The Placebo Effect in Veterinary Medicine: Interpreting Clinical Study Results | Animal Health Insights Town Hall

5 Key Takeaways for Designing Better Clinical Studies

Insights from Dr. Marie-Paul Lachaud, veterinary clinical development consultant with 35+ years of experience across companion animal, livestock, and regulatory affairs — shared during Prelude’s Animal Health Insights webinar series.

When a placebo-controlled pain study in dogs shows 60% improvement in the placebo group and 72% in the active treatment group, you haven’t failed. But you have encountered one of the most underestimated forces in veterinary clinical development: the placebo effect.

Dr. Marie-Paul Lachaud has spent over three decades navigating this exact challenge. A French veterinarian who co-founded one of Europe’s first veterinary CROs in 1990, she went on to lead animal health clinical development at ICON Clinical Research, held senior roles at Aratana Therapeutics (now part of Elanco) and One Health Pharma, and has contributed to product approvals across more than ten therapeutic areas. In 2018, she received the inaugural Feather Intercap Award at VMX.

During a recent Animal Health Insights webinar, Dr. Lachaud shared what she’s learned about why the placebo effect in veterinary medicine may actually be more powerful than its human counterpart and what smart teams do about it.

1. The Placebo Effect in Animals May Be Stronger Than in Humans

In human medicine, the placebo dynamic involves two parties: the patient and the practitioner. In companion animal studies, there are three: the patient, the owner, and the veterinarian. Each influences the others and the bond between pet and owner amplifies the effect in ways that human trials don’t have to account for.

Dogs are particularly sensitive to their owners’ expectations and emotional states. When an owner enrolls their pet in a clinical study, they become more attentive, more encouraging, and more hopeful, all of which the animal responds to. As Dr. Lachaud put it, the dog wants to please its owner, and that desire can manifest as apparent clinical improvement even without any active treatment.

This three-way dynamic means that study teams designing placebo-controlled trials in companion animals need to anticipate placebo response rates that can rival and sometimes exceed what human pharmaceutical teams encounter.

2. In Pain Studies, Expect 30–35% Placebo Response And Plan Accordingly

Dr. Lachaud cited well-documented placebo response rates of 30–35% in canine osteoarthritis pain studies that use the Canine Brief Pain Inventory (CBPI) — a validated owner-assessment tool now accepted by both the FDA’s Center for Veterinary Medicine (CVM) and European regulators as a primary endpoint.

The irony is worth noting: the endpoint both regulatory bodies have converged on is the one most susceptible to owner bias. But objective alternatives like force plates don’t work reliably with privately owned dogs who aren’t trained for laboratory conditions. The CBPI represents a practical compromise and when combined with the concept of clinical success (evaluating each individual animal as a success or failure rather than relying on group mean scores), it produces clinically meaningful results.

The practical implication: if you’re designing a placebo-controlled pain study in dogs, a 30–35% placebo response rate isn’t a surprise to plan around. It’s a baseline assumption to build into your statistical analysis plan from day one. And that has direct consequences for sample size calculations and study duration.

3. Baseline Diagnosis Is Your Most Powerful Tool Against Placebo Inflation

When Dr. Lachaud was asked how to reduce the placebo effect through study design, her answer was emphatic: start with baseline diagnosis.

In non-inferiority studies with two active treatments, investigators face pressure to enroll patients. A dog that scores a 4 might get recorded as a 5 to meet the inclusion threshold. In that context, the imprecision is less consequential. Both groups are receiving active treatment.

In placebo-controlled studies, that same shortcut inflates the placebo response. If the disease isn’t severe enough at baseline, milder cases are more likely to show improvement regardless of treatment and your ability to detect a real treatment effect erodes.

Dr. Lachaud’s hierarchy for controlling placebo response through design:

1. Rigorous baseline diagnosis — Ensure enrolled animals genuinely meet the severity threshold defined in the protocol. No fudging.
2. Validated primary efficacy endpoint — Chosen based on previous data, not intuition. The wrong endpoint at the wrong time point can render a study inconclusive.
3. Correct time point selection — Regulators won’t accept a post-hoc change from day 60 to day 30 because that’s where your significance happened to fall. This must be defined prospectively based on prior work.
4. Appropriate statistical plan — Models that account for the realities of animal health sample sizes (hundreds, not thousands) and that support clinical success endpoints.
5. Investigator site selection — Experienced sites that understand the rigor required for placebo-controlled work, which differs meaningfully from positive-control studies.

4. Pilot Studies Aren’t Optional — They’re the Difference Between Success and Expensive Surprise

A recurring theme throughout the discussion was the critical importance of prior work before committing to a pivotal study. Dr. Lachaud was direct: you cannot reliably predict what the placebo response will look like in your specific study from literature alone.

She described how the Aratana team (where she held a senior role) ran a single pilot study with 80 animals per group, over 200 animals total across four treatment groups, before proceeding to the pivotal. That investment in pilot data gave them a realistic picture of what to expect from the placebo group, how to power the pivotal study, and which endpoints would hold up under regulatory scrutiny.

The alternative, guessing, is how teams end up with pivotal studies that produce ambiguous results and regulatory requests to start over.

Dr. Lachaud also noted the distinction between placebo response and natural recovery. Diseases like osteoarthritis cycle through flares and remissions. Without robust baseline data and carefully chosen evaluation time points, it’s difficult to separate genuine treatment effects from the natural fluctuation of the disease. Another reason pilot data is indispensable.

5. Corporate Practice Consolidation Is Making Site Selection Harder

When asked what excites her about where veterinary clinical development is headed, Dr. Lachaud offered a candid concern instead: the consolidation of veterinary practices into corporate groups is creating new obstacles for clinical trial feasibility.

In theory, larger practices should mean bigger patient pools and faster enrollment. In practice, corporate veterinary groups often have standardized treatment protocols that conflict with study protocols, employee veterinarians with less financial motivation to participate in trials than independent practice owners, heavy administrative overhead that makes study participation logistically difficult, and managers who are too stretched to even review a study synopsis.

This is particularly consequential for placebo-controlled studies, which demand more diagnostic rigor at baseline than positive-control designs. Finding investigator sites that have the time, expertise, and institutional freedom to run these studies properly is becoming a real constraint, especially in Europe, where corporate consolidation of veterinary practices is accelerating.

The Audience Questions That Mattered Most

With 17 audience questions submitted during the session, several threads deserve attention beyond what the main takeaways cover.

Can you differentiate placebo effect from natural recovery? Dr. Lachaud pointed to diseases like osteoarthritis that cycle through flares and remissions. Without a sufficiently severe baseline diagnosis, what looks like placebo response may simply be the natural fluctuation of the disease. This is another reason time-point selection matters so much. A treatment that shows significance at day 14 may lose it by day 28 as the disease naturally oscillates, and regulators won’t accept a post-hoc endpoint change.

Does double-blinding help? In theory, yes, but it doesn’t eliminate the core issue. The owner still knows their pet is in a study, still becomes more attentive, and still unconsciously encourages the behaviors being evaluated. Double-blinding addresses investigator bias, but the pet-owner dynamic remains intact.

What about behavioral studies? Expect an even larger placebo effect. Dr. Lachaud described working on cognitive dysfunction studies where cultural differences between US and European investigators made it difficult to even agree on assessment criteria. American investigators focused on discrete observable behaviors (does the dog go to the right side of the door?), while European investigators looked at overall behavioral patterns. Finding validated, cross-culturally reliable assessment tools for behavioral endpoints remains an active challenge.

Is there a way to design owner assessments that minimize bias? This is where validated instruments like the CBPI become critical. Dr. Lachaud credited the years of work by Dottie Brown at the University of Pennsylvania (now at Mars) for developing and validating the CBPI. The Aratana team partnered with Brown to access her data and co-publish validation papers, a model for how sponsors and academics can collaborate to build the assessment infrastructure that placebo-controlled studies require. A feline version has also been developed, though cat pain assessment remains particularly difficult given how subtly cats express discomfort.

What This Means for Your Next Study

The placebo effect in veterinary clinical trials isn’t a flaw to be eliminated. It’s a structural feature of how companion animals, their owners, and clinical evaluators interact and it has to be accounted for from the earliest stages of study design.

Dr. Lachaud closed with a framing that stuck: placebo-controlled studies remain the gold standard for demonstrating true treatment effect. More studies in Europe are moving to placebo-controlled designs as regulatory requirements tighten, and the era of underpowered non-inferiority studies with ambiguous conclusions is fading. As she put it, if the best drug doesn’t make a difference versus placebo, you haven’t got a drug.

For teams designing their next placebo-controlled study, the takeaways are actionable: invest in pilot work, choose your endpoints based on data rather than convenience, power for the placebo response you’ll actually encounter, and select sites with the experience and capacity to execute rigorously. And when you sit down with your statistician, make sure you understand each other. The statistical analysis plan has to account for the realities of animal health sample sizes, where hundreds of animals is a large study, not a rounding error.

---

Watch the Full Conversation

This article is based on Prelude’s Animal Health Insights webinar featuring Dr. Marie-Paul Lachaud. Watch the full replay here.

Animal Health Insights is a monthly webinar series hosted by Prelude CEO Tommy Jackson, bringing together experts from across the animal health R&D community to discuss the topics that matter most to clinical development teams. Subscribe to our newsletter to get notified about upcoming episodes.