Appendix C: The Evidence Hierarchy¶

How to Evaluate Claims About Oral Health¶

I've been watching humans make claims about oral health for millennia. I've seen promising approaches validated and cherished beliefs overturned. What I've learned is that enthusiasm is no substitute for evidence. This appendix is designed to help you think critically about the claims you encounter—including the ones I've made in this book.

The Pyramid of Evidence¶

Not all evidence is created equal. Scientific evidence exists in a hierarchy, from weakest to strongest:

                    /\
                   /  \
                  /    \
                 / SRs  \
                / & MAs  \
               /----------\
              /   RCTs     \
             /--------------\
            / Cohort Studies \
           /------------------\
          /Case-Control Studies\
         /----------------------\
        /    Case Reports &      \
       /     Case Series          \
      /----------------------------\
     /      In Vitro & Animal       \
    /          Studies               \
   /----------------------------------\
  /       Mechanistic Reasoning &      \
 /        Expert Opinion                \
/----------------------------------------\

From bottom to top:

Level 1: Mechanistic Reasoning and Expert Opinion¶

What it is: Theoretical arguments about why something should work, based on known biology and chemistry; opinions of experts in the field.

Example: "Baking soda is alkaline, and acidic conditions cause demineralization, therefore baking soda should help prevent cavities."

Strength: This is where hypotheses come from. It's essential for generating ideas to test.

Weakness: Human biology is complex. Things that "should" work often don't; things that "shouldn't" work sometimes do. Experts have been wrong many times.

The Tooth Fairy's note: Much of traditional herbal medicine sits at this level—mechanistic reasoning plus generations of anecdotal observation. That's not nothing, but it's not proof.

Level 2: In Vitro and Animal Studies¶

Red flags to beware of

Claims based solely on in vitro studies ("kills bacteria in the lab!")
Extrapolating animal results directly to humans
Ignoring concentration differences between lab and clinical use

What it is: Studies conducted in test tubes, petri dishes, or laboratory animals.

Example: "Compound X killed S. mutans in a petri dish at concentrations of 0.1%."

Strength: Allows controlled testing of specific mechanisms. Can test things that would be unethical in humans. Relatively inexpensive and fast.

Weakness: The petri dish is not a mouth. Concentrations that work in the lab may be impossible to achieve in vivo. The mouse is not a human. Bioavailability, tissue penetration, and metabolism all affect whether laboratory effects translate to clinical reality.

The Tooth Fairy's note: A lot of the essential oil and botanical research is at this level. The antimicrobial effects are real in the lab; whether they matter in your mouth is a different question.

Level 3: Case Reports and Case Series¶

What it is: Descriptions of what happened to individual patients or small groups of patients.

Example: "We treated 12 patients with propolis mouthwash and observed improvement in their gingivitis."

Strength: Can identify new phenomena, generate hypotheses, document rare outcomes.

Weakness: No control group—you don't know what would have happened without the treatment. Subject to selection bias (maybe only the patients who improved were reported). Placebo effects, natural disease course, and regression to the mean all confound interpretation.

Level 4: Case-Control Studies¶

What it is: Researchers identify people with a condition (cases) and people without it (controls), then look backward to see what exposures differ between the groups.

Example: "We compared 100 people with severe caries to 100 people with healthy teeth and found that the caries group consumed more sugary beverages."

Strength: Efficient for studying rare conditions; can examine multiple exposures.

Weakness: Retrospective—relies on memory and records. Prone to recall bias (sick people may remember exposures differently). Establishes association, not causation.

Level 5: Cohort Studies¶

What it is: Researchers follow a group of people over time, tracking exposures and outcomes.

Example: "We followed 5,000 people for 10 years, measuring their fluoride exposure and tracking cavity development."

Strength: Can establish temporal sequence (exposure came before outcome); measures actual incidence.

Weakness: Expensive and time-consuming; participants may drop out; confounding factors may not be adequately controlled.

Level 6: Randomized Controlled Trials (RCTs)¶

What it is: Participants are randomly assigned to receive either the treatment being tested or a control (placebo or standard treatment). Neither participants nor assessors know who received what (double-blinding).

Example: "We randomized 200 participants to either xylitol gum or placebo gum for 2 years and measured new cavity development."

Strength: The gold standard for determining whether a treatment causes an effect. Randomization minimizes confounding; blinding minimizes bias.

Weakness: Expensive; may not be feasible for some questions; artificial conditions may not reflect real-world use.

Key questions to ask:

How were participants randomized?
Was the study double-blinded?
How many participants completed the study?
Were outcomes assessed objectively?
Was the study registered before it was conducted?

Level 7: Systematic Reviews and Meta-Analyses¶

What it is: Researchers systematically search for all available studies on a question, assess their quality, and synthesize findings—sometimes mathematically combining results (meta-analysis).

Example: "We identified 23 RCTs of xylitol for caries prevention, pooled their results, and found a statistically significant 25% reduction in new cavities."

Strength: Reduces impact of individual study biases; provides more precise estimates; can identify patterns across studies.

Weakness: Only as good as the underlying studies; publication bias means positive studies are more likely to be published.

The gold standard organizations:

Cochrane Collaboration: Produces the most rigorous systematic reviews in healthcare
PRISMA guidelines: Standard for how systematic reviews should be conducted and reported

Practical Translation: What This Means for You¶

When you encounter a claim about oral health—whether in this book, on a product label, or from a friend—ask these questions:

Question 1: What's the actual evidence?¶

Claim Type	What to Look For
"Kills 99% of germs"	In vitro data—probably; clinical relevance—uncertain
"Clinically proven"	Look for published RCTs; "clinically proven" is often marketing
"Traditional use"	Historical precedent, but not proof of efficacy
"Recommended by dentists"	Often paid endorsements; look for independent studies
"Natural"	Not an evidence claim at all

Question 2: In whom was it tested?¶

A study conducted in:

Healthy young adults may not apply to elderly people with dry mouth
People with active disease may not apply to prevention in healthy people
Laboratory conditions may not apply to real-world use

Question 3: What was the comparison?¶

"Better than nothing" is different from "better than current best practice."

If a mouthwash is compared to plain water, beating the control is unimpressive. If it's compared to chlorhexidine and shows equivalence, that's meaningful.

Question 4: How big was the effect?¶

Statistical significance is not the same as clinical significance.

A study might show a "statistically significant" 2% reduction in plaque. But is 2% meaningful for your oral health? Probably not.

Look for:

Absolute differences (not just percentages)
Number needed to treat (how many people need to use this for one person to benefit)
Clinical outcomes (actual cavities, not just surrogate markers like bacterial counts)

Question 5: Who funded the study?¶

Industry-funded studies are not automatically wrong, but they do show a systematic tendency toward favorable results for the sponsor's product. Independent replication increases confidence.

Why Traditional Remedies Often Lack "Evidence"¶

A note on fairness: The absence of high-level evidence doesn't mean something doesn't work. It often means no one has funded the research.

Clinical trials are expensive—often millions of dollars for a single RCT. Pharmaceutical companies invest in trials because they expect to recoup costs through patented products. But you cannot patent:

Salt
Baking soda
Sage tea
The miswak tree

No one makes money from these, so no one funds rigorous trials. This creates a systematic bias in the evidence base toward patentable, profitable interventions.

This is why I've tried, throughout this book, to:

Distinguish between "no evidence" and "evidence of no effect"
Give traditional remedies the benefit of biological plausibility when mechanisms are sound
Be transparent about the evidence level for each claim

The Tiered Approach I Use in This Book¶

Throughout this book, I've used an informal tiering system:

Tier 1: Strong Evidence¶

Multiple RCTs or systematic reviews support efficacy
Clear mechanism of action
Widely accepted in the scientific community

Examples: Fluoride for caries prevention; chlorhexidine for plaque/gingivitis reduction; xylitol for caries prevention

Tier 2: Moderate Evidence¶

Some RCTs, though perhaps small or short-term
In vitro data supports mechanism
Traditional use is extensive
Expert consensus is generally positive

Examples: Nano-hydroxyapatite for remineralization; propolis for gingivitis; green tea catechins; oil pulling

Tier 3: Promising but Limited¶

Mostly in vitro and animal data
Limited or no clinical trials
Strong traditional use or mechanistic plausibility
Worth trying, but don't expect certainty

Examples: Many individual essential oils; some botanicals like myrrh or calendula

Tier 4: Traditional Use Only¶

Long history of use
Mechanistic plausibility
No modern research
May be effective, but we genuinely don't know

Examples: Some traditional chew sticks from species other than Salvadora persica

Tier 5: Unlikely or Disproven¶

Evidence suggests no benefit
Or potential for harm outweighs plausible benefit

Examples: Activated charcoal for whitening (unproven, potentially harmful); high-dose hydrogen peroxide rinses (tissue damage)

How to Stay Updated¶

Science evolves. What I've written in this book reflects evidence available at the time of writing. To stay current:

Reliable Sources¶

PubMed (pubmed.ncbi.nlm.nih.gov) - Free database of biomedical literature - Search for: [topic] AND (systematic review OR meta-analysis) for highest-level evidence

Cochrane Library (cochranelibrary.com) - Gold-standard systematic reviews - Oral health reviews available free or via library access

Journal of Dental Research - Peer-reviewed primary research and reviews - High-quality but may require library access

AADR/IADR - American and International Associations for Dental Research - Conferences feature cutting-edge research (often before publication)

Red Flags for Unreliable Sources¶

Claims of miracle cures
Hostility toward "mainstream" medicine
Selling products while providing health information
No citations or references to scientific literature
Conspiracy theories about suppression of natural cures
Testimonials as primary evidence
"Doctors don't want you to know"

A Final Note on Certainty¶

I've lived long enough to watch scientific consensus change many times. Things that were "proven" have been overturned. Things that were dismissed have been validated.

What I've learned is to hold conclusions loosely—to follow the evidence where it leads while acknowledging uncertainty. The best scientific thinkers are not those with the most certainty, but those who update their beliefs appropriately as new evidence emerges.

When I tell you that something has "strong evidence," I mean that the current weight of evidence supports it—not that it's unquestionable truth forever. When I tell you something is "promising but limited," I'm being honest about the gaps in our knowledge.

This uncertainty isn't a weakness of science. It's a feature. It's what allows knowledge to grow and improve over time.

Read critically. Ask questions. Update your beliefs when the evidence changes.

And when someone tells you they have all the answers about oral health—or anything else—that's precisely when you should be most skeptical.

The truth is out there, as they say. But finding it requires patience, humility, and a willingness to be wrong. I've been wrong many times in my long existence. I'll be wrong again. What I won't do is stop learning.