Democratizing
medicine

The paradigm of how we discover medicines is changing.

The ways we deliver them to patients are changing, too. Together, these developments have potentially seismic implications—for drug developers, clinicians, and patients.

What’s driving these changes? There are many factors, but it’s possible to point to a few key enablers: innovations in drug development, new solutions for drug delivery, and easier access to the tools needed for both.

The earliest medicines were stumbled upon in nature. A classic example is willow bark, used as a pain reliever in ancient times. Later, the small molecule responsible for willow bark’s therapeutic effects was isolated, and aspirin was born. For much of human history, this is how drug development proceeded:

“I found some new plant. Let’s see what we can isolate from this plant and [whether] that has any therapeutic effect,” says Tim Leaver, Senior Director of Nanomedicine, Lipids, and Services at Cytiva.

We’ve come a long way since then. Computational modeling and AI tools accomplish screening tasks in a fraction of the time needed for traditional methods, while advances in gene sequencing and molecular analysis have blurred the lines between discovery and optimization. But what we’re seeing now is a more dramatic directional shift: whereas in the past we’ve often known that a drug works before learning how, today we’re increasingly able to reverse-engineer therapies based on how we want them to work. In other words, the route from discovery to mechanistic understanding is now a two-way street.

That street is how we arrived at CAR T cell therapy, for example, where a patient’s immune cells are modified to express chimeric antigen receptor (CAR), a protein that directs the patient’s immune cells to fight the cancer from within. The approach is not only much more targeted than traditional methods like chemotherapy and radiation but has also provided durable remission for many patients (1).

But producing this therapy is resource-intensive (2), both because of the viral vectors used to deliver it and because the cells are modified outside the patient’s body.

So what does this shift look like in practice? What’s driving it? And, crucially, what’s different this time around? Why now? One key step is the ability to design drugs that can change how the body’s cells function at the molecular level.

One way to think of it is like software. A software developer writes lines of code, telling a computer what to do. In much the same way, drug developers are now able to use “code” to give the body instructions. This is possible in a range of applications. Click the images below to see how:

These interactive games need more room. For the best experience, view on a screen at least 1024 px wide.

Using messenger RNA (mRNA) to carry instructions to cells, telling them which proteins to make. In the case of an mRNA vaccine, the code tells the cell how to make a small, harmless piece of a virus (say Covid-19) so the body can learn how to fight it.

CRISPR technology that uses a protein to edit DNA. Harmful gene variants can be turned off or silenced; mutations can be fixed; or genes can be modified to, for example, help combat a cancer.

CAR T therapy that modifies a patient’s immune cells to recognize and destroy cancer cells.

Until recently, however, most of these approaches have involved engineering cells outside the body, then reinfusing them. The next step: in vivo medicine. As Tim Leaver says,

In vivo is when those genetic steps are taking place inside the body rather than outside.

Already, clinical results for LNPs in in vivo therapy are showing promise. In results published in the New England Journal of Medicine in October 2025, a team in China used LNPs and mRNA to effect in vivo generation of CAR T cells in five people with lupus (3). The authors noted “reduced disease activity, and no major toxic effects”—a promising initial result that underscores the real-world potential of LNP technology.

Peer predicts still more breakthroughs in the near future. “We’re going to see more data on cancer,” especially blood cancers, and other applications, including therapeutic proteins (such as antibodies or ‘intrabodies’) in different diseases, he says.

These innovations are also surfacing questions about how key players will adapt. Regulation is a concern shared by all the experts interviewed for this article. Right now, regulators license drugs to treat specific diseases. A treatment might be approved for use with pancreatic cancer, for example, and separately for bladder cancer. It’s not clear how regulators would handle cases where a generic delivery vehicle can be loaded interchangeably with different types of personalized therapeutic molecules.

But the FDA has signaled openness to such medicines. In late 2025, the agency said it would be open to the exploration of individualized pathways where randomized clinical trials aren’t feasible (4), but specifics remain under development. The question can be framed as:

If all the pieces are in place, Leaver continues, “the logical extreme of this is actually personalized medicine.” It could even extend to breakthroughs like personalized cancer vaccines, or treatments for ultra-rare diseases, where “an entire clinical trial process just doesn’t make sense.”

Such a shift would be a radical one, requiring regulatory flexibility and a certain amount of rethinking, not just in how patients are treated, but in how drugs are manufactured and potentially in how insurers assess costs.

The status quo is expensive in many cases: ex vivo gene therapy is costly not only because of the labor and equipment needed to produce it, but because patients in some cases spend many weeks in the hospital, preparing for and then receiving the therapies. Traditional cancer treatments like chemotherapy and radiotherapy are more widely available, but they also have deep associated costs, financially and in their impacts on the health (and productivity, and happiness) of recipients and their loved ones.

Democratized medicine with more in vivo development wouldn’t eliminate costs, but could it move them? For patients, the treatment process would be simplified, even if it takes longer for the price to come down. Moreover, there’s reason to be optimistic about costs as well.

The modular nature of LNPs lends itself to a multidrug manufacturing model that could achieve economies of scope rather than scale. This would be analogous to the establishment of “foundries” in the computer chip manufacturing industry that precipitated rapid innovation and lower costs in that field (5). And in similar ways, an LNP foundry allows the capital costs to be shared among numerous smaller drug developers who all manufacture their LNP drugs in the same facility. Coupled with a platform-based regulatory framework, this could open up the bottlenecks drug developers face in translating their work to the clinic.

The last decade has shown, more clearly than ever, that when the imperative is strong enough and the key players are on board, the world of drug development can marshal its forces to radically change things for the better.

We don’t have to wait for another pandemic.

We can start now.