Special Article 5, Issue 13.2

Critical Appraisal in Dental Sleep Medicine: A Guide for Clinicians

Abstract:

Dental sleep medicine clinicians are inundated with new research findings, often directly from industry. This article presents a practical framework for critically evaluating these materials by asking straightforward questions: Does the study measure what matters? Is the study’s design strong enough to give reliable results? Are the reported statistically significant differences clinically relevant?  Do the conclusions match what the data show? Are there financial relationships that could influence the findings? By applying these questions systematically, clinicians can distinguish reliable evidence from marketing claims and make better-informed treatment decisions to benefit their patients.

Citation:

Falardo Ramos S, et al. Critical Appraisal in Dental Sleep Medicine: A Guide for Clinicians. J Dent Sleep Med. 2026;13(2).

INTRODUCTION 

Evidence-based practice in dental sleep medicine depends on the ability to interpret and apply research findings to patient care. However, most dentists encounter research findings not just through peer-reviewed research, but through industry materials, conference presentations, and sales representatives. These sources may unintentionally introduce bias by selectively framing results in ways that favor products. Because of the competing time restraints of clinical practice, practitioners involved in dental sleep medicine need efficient ways to evaluate whether a study's findings are trustworthy and applicable to their patients.

This article offers a structured approach organized around five practical questions:

• Does the study measure what matters?
• Is the study’s design strong enough to give reliable results?
• Are the statistically significant findings clinically significant?
• Does the conclusion align with the data presented?
• Are there financial relationships that could influence the findings?

Question 1: Does the study measure what matters?

Before anything else, consider whether the study is measuring something clinically relevant. This is particularly important in dental sleep medicine, where studies sometimes report surrogate outcomes. These are indirect measures, such as biomarkers or imaging findings, used as stand-ins for other clinically relevant outcomes.¹ For example, a study might report changes in airway dimensions; however, improvements in the apnea-hypopnea index (AHI), daytime sleepiness, snoring, and quality of life are significantly more important in practice.

 Some questions to consider are:

• Are the outcomes measured relevant to clinical practice?
• Would improvement in the reported outcomes matter to your patients?

Studies relying on surrogate outcomes should ideally also report patient-centered measures. When they do not, practitioners should consider whether the surrogate has been validated against the outcomes that matter to their patients. Novel surrogate measures may deserve consideration, because they may reveal important insights. However, their clinical value remains uncertain until compared to established standards. Changes in an unvalidated surrogate measure tells very little about whether patients will benefit. Table 1 outlines examples of surrogate, intermediate, and patient-centered outcomes related to obstructive sleep apnea (OSA).

Question 2: Is the study’s design strong enough to give reliable results?

Even when a study measures the intended outcome, the design must be rigorous enough to provide a trustworthy answer. Study design determines whether findings reflect a true treatment effect or result from bias, chance, or confounding. Several aspects of study design can be signals to be careful with interpreting results. Limitations in the study design can be signals to be careful when interpreting and applying the results.

Small sample size: Small studies are poorly suited to detect modest effects and are more prone to producing exaggerated or spurious findings.² A study with 15 participants showing dramatic results should be viewed with more skepticism than a study with 150 participants showing modest results. When evaluating smaller studies, whether the researchers provided a justification for their sample size should be considered.³ This can help determine if the study was adequately designed to detect meaningful effects. When a study with few participants reports dramatic results for an outcome where only modest differences are expected, these findings should be viewed with caution. Although case reports continue to have value in the research world, studies supported by power analyses or meta-analyses tend to provide the most definitive evidence.

Control groups: A control group is a set of participants used as a comparison for those receiving the treatment being studied. Control patients might receive no treatment (negative control groups), a placebo, standard care (positive control groups), or an alternative active treatment, depending on the research question. Without a comparison, a practitioner cannot know what would have happened in the absence of the intervention. Improvements might reflect placebo effects (feeling better because that is the expectation), natural fluctuation in symptoms, or regression to the mean (the tendency for extreme values to move closer to average over time).⁴ Findings should be interpreted with caution if no control group or comparison is noted. When control groups exist, the practitioner must check whether they are comparable with the treatment group at the start of the study for characteristics such as age, sex, and baseline severity. If groups differ in important ways before treatment begins, and those differences are not controlled during the statistical analysis, differences at the end may not be due to the treatment itself.

Randomization: To reduce the potential for selection bias, randomization uses chance to determine who receives the new treatment and who receives the comparison.⁵ The expected goal is that the groups are similar at the start so that any differences in outcomes are more likely due to the treatment itself, not preexisting differences between participants.⁵  Selection bias occurs when the people in one group systematically differ from those in the other because of how they were selected, which can positively or negatively influence results, making a treatment appear better or worse than it really is. Randomization helps prevent this by removing choice from the assignment process. However, randomization is not foolproof, because it is only reliable when group assignments are concealed from researchers and when similar numbers of patients in each group complete the study, which is known as a double-blind study design.⁶

Preregistration: Even a seemingly well-designed study can be misleading if what is reported does not match what was originally planned. One key tool in preventing this is preregistration, in which investigators specify exactly how a study will be conducted, including its primary outcomes and core analysis plans, before the study is ever conducted. Preregistration can help with preventing selective reporting by clarifying what was planned and what was added after the results were known.^7,8 It is important to check the relevant preregistration database and compare the analyses that were originally planned to what is reported in the final study. Common pre-registration platforms include ClinicalTrials.gov for clinical trials, PROSPERO for systematic reviews, and general registries such as the Open Science Framework.^9–11

Question 3: Are the statistically significant findings clinically significant?

Statistics can be easily misused to make findings seem more impressive than they appear. Understanding a few key concepts can help identify when to be skeptical of results.

Beyond P values and statistical significance: The dental sleep medicine practitioner should be wary of “statistically significant” results. A P value answers one specific question: If the treatment had no effect at all, how surprising are these results?¹² For example, a value of P = 0.03 means that if the treatment truly did nothing, results in this extreme would be seen approximately 3 times out of 100.¹³ Because the conventional measure of “statistical significance” is a value of P < 0.05, a common mistake is treating P < 0.05 as proof that a treatment works and P ≥ 0.05 as proof that it does not. Neither is correct.

Effect sizes. Because P values can easily be misinterpreted, how large the effect is (known as the effect size) and whether that difference would matter in practice should always be considered.¹⁴ The effect size indicates how large the difference is between groups. For example, a study might report a statistically significant reduction in AHI, whereas the actual reduction is only two events per hour (the effect size), a difference that is likely too small for any discernable effect. Not every statistically significant result is automatically clinically significant. The difference is noted, but its magnitude would not be a relevant difference maker in practice or for the patient.

Graphs and figures. Visual presentation can exaggerate or minimize findings.¹⁵ The dental sleep medicine practitioner should watch for y-axes that do not start at zero. In statistics, this is called truncating and can create the impression of dramatic results, making small differences look large. The dental sleep medicine practitioner should check that scales are consistent across compared graphs, and look for error bars or confidence intervals; their absence makes it harder to judge uncertainty.

Question 4: Do the conclusions match what the data show?

One of the most common problems in research is making exaggerated claims that go beyond what the actual findings were. This is known as "spin," the use of language that distorts or misrepresents findings, often to make results appear more favorable than they are.¹⁶

Compare carefully:

• What do the data actually show?
• Is there a match between what was measured and what was concluded?
• Do these data truly support the claims made in the paper?

Authors sometimes report modest or nonsignificant effects but describe the treatment as "effective" or "promising" in the abstract or conclusion. Some may also highlight other outcomes that looked favorable while glossing over more important outcomes that may show no effect or even harm.

Question 5: Are there financial relationships that could influence the findings?

CONCLUSION

Evaluating a study for evidence-based practice does not require exhaustive statistical knowledge or formal training in research methods. Instead, it requires asking the right questions based on prior clinical experience. When encountering study data, whether in a journal, conference, or through industry materials, working through these questions can help separate reliable findings from overstated claims. Table 2 provides a quick reference for applying these questions in practice.

It is important to note that the goal of critically appraising research as practitioners is to calibrate the level of confidence in the findings and temper down or up their applicability to day-to-day clinician decisions. Doing so helps practitioners involved in dental sleep medicine discern between robust findings that would apply to clinical practice and preliminary data that should be interpreted with caution. Maintaining a discerning approach to both scientific and industry data ensures that the future of dental sleep medicine continues to advance through strong evidence, integrity, and patient-centered care.

REFERENCES

1. Christensen R, Ciani O, Manyara AM, Taylor RS. Surrogate endpoints: A key concept in clinical epidemiology. J Clin Epidemiol. 2024;167:111242. doi:10.1016/j.jclinepi.2023.111242
2. Faber J, Fonseca LM. How sample size influences research outcomes. Dent Press J Orthod. 2014;19(4):27-29. doi:10.1590/2176-9451.19.4.027-029.ebo
3. Lakens D. Sample Size Justification. Ravenzwaaij D van, ed. Collabra Psychol. 2022;8(1):33267. doi:10.1525/collabra.33267
4. Hróbjartsson A, Gøtzsche PC. Is the placebo powerless? An analysis of clinical trials comparing placebo with no treatment. N Engl J Med. 2001;344(21):1594-1602. doi:10.1056/NEJM200105243442106
5. Tripepi G, Jager KJ, Dekker FW, Zoccali C. Selection bias and information bias in clinical research. Nephron Clin Pract. 2010;115(2):c94-99. doi:10.1159/000312871
6. Sverdlov O, Rosenberger WF. Randomization in clinical trials: Can we eliminate bias? Clin Investig. 2013;3(1):37-47. doi:10.4155/cli.12.130
7. Nosek BA, Ebersole CR, DeHaven AC, Mellor DT. The preregistration revolution. Proc Natl Acad Sci U S A. 2018;115(11):2600-2606. doi:10.1073/pnas.1708274114
8. Lakens D, Mesquida C, Rasti S, Ditroilo M. The benefits of preregistration and registered reports. Evidence-Based Toxicology. 2024;2(1):2376046. doi:10.1080/2833373X.2024.2376046
9. National Library of Medicine. ClinicalTrials.gov. https://clinicaltrials.gov/. Accessed January 26, 2026.
10. University of York. PROSPERO: International Prospective Register of Systematic Reviews. Centre for Reviews and Dissemination. https://www.crd.york.ac.uk/prospero/. Accessed January 26, 2026.
11. Center for Open Science. Open Science Framework. https://osf.io/. Accessed January 26, 2026.
12. Rafi Z, Greenland S. Semantic and cognitive tools to aid statistical science: Replace confidence and significance by compatibility and surprise. BMC Med Res Methodol. 2020;20(1):244. doi:10.1186/s12874-020-01105-9
13. Greenland S, Senn SJ, Rothman KJ, et al. Statistical tests, P values, confidence intervals, and power: A guide to misinterpretations. Eur J Epidemiol. 2016;31(4):337-350. doi:10.1007/s10654-016-0149-3
14. Sullivan GM, Feinn R. Using effect size—or why the P value is not enough. J Grad Med Educ. 2012 Sep;4(3):279–282. doi:10.4300/JGME-D-12-00156.1
15. Cabanski C, Gilbert H, Mosesova S. Can graphics tell lies? A tutorial on how to visualize your data. Clin Transl Sci. 2018;11(4):371-377. doi:10.1111/cts.12554
16. Chiu K, Grundy Q, Bero L. “Spin” in published biomedical literature: A methodological systematic review. PLoS Biol. 2017;15(9):e2002173. doi:10.1371/journal.pbio.2002173
17. Dunn AG, Coiera E, Mandl KD, Bourgeois FT. Conflict of interest disclosure in biomedical research: A review of current practices, biases, and the role of public registries in improving transparency. Res Integr Peer Rev. 2016;1:1. doi:10.1186/s41073-016-0006-7
18. Crossley JR, Wallerius K, Hoa M, Davidson B, Giurintano JP. Association between conflict of interest and published position on hypoglossal nerve stimulation for sleep apnea. Otolaryngol Head Neck Surg. 2021;165(2):375-380. doi:10.1177/0194599820982914
19. Hu Q, Acharya A, Leung WK, Pelekos G. Sponsorship bias in clinical trials in the dental application of probiotics: A meta-epidemiological study. Nutrients. 2022;14(16):3409. doi:10.3390/nu14163409

SUBMISSION & CORRESPONDENCE INFORMATION

Submitted for publication January 28, 2026
Accepted for publication February 27, 2026

Address correspondence to: Susana Falardo Ramos, DDS, MSc, PhD; Email: susana.falardo@gmail.com
 

DISCLOSURE STATEMENT