Dentists today have more clinical information available to them than ever, and not all of it carries the same weight. Dr. Richard H. Nagelberg, director of medical affairs at OraPharma, a division of Bausch Health US, and a general dentist in suburban Philadelphia for more than 40 years, has spent much of his career helping clinicians think critically about the evidence behind their decisions. I spoke with Nagelberg about how to evaluate clinical evidence, no matter where it comes from.
Kevin Henry: Dentists today are getting clinical information from more sources than ever, such as journals, continuing education courses, conferences, manufacturers, digital platforms, even AI tools. What's the central challenge in that environment?
Dr. Nagelberg: It's not deciding which sources to trust. It's understanding how to evaluate any piece of clinical evidence, no matter where it comes from. Not all evidence is created equal, and the same conclusion can rest on a large, well-designed randomized trial or on a single uncontrolled study. The credibility of the source and the quality of the underlying evidence are two different things, and clinicians need to be able to tell them apart.
Let's talk about study design. Where should someone start when they're trying to gauge how much weight to give a study?
Dr. Richard Nagelberg
Start with what type of study it is, because that determines what it can actually tell you. Case reports and case series describe outcomes in one patient or a small group. They're useful for generating hypotheses or flagging unusual findings, but they can't establish that a treatment works, only that it appeared to work under specific, uncontrolled circumstances.
What about observational studies, like cohort studies, case-control studies, that kind of design?
Those can identify associations between an exposure and an outcome, but they can't establish causation on their own. If patients who received a particular treatment did better than those who didn't, that difference might reflect how patients were selected for treatment in the first place rather than the effect of the treatment itself.
And randomized controlled trials?
Randomized controlled trials (RCTs) are the reference standard for evaluating treatment efficacy. Randomization distributes known and unknown confounders across groups so the treatment itself becomes the most plausible explanation for any difference you observe.
A well-designed RCT -- including adequate sample size, appropriate controls, blinding where feasible, prespecified primary endpoints -- provides the strongest basis for a clinical conclusion that a single study can offer.
Where do systematic reviews and meta-analyses fit in?
They synthesize findings across multiple studies using explicit, reproducible methodology. When they're available and well conducted, they give you the most complete picture of what the evidence as a whole shows, including where it's inconsistent or limited, which a single study can't do.
If a dentist wants a quick mental checklist before trusting a piece of evidence, what should be on it?
A few questions cover most of it: What type of study is this, and what can that design actually establish? Was it randomized and controlled or observational? How large was the sample, and was it powered to detect the difference being reported? Was there a control group, and was it the right comparator? Were the outcomes assessed by someone other than the treating clinician? And was the study peer-reviewed, and was it externally funded?
Asking those questions consistently, on any evidence from any source, is most of the work.
You draw a distinction in your writing between statistical significance and clinical relevance. Why does that trip people up?
Because they sound like the same thing and they're not. A result is statistically significant when the probability of observing it by chance falls below a predetermined threshold, conventionally a p-value less than .05. That tells you nothing about whether the size of the effect actually matters to a patient.
A treatment can produce a statistically significant 0.3 mm reduction in a clinical measurement in a large enough study without that translating into a meaningful change for the patient. The reverse happens too, for example, a large, clinically meaningful effect can fail to reach statistical significance in an underpowered study. Both show up in the literature, and both get cited selectively.
What should someone look for instead of the p-value?
Look for effect size, which is the actual magnitude of the difference between groups, alongside the p-value. And ask whether the primary endpoint is a clinical outcome that matters to patients, like pain, function, or disease resolution, or a surrogate measure, like a lab value or technical metric assumed to correlate with clinical outcomes. That distinction isn't always spelled out in summaries, so it's worth checking.
You’ve mentioned a few recurring patterns where a conclusion goes beyond what the underlying evidence can support. What are the most common ones?
Overgeneralization of population is one. A study done in a narrowly defined group, by severity, age, or systemic health status, may be extrapolated for more general use.
Conflating association with causation is another; observational data can support a hypothesis, but it rarely proves one, so when you see a causal conclusion built on correlational findings, that's worth a closer look at the study design.
Selective citation is a third: A single favorable study cited on its own can tell an incomplete story if other studies on the same question reached different conclusions, which is exactly why systematic reviews and clinical practice guidelines matter, because they synthesize the full body of evidence rather than the highlights.
And follow-up duration?
That's the fourth pattern. Studies measuring outcomes at one, three, or six months may not reflect how durable a treatment effect is over the time frame that actually matters to patients. A treatment with strong short-term results and uncertain long-term durability is a different proposition from one with both, and that gap doesn't always get flagged.
What are the practical habits a busy clinician can actually build into their routine?
A few things. When a piece of evidence comes up in a CE program, product material, or a colleague's recommendation, ask what study type it's based on and whether you can get to the original publication, as abstracts and summaries can leave out limitations that are disclosed in the full text.
PubMed gives free access to most abstracts and a lot of full-text articles, and a targeted search filtered by study type can change how that evidence lands. For industry-sponsored research, note the disclosure and ask whether the study design and analysis plan were preregistered, ideally through ClinicalTrials.gov, before data collection -- that's a meaningful signal of methodological integrity.
For bigger clinical decisions, is a single well-designed study ever enough on its own?
For decisions with significant patient impact, look beyond a single study. Clinical practice guidelines from organizations like the ADA or specialty academies apply systematic methods to synthesize the available evidence and grade the strength of their recommendations.
They'll distinguish between strong evidence supporting an approach and expert consensus in the absence of robust trial data. That distinction matters, and it's not something a single study can give you.
If you had one message for the dental community about reading clinical evidence, what would it be?
Judge evidence by its strength, not by where it appears. Evidence from a well-designed RCT deserves confidence whether it's in a peer-reviewed journal or a manufacturer's clinical summary.
Evidence that doesn't meet that bar deserves scrutiny in either context. It’s not about approaching new evidence with doubt. It's about bringing the same structured, consistent inquiry to every source, which is ultimately what lets clinicians apply new evidence responsibly and have more confident, well-grounded conversations with their patients.
Disclosure: Dr. Richard Nagelberg is the director of medical affairs at OraPharma, a division of Bausch Health.
The comments and observations expressed herein do not necessarily reflect the opinions of DrBicuspid.com, nor should they be construed as an endorsement or admonishment of any particular idea, vendor, or organization.
