Levels of evidence in HOF: how to read the literature

Anyone working in orofacial harmonization lives with a flood of information: papers, congress abstracts, recorded lectures, colleagues' posts and manufacturer materials. Almost all of it arrives with the same visual authority, and it is easy to treat a series of ten cases as if it were a clinical trial. Knowing how to read the level of evidence is the filter that separates what deserves to change your practice from what deserves only curiosity.

This page is a reading guide, not a list of studies. It does not cite numerical research results or rank products. The goal is to give professionals a mental map: what kind of study the information came from, what that design can prove, where it tends to mislead, and how to turn all of that into clinical decisions and honest conversations with patients.

What does "level of evidence" mean?

The level of evidence is a ranking of the study design, ordered by the protection it offers against systematic errors (bias). The more a design isolates the effect of an intervention from chance, from the patient's expectations and from the researcher's expectations, the higher it usually sits in the hierarchy. The ranking is a compass, not a verdict: a poorly run trial can be worth less than a good cohort.

The classic hierarchy, from strongest to weakest, appears in the table.

Notice that no row is useless. Each one answers a different question, and the most common mistake is asking a design for an answer it cannot give.

Why is HOF a hard field to study?

Some obstacles are structural and help explain why much of what is taught in courses comes from accumulated experience rather than controlled trials.

• Imperfect blinding. The person injecting knows what was injected. Even when the outcome assessor is blinded, the look of the volume and local reactions can give the group away.
• Subjective outcomes. "A more harmonious face" depends on who is looking. Without a validated scale and trained raters, the outcome becomes an impression. See the validated aesthetic assessment scales.
• Short follow-up. Many relevant questions (real durability, late nodules, effects of repeated retreatments) require years of follow-up.
• Serious events are rare. Precisely because they are rare, trials are not large enough to detect them, and the knowledge comes from reports and registries.
• Heterogeneity. Product, technique, area, injector experience and patient profile all vary, and pooling everything into an average can hide differences.
• Funding. A good part of the research is sponsored by the people who make the product.

None of this invalidates the literature. It only asks that your reading be proportionate: modest claims for modest evidence.

How do you read a paper in five questions?

A short routine keeps the abstract and the conclusion (the most optimistic parts) from steering your opinion. Detailed reading of trials is covered in how to read a clinical trial of injectables; here is the essential.

1. What was the question, and was it defined in advance? Studies with the primary outcome registered before data collection are less likely to "fish" for results.
2. Who were the patients? Age, skin phototype, sex and procedure history define who the result applies to. A study of young patients with no prior treatment may say nothing about your 60-year-old patient with old fillers.
3. What was the comparator? Comparing with nothing, with placebo, with another product or with the opposite side of the same face are different situations, with different strengths.
4. Who measured the outcome, and how? A blinded assessor, a validated scale and standardized photos weigh heavily. See the importance of standardized clinical photography.
5. What was left out? Dropouts, adverse events and analyses that did not work out usually appear in the tables, not in the abstract.

What are the most common pitfalls?

Confusing association with cause. A cohort may show that patients who had a certain technique had less swelling, but they may have been different patients from the start. Only randomization balances what nobody measured.

Surrogate outcome. Measuring the thickness of a layer on an image is not the same as measuring satisfaction, natural look or safety. The surrogate can be useful, but it does not prove the outcome that matters to the patient.

Statistical significance without clinical importance. A difference can be statistically detectable and too small for the patient to notice. Always ask whether a difference of that size would change your approach.

Publication bias. Positive results get published more easily. What did not work tends to stay in the drawer, which makes the literature more optimistic than reality.

Citation chains. One group publishes a report, another cites the report, and soon the statement becomes "established knowledge" without any controlled study behind it. Trace the claim back to the primary source.

Conclusion bigger than the data. It is common for an abstract to conclude "safe and effective" from a sample that would not even have the power to detect rare events.

How does the evidence vary by HOF topic?

Each area of the practice has its own evidence profile, and that changes how much you can lean on the literature.

This table is general guidance, and it is worth checking the current literature on each topic before deciding. For an example of how a specific question is debated, see the page on the evidence for preventive toxin.

What to do when there is no good evidence?

In HOF this happens all the time, and the most honest path is not to pretend to certainty. In general, three sources are combined: the best evidence available (even if modest), your own clinical experience and that of trusted colleagues, and the patient's informed preference. The weight of each shifts with the risk: the greater the risk of an approach, the greater the evidence required to adopt it.

• In high vascular risk areas, prefer conservative approaches and tested safety protocols, even if a new technique looks attractive.
• With recent techniques, consider tracking the learning curve with selected cases, careful records and standardized photos.
• When the approach is off-label (outside the label indications), document your reasoning and inform the patient, according to your professional council's rules.
• Review your approaches as new evidence appears. The literature changes, and what was usual a few years ago may have been revised.

How do you turn reading into a habit without being overwhelmed by it?

Nobody can keep up with everything. A simple routine helps keep your knowledge current without consuming your day.

• Choose a few quality sources and follow them regularly: journals in the field, society guidelines and recent reviews. The page how to keep up with the scientific literature details the process.
• Record what changed in your practice and why, in a notebook or file. This keeps you from adopting fads and makes it easier to review decisions.
• Discuss cases with colleagues in study groups, where critique of the study design is part of the conversation.
• Be wary of absolute certainties. Phrases like "always", "never" and "the best product" usually signal opinion, not evidence.
• Value patient-reported outcomes. Instruments such as FACE-Q bring the perspective of the person receiving the treatment into the research.

Where does Harmy fit in this conversation?

Harmy is not a study and does not provide evidence about procedures. It is a communication tool for the practice.

This distinction matters. Showing a simulation to the patient is part of good informed consent; presenting the simulation as proof that the result will look like that is the kind of promise that careful reading of the literature teaches you to avoid.

Summary to take into practice

Levels of evidence are a tool of intellectual humility. They remind us that each source has a strength and a limit, and that the clinical decision comes from the sum of available evidence, experience and the patient's informed will. In HOF, where much of the knowledge still rests on case series and expert opinion, the safest stance is to claim little, document well, always review and explain the uncertainties to the patient. When an approach depends on a study, it is worth asking, before adopting it, what the design was, who the patients were and what was measured.

To go further, also read about injectable traceability and about adverse event reporting, two ways for the practice to contribute to the surveillance that feeds the field's knowledge.