Every trial sponsor has felt it: the quiet dread when a study that should enroll on time starts falling behind. You've done the math. You have enough sites. The patient population exists. And yet.

The temptation is to blame randomness. To chalk it up to unpredictable site performance, patient hesitancy, or bad luck. But here's the thing: enrollment isn't random. It follows laws. And once you see those laws clearly, seemingly chaotic trials become explainable systems.

The key is understanding one number: patients per site per month (PSM). It's the fundamental particle of clinical trial enrollment. Get this right, and everything else follows.

The Fundamental Particle

In physics, atoms are the building blocks of matter. In clinical trials, PSM is the building block of enrollment.

The math looks simple:

Total Enrollment = Sites × Time × PSM

If you need 300 patients, and you have 30 sites running for 10 months, you need an average PSM of 1.0. Elementary arithmetic.

But here's where the simplicity becomes deceptive: averages lie.

A study might report PSM of 1.0 across its sites. But that average hides massive variance. Five sites might be running at PSM of 3.0, carrying the study. Twenty sites might be limping along at 0.2. And five sites might have enrolled exactly zero patients despite being open for six months.

Like atoms, PSM looks simple from a distance. Up close, it contains hidden complexity—internal structure that determines whether your trial succeeds or fails.

This is why experienced operations teams obsess over PSM at the site level, not the study level. The study-level average is an output. The site-level distribution is the input you can actually influence.

Forces That Drive Enrollment

In classical mechanics, acceleration requires force. Objects at rest stay at rest unless something acts on them. The same is true for clinical trials.

Applied force is what accelerates enrollment. Three forces matter most:

Site selection quality. This is your largest lever. Choosing sites with demonstrated enrollment capability—sites that have enrolled similar patients in similar trials—applies immediate force to your enrollment curve. A site that enrolled well in your competitor's trial will probably enroll well in yours.

Protocol simplicity. Every additional visit, every extra blood draw, every complex eligibility criterion adds friction. Simple protocols enroll faster because they reduce the per-patient burden on sites. The more complex the protocol, the harder you have to push.

Competitive positioning. Sites have finite capacity. If your trial is one of three competing for the same patients, you're splitting the available force three ways. Being the preferred trial at a site—through better relationships, easier protocols, or more compelling science—concentrates force in your direction.

The enrollment equation, roughly:

Enrollment Velocity = (N sites) × (avg PSM) × (activation rate)

More sites, higher PSM, faster activation—these are your levers. But as with any physical system, there's another side to the equation.

Friction: What Slows Everything Down

Every physics student learns about friction. It's the force that opposes motion—the reason objects eventually stop, the reason engines generate heat, the reason perpetual motion machines are impossible.

Clinical trials have friction too. It comes in two forms.

Static friction is the resistance to starting. Before a site can enroll a single patient, it must overcome:

• IRB/ethics committee review and approval
• Contract and budget negotiations
• Site initiation visits and training
• System access, supply shipments, regulatory documentation

Static friction is why "site count" is a misleading metric. A site that's selected but not activated contributes nothing. And the gap between selection and activation often stretches months longer than anyone planned.

Kinetic friction is the resistance during motion. Even after a site starts enrolling, constant forces work against progress:

• Competing trials recruiting from the same patient pool
• Protocol amendments that require re-training and re-consenting
• Screen failures that consume site effort without contributing patients
• Patient dropout that erodes hard-won enrollment gains

Here's the insight that planning often misses: friction is invisible in projections but dominates execution. Enrollment plans typically assume smooth ramp-ups and steady-state performance. Reality delivers delays, interruptions, and constant drag.

The physics parallel is exact: a car's engine might produce 300 horsepower, but if you're driving through mud, most of that power converts to heat rather than forward motion. A trial might have 50 activated sites, but if half are fighting excessive friction, they're consuming resources without contributing proportionally to enrollment.

Momentum and Inertia

Newton's first law: an object in motion stays in motion. An object at rest stays at rest.

Momentum in clinical trials is real. A site that's enrolling well tends to continue enrolling well. Why?

• The study team becomes efficient at the protocol
• Referral networks activate and strengthen
• Sponsor-site trust builds, enabling faster problem resolution
• The site prioritizes your trial because it's succeeding

Momentum compounds. Early success creates conditions for continued success.

Inertia is equally real, and more problematic. Enrollment patterns resist change.

A struggling site rarely recovers. By the time you've identified underperformance, diagnosed the cause, implemented interventions, and waited to see results, months have passed. The rescue almost never arrives in time.

Adding sites late doesn't accelerate enrollment proportionally. New sites face full static friction—the entire activation timeline—while your study clock keeps running. Late sites also join when the easy-to-find patients have already been identified, leaving only harder-to-reach populations.

This is why early site selection quality matters more than late intervention. The physics favor getting it right at the start, not fixing it later.

There's an asymmetry worth noting. Building momentum takes sustained effort over months. Losing momentum can happen in weeks—a key coordinator leaves, a competing trial opens, a protocol amendment disrupts workflows. Momentum is hard to build and easy to destroy.

The Planning Paradox

Most enrollment plans make a fatal assumption: that PSM will be roughly uniform across sites.

They project that if the protocol supports PSM of 1.0, all 30 sites will perform somewhere close to 1.0. Maybe some variance—0.8 here, 1.2 there—but basically uniform.

Reality follows a power law.

In most trials, 20% of sites contribute 50-80% of patients. The distribution isn't normal; it's heavily skewed. A few sites massively outperform. Many sites barely contribute. Some contribute nothing at all.

The physics implication is counterintuitive: optimizing the top 20% of sites matters more than expanding the bottom 50%.

If five sites are enrolling at PSM of 3.0, helping them reach 4.0 adds more patients than activating ten new sites that might achieve 0.5. Resources spent rescuing underperformers often generate less return than resources spent enabling high performers to do even more.

This doesn't mean you abandon struggling sites. But it does mean you allocate attention and support proportionally to expected return—not equally across the portfolio.

Think of it like investing. You wouldn't allocate capital equally across all stocks regardless of performance. You'd weight toward your highest-conviction positions. Site management follows the same logic: concentrate resources where the physics are most favorable.

Implications for Trial Design

Seeing enrollment as a physics problem changes how you approach it.

Measure variance, not just averages. A PSM average of 1.0 with low variance is a different situation than 1.0 with high variance. The latter indicates a few sites carrying dead weight—fragile performance that depends on continued high performance from your top enrollers.

Apply force where friction is lowest. Your highest-performing sites are already demonstrating low friction. They have the processes, the patient flow, the motivation. Help them do more before trying to transform sites with fundamental friction problems.

Build momentum early. The compounding effect of early enrollment success is significant. Prioritize fast site activation. Consider starting with your highest-confidence sites rather than activating everything at once.

Treat PSM as a distribution, not a point estimate. When planning, model the range of outcomes. What if your top 10% of sites hit PSM of 3.0 and your bottom 30% hit 0.2? That's often closer to reality than assuming everyone hits 1.0.

Reduce static friction aggressively. Every week saved in site activation is a week gained for enrollment. Streamline contracts. Standardize budgets. Pre-qualify sites so they're ready to launch. The fastest path to enrollment isn't always more sites—sometimes it's faster activation of fewer, better sites.

Clinical trial enrollment isn't a hope problem. It's a physics problem. The forces are knowable. The friction is measurable. The dynamics follow laws.

Once you see it this way, you stop asking "why is this trial struggling?" and start asking "what forces are we applying, and what friction are we fighting?" The answers, it turns out, are usually hiding in the PSM distribution—that fundamental particle that determines everything downstream.