Test new medicines and medical technologies on a wide range of people, and spread them only after checking the data
Steps to make it happen
- Makers of medicines and medical tools first submit a plan that says what kinds of people they will test on.
- They set goals so that age, sex, body type, and home region are not lopsided.
- Older people and people who live far away can join tests from a nearby hospital or from home.
- Genetic data and data used to train health apps come with a table that shows whose data it is.
- When a medicine is approved, its instructions list who was tested and which groups are missing.
- For the missing groups, makers keep studying the effects and side effects after the medicine goes on sale.
- Making sure a medicine is safe for everyone before it spreads becomes the normal way.
Water so far 💧 0
The proposal in detail
This proposal covers technologies used directly in decisions about people's health: new medicines and vaccines, treatments and tests that use genes, and AI and health-management apps that help judge illness. Before starting trials that check effects in people, known as clinical trials, the companies and research institutions that develop these technologies submit a plan for who will take part to a public body that reviews whether use should be allowed. The goals are set by looking at the people who actually have the disease: their age, sex, body size, genetic background, other illnesses and other medicines they take, and the region and income level where they live. For gene databases and AI training data, a note is attached, like the ingredient list on food, that states where the data came from, what it is made of, and which people are not included. When the review body approves a technology, it requires that the makeup of the trial participants and the groups for whom the evidence is still thin be stated clearly in the package insert and in the explanations given to users. For groups with gaps, follow-up studies after the technology goes into use are made a condition. For rare diseases with very few patients, and in emergencies such as a large outbreak of infectious disease, approval is not delayed, and filling the gaps through studies after use begins is allowed. Children and pregnant people are added step by step, with extra care for safety. Information about participants' traits is handled anonymously and with their consent, and the review bodies of different countries agree on a common record format so that results can be compared across borders.
BenefitsBrings us closer to medicine that is safe and works for everyone
The biggest benefit is that checking is done on groups close to the people who will actually use the medicine or technology, so side effects or weak effects that occur only in certain people are easier to find before sale or at an early stage. Because the instructions say for which people the evidence is still thin, doctors and patients can choose carefully, knowing what is uncertain. For older people, people with other illnesses, and people in distant regions, who have not been included in trials until now, there will be more information that applies to them. For gene tests and AI as well, showing what the data is made of lets users more easily spot situations where wrong results are likely. Because trial sites will need to be placed in many regions of the world, research strength is less concentrated in a few countries, and it can also become a chance for many countries to build up medical workers and equipment. If less is wasted on continuing treatments that do not work for the people they do not suit, medical costs can be saved as well. And the reassurance that the technology has been checked on people like ourselves supports society's trust in new medical technology.
DrawbacksDevelopment may become slower and costlier, hurting patients who wait
The heaviest concern from opponents is that gathering diverse participants takes time and money, so new treatments reach patients later. For a person who has no treatment now and is in danger of dying, a delay of months or years before approval cannot be made up. The added cost may be passed on to the price of the medicine, and small companies, universities, and researchers in poorly funded countries, which have less strength, are more likely to give up on development, so there is a worry that the makers of technology may end up concentrated in a few large companies instead. Also, putting heavy weight on fine categories such as genetic background or ethnic group could make biological differences that are really small look large and strengthen prejudice. The more information about participants' traits and genes is collected, the higher the risk that personal information leaks or is misused. In addition, some criticize that spreading trials around the world could use people in low-income countries as cheap trial participants while the finished medicine never reaches them, creating unfairness. There is also a chance that goals set just to hit numbers will be met only in form.
Learn the background
How well a medicine or medical technology works, and what side effects it causes, can change with age, sex, body size, genetic background, other medicines a person takes, diet, and living conditions. Yet it has long been pointed out that clinical trial participants lean toward people in a few regions with many research facilities, who are relatively young and have few other illnesses, while older people, women, pregnant people, children, and people in low-income countries are few. There are reports of cases where, after a product went on sale, it turned out that certain groups had more side effects or needed a different dose. In gene medicine too, the genetic information used in research leans toward people with certain ancestries, and it has been pointed out that test results are harder to interpret for everyone else. Among AI tools that help judge illness, it is known that accuracy drops for skin colors or imaging equipment that were rare in the training data. Today a single medicine or app is used all over the world across borders. Spreading it without asking who it was tested on could place hidden dangers on the people who were never checked.
Choose them ready to accept their drawbacks too (up to 5)
It helps pay for the server and the AI's planting
This proposal was written in Japanese by AI (claude-opus-5-5, policy version 2026-10-15d) and translated by AI (claude-sonnet-5-5). Nobody has checked the facts. Read the original (日本語) How the AI writes