy from the aggressive marketing of innovation.
Calibrating Institutions as a Shield Against Information Asymmetry
Deborah Cohen argues that patients cannot independently verify the quality of modern technologies. Expecting a layperson to evaluate the construction of an algorithm or the reliability of a biomarker transforms autonomy into an overwhelming burden of constant auditing.
Medicine involves so-called credence goods—services whose quality cannot be assessed even after they have been delivered. An example is an unnecessary diagnostic test; a patient may never know that it was redundant.
This is why calibrating institutions are essential. These are structures that correct information asymmetry and enforce a minimum standard of evidence. Their role is to protect individuals from adopting technologies with merely hypothetical benefits too early.
The Regulatory Gap Between Wellness and Medicine
Current regulations often fail because the law classifies categories, while innovations transcend them. Wellness products can evade medical rigors by using "lifestyle" language, while in reality, they strongly influence users' health decisions.
Deborah Cohen identifies this as the regulatory perimeter problem. A manufacturer may claim a wellness purpose to bypass clinical evidence requirements, even though the product de facto suggests medical solutions.
To counter this, regulation must address not only the product's name but its actual function within the decision-making chain. It is crucial that the strength of claims in marketing corresponds to the actual strength of evidence, preventing the phenomenon of evidentiary inflation.
Lifecycle Regulation Instead of One-Time Certification
Traditional, one-time certification is insufficient for AI technologies that evolve after entering the market. Deborah Cohen points to the pacing problem—technology develops faster than legislative cycles.
The solution is lifecycle governance, or oversight across the entire product lifecycle. Instead of asking "was the product approved?", we should examine its real-world performance and update requirements as new data emerges.
In this model, tools such as Health Technology Assessment (HTA) are key, filtering clinical value against marketing promises. HTA does not only ask about safety, but whether a technology offers a real benefit relative to costs and alternatives.
In the case of AI, procedural regulation is necessary. Rather than controlling every single output, oversight should focus on how models are trained, their auditability, and error escalation mechanisms.
Summary
Calibrating institutions do not restrict patient freedom; rather, they create the infrastructure for genuine autonomy. Through them, choice is based on reliable data rather than marketing proficiency.
We are faced with a question about the limits of self-improvement in a world where every imperfection is treated as a product. Is the goal of medicine to create a maximally optimized and eternally improved human being?
Perhaps the true luxury of the future will be the state of being healthy enough to finally stop thinking about health altogether.
Frequently Asked Questions
Why can't a patient independently assess the quality of modern health technologies, and whom do they need for this verification?
Independent assessment of the quality of health technologies by the patient is impossible due to information asymmetry and the fact that many medical services are credence goods, whose value cannot be verified even after they have been delivered. Calibrating institutions and regulatory systems are necessary for verification, as they correct these deficiencies and enforce a minimum level of evidence.
Why do current regulations fail to protect patients from products that are not medicines but suggest medical solutions?
Current regulations rely on category classification and the manufacturer's declared intended use, whereas innovations often blur these boundaries. This allows manufacturers to use wellness language instead of medical terminology, meaning products can strongly influence users' health decisions while remaining outside the regime of medical regulations.
Why are traditional regulations insufficient for controlling AI technologies in healthcare, and how can they be replaced?
Traditional regulations are insufficient due to the pacing problem—the fact that AI technology evolves faster than the legislative and evidentiary cycle. They can be replaced by lifecycle governance and models such as responsive regulation or agile regulatory governance, which rely on post-implementation system monitoring and adaptive adjustment of requirements.
What is HTA, and why is product safety alone not enough to protect patients from manipulation?
HTA is an interdisciplinary assessment of the clinical benefits and safety of a technology in relation to costs, system impact, and social and ethical aspects. Product safety alone is insufficient because a product with low physical risk can generate high behavioral risk, e.g., through incorrect recommendations leading to delayed treatment.
How can we prevent the marketing exaggeration of the effects of new health technologies while maintaining the pace of innovation?
Marketing language should be calibrated so that the strength of claims is proportional to the strength of the research design and evidence. A 'regulatory sandbox'—a controlled testing environment under the supervision of a regulator—can also be used, allowing risks and opportunities for innovation to be safely explored before full market deployment.
How can calibrating institutions practically counteract information asymmetry and the influence of industrial lobbying?
Calibrating institutions can counteract information asymmetry by creating public registries of claims and evidence and implementing transparent data disclosure standards that are easy for users to process. To avoid the influence of industrial lobbying, a pluralization of knowledge sources should be applied, involving independent academics, patient organizations, and various experts in the assessment processes.
Does the fact that a technology works and is safe automatically mean that it is worth introducing into the healthcare system?
No, because a technology can be effective and safe, but at the same time too costly or likely to deepen social and health inequalities. The decision to implement it requires an assessment of the systemic impact and an analysis of whether the benefits are distributed fairly.
What should accountability and oversight of AI systems in healthcare look like to avoid the blurring of blame and systemic errors?
Responsibility should be shared between the professional (application of the tool), the platform (presentation of content and data), and the patient (informed decisions). To avoid the blurring of blame, it is essential to implement principles of traceability and contestability, allowing for the reconstruction of events and the ability to challenge algorithmic recommendations. Oversight must shift from controlling individual pieces of content to regulating processes, taking into account latent conditions (e.g., design flaws) rather than only the errors of frontline personnel.
How can systemic regulations protect patient freedom instead of restricting it?
Regulations protect patient freedom by creating an infrastructure that eliminates information asymmetry and ensures reliable data (e.g., through labeling advertisements or validating models). This ensures that patient autonomy is based on real choices, rather than susceptibility to manipulation and persuasive promises.
