Beyond the P Value: From Statistical Metrics to Clinical Decision-Making

Luciano Rossi, PhD, ARGENTINA Ignacio Pasqualini, MD, ARGENTINA

ISAKOS eNewsletters   Current Perspective 2026   rating

Introduction

Two randomized trials evaluating different techniques for rotator cuff repair are published the same month. The first reports a statistically significant 2-point improvement in ASES scores (p = 0.01). The second reports a 12-point improvement that fails to reach significance (p = 0.07).

Which study should change your practice?

For decades, surgeons have been trained to answer this question by looking first at the p value. Statistical significance remains the cornerstone of evidence-based medicine, helping us to distinguish true treatment effects from random variation. But it answers only one part of a much larger clinical question.

Patients do not experience p values. They experience pain relief, functional recovery, return to sport, and quality of life. Surgeons, in turn, do not decide based solely on statistical tests; every recommendation integrates evidence with patient expectations, expertise, risk, cost, and judgment.

Over the past decade, orthopaedic sports medicine has progressively shifted from interpreting research through statistical significance alone toward a more patient-centered approach. Concepts such as the Minimum Clinically Important Difference (MCID), Substantial Clinical Benefit (SCB), and Patient Acceptable Symptom State (PASS) increasingly appear in clinical research because they better reflect what matters to patients rather than investigators.

Yet even these measures do not fully answer the question that every surgeon ultimately faces. A statistically significant result may fail to improve a patient's life; a clinically meaningful improvement may not justify changing surgical practice. Rather than competing concepts, statistical and clinical significance should be viewed as sequential steps within a broader framework of clinical decision-making (Figure 1).

This article proposes a practical framework built around three questions:

• Is the difference real?
• Will my patient notice the difference?
• Should this evidence change my clinical practice?

Figure 1. The three levels of evidence interpretation—from statistical significance, to clinical significance, to clinical relevance.

Is the Difference Real?

Every clinical investigation begins with the same challenge: determining whether an observed effect is genuine or merely reflects random variation. Statistical significance addresses scientific validity, not clinical importance. Four concepts deserve attention here: the p value, confidence intervals, effect size, and statistical power.

The p value. For generations, clinicians have been taught that a p value below 0.05 means an observed difference is unlikely to have occurred by chance. This is important, but incomplete. The p value does not quantify the magnitude of the effect, does not indicate whether patients perceive the improvement, and does not determine whether findings justify changing practice. It answers only one question: is the observed difference likely to be real? Confusing statistical significance with clinical importance remains one of the most common errors in interpreting orthopaedic research. Strictly defined, the p value is the probability of obtaining a difference at least as large as the one observed, assuming that the null hypothesis (i.e., that there is no true difference between the treatments compared) is true. When this probability falls below a prespecified threshold (conventionally 0.05), the null hypothesis is rejected. Importantly, the p value is neither the probability that the null hypothesis is true nor the probability that the observed result is due to chance alone.

Confidence intervals. Rather than fixating on the p value, clinicians should examine the confidence interval, which estimates both the magnitude and precision of an effect. A narrow interval suggests a precise, reproducible estimate; a wide one reflects uncertainty. Two ACL reconstruction studies might report an identical average improvement, yet one shows a confidence interval of 8–10 points and the other 1–17 points—both statistically significant, but very different in certainty.

Effect size. Once significance is established, the next question is how large the effect actually is. Effect size reflects practical magnitude independent of sample size. In large multicenter cohorts of hundreds or thousands of patients, minimal differences can achieve highly significant p values despite offering little real benefit; conversely, smaller studies may show substantial improvements that fail to reach significance because of limited statistical power.

Statistical power. Failure to reach significance should never be read as proof of equivalence. Every test carries the risk of a Type-II error—a false-negative result driven by insufficient power. This is especially relevant in sports medicine trials, which often enroll small cohorts because of strict eligibility criteria and long follow-up. A non-significant result may simply mean that the study lacked a sufficient number of patients needed to detect a real difference.

When significance misleads. Consider a multicenter study of 2,000 patients undergoing rotator cuff repair, reporting a statistically significant 2-point ASES improvement after biologic augmentation (p < 0.001). Would any patient truly perceive that improvement? Probably not. Now consider forty collision athletes undergoing a novel stabilization procedure who improve by 12 ASES points, with lower recurrence and earlier return to sport—but a p value of 0.07 because of limited sample size. Should those findings be dismissed? Again, probably not (Figure 2).

Statistical significance is the beginning of evidence interpretation, not the end. Knowing that an effect is likely real does not mean that it is meaningful to patients—to answer that, we must move beyond statistics.

Figure 2. Statistical significance depends on sample size as much as on the true magnitude of benefit: a small effect in a large trial can be “significant,” while a large effect in a small trial may not be.

Will My Patient Notice the Difference?

Two patients return to your office 6 months after shoulder surgery. The first has excellent radiographs, improved motion and strength, and significantly better PROM scores—yet hesitates before saying, “Doctor... I don't really feel that much better.” The second shows only moderate improvement in outcome scores but has returned to tennis, sleeps pain-free, and tells you, “I finally have my shoulder back.”

Which patient had the better outcome?

Traditional statistics compare group averages, not individual perception. This limitation drove one of the most important paradigm shifts in outcomes research: clinical significance — not whether a difference exists, but whether it matters to the patient.

MCID—Did the patient notice any improvement? The Minimum Clinically Important Difference is the smallest improvement that patients perceive as beneficial. It does not define an excellent outcome; it marks the threshold beyond which patients begin to recognize meaningful change. A rotator cuff repair patient whose ASES score rises from 42 to 50 may show a statistically significant 8-point gain—but the clinically relevant question is whether that change exceeds the established MCID for that population. The MCID can be calculated using anchor-based or distribution-based methods. Anchor-based methods relate the change in PROM scores to an external criterion, typically the patient’s own global rating of change, and define the MCID as the mean change in score among patients who report being “slightly” or “somewhat” better. Distribution-based methods rely on the statistical variability of the scores; the most widely used approach defines the MCID as one-half of the standard deviation (0.5 SD) of the group’s baseline score. Because both approaches are derived from group means and variability, the MCID describes an average threshold for a population and should not be interpreted as a guaranteed cutoff for any individual patient.

SCB—Was the improvement truly meaningful? The Substantial Clinical Benefit represents a considerably higher threshold, distinguishing patients who are merely “better” from those who are “doing well.” A recreational runner may consider moderate pain relief satisfactory; a professional athlete returning to elite competition typically requires far greater recovery. SCB is particularly useful when evaluating athletic populations.

PASS—Has the patient reached an acceptable state? Rather than measuring improvement, PASS asks whether the patient’s current condition is satisfactory. Consider two patients undergoing the same stabilization procedure, both improving by 25 PROM points. One reaches a postoperative score of 92 points and resumes competitive sport, whereas the other improves from a very poor baseline but reaches only 65 points, remaining unable to return to prior activity. Both achieved the same improvement—but only one reached an acceptable state (Figure 3).

Figure 3. MCID, SCB, and PASS are complementary, not interchangeable, milestones along the patient's recovery journey.

These metrics are complementary, not interchangeable: MCID identifies whether meaningful improvement occurred, SCB determines whether the improvement was substantial, and PASS determines whether the patient reached a satisfactory state. Modern investigators increasingly report how many patients achieved MCID, SCB, or PASS rather than relying solely on mean PROM changes.

Clinical significance is also not universal. Thresholds are procedure- and population-specific: an MCID established after ACL reconstruction cannot be applied to rotator cuff repair, and a PASS threshold validated for reverse shoulder arthroplasty should not be extrapolated to instability surgery. Baseline disability, age, athletic demands, follow-up duration, and cultural expectations all shape what “meaningful” means.

Ultimately, patients rarely ask whether a treatment achieved statistical significance. They ask, “Doctor, what are the chances that I will actually feel better?” Clinical significance brings us closer to answering that question—but even meaningful improvement does not automatically justify changing surgical practice.

Should This Evidence Change My Clinical Practice?

By this point, surgeons know whether a difference is likely real and whether patients are likely to perceive it as meaningful. One might assume that such information is enough to guide practice. It is not. Every surgeon has seen statistically significant, clinically meaningful findings that never changed daily practice—as well as innovations that were adopted with enthusiasm but were later abandoned once longer follow-up or cost-effectiveness data emerged. Evidence alone does not change clinical practice; that requires interpretation, judgment, and context.

We propose calling this missing step clinical relevance—the practical value of a treatment within the real-world environment in which surgeons and patients actually make treatment decisions. A treatment may be statistically significant and help more patients reach MCID or PASS, yet if those benefits come at substantially higher cost, complexity, or complication risk, widespread adoption may not be justified. Conversely, a simple, low-complication procedure with modest but consistent benefit may carry enormous clinical relevance.

Unlike statistical significance (summarized by a p value) or clinical significance (quantified through MCID, SCB, or PASS), clinical relevance cannot be reduced to a single metric. It emerges from the integration of magnitude of benefit, durability, complication profile, surgical complexity, learning curve, cost-effectiveness, return-to-sport outcomes, patient expectations, resource availability, surgeon experience, and shared decision-making (Figure 4). No single factor is sufficient alone.

Figure 4. Clinical relevance is multifactorial—it emerges from the integration of clinical, practical, and patient-specific considerations, not from any single measure.

Pearls and Pitfalls

Pearls

• Never interpret a p value without considering the magnitude and precision of the treatment effect.
• Report the proportion of patients achieving MCID, SCB, and PASS rather than relying solely on mean PROM improvements.
• Remember that clinically meaningful thresholds are procedure- and population-specific.
• Consider return-to-sport rates, satisfaction, and complications alongside PROMs.
• Integrate evidence with shared decision-making rather than replacing clinical judgment.

Pitfalls

• Assuming statistical significance automatically implies clinical importance.
• Ignoring confidence intervals or effect size.
• Applying MCID or PASS values validated for a different pathology.
• Equating non-significant findings with treatment equivalence.
• Adopting new technologies before considering cost, learning curve, and durability.

Take-Home Messages

• Statistical significance answers whether an observed difference is likely real.
• Clinical significance determines whether patients actually perceive that difference as meaningful.
• Clinical relevance evaluates whether findings justify changing clinical practice.
• These three concepts are complementary, not competitive.
• The best clinical decisions integrate scientific evidence, patient-centered outcomes, and sound surgical judgment.

Future Perspectives

Orthopaedic sports medicine is entering an era of increasingly personalized care. Future research will move beyond average treatment effects toward predicting individualized probabilities of achieving clinically meaningful outcomes. Artificial intelligence, machine learning, and large international registries will let surgeons estimate the likelihood that a specific patient—not merely the average study participant—will achieve MCID, PASS, return to sport, or long-term success.

At the same time, interpretation of clinical research is likely to evolve from isolated statistical comparisons toward comprehensive models integrating statistical significance, clinically meaningful improvement, patient preferences, costs, and value-based care. Rather than asking whether one treatment is statistically superior to another, future investigations should help clinicians to identify which treatment is most likely to benefit the individual patient sitting in front of them.

Ultimately, the future of orthopaedic research lies not in generating more statistically significant results, but in producing evidence that meaningfully improves clinical decision-making and patient care.

References

1. Copay AG, Subach BR, Glassman SD, Polly DW Jr, Schuler TC. Understanding the minimum clinically important difference: a review of concepts and methods. Spine J. 2007;7(5):541-546.
2. Bloom DA, Kaplan DJ, Mojica E, et al. The minimal clinically important difference: a review of clinical significance. Am J Sports Med. 2023;51(2):520-524.
3. Kim DM, Kim TH, Kholinne E, et al. Minimal clinically important difference, substantial clinical benefit, and patient acceptable symptomatic state after arthroscopic rotator cuff repair. Am J Sports Med. 2020;48(11):2650-2659.
4. Nwachukwu BU, Chang B, Adjei J, Schairer WW, Ranawat AS, Kelly BT, Nawabi DH. Time required to achieve minimal clinically important difference and substantial clinical benefit after arthroscopic treatment of femoroacetabular impingement. Am J Sports Med. 2018;46(11):2601-2606.
5. Sun B, Vivekanantha P, Abdel Khalik H, de SA D. Several factors predict the achievement of the patient acceptable symptom state and minimal clinically important difference for patient-reported outcome measures following anterior cruciate ligament reconstruction: a systematic review. Knee Surg Sports Traumatol Arthrosc. 2025;33(5):1617-1632.

Please note: ISAKOS Newsletter Current Perspectives are not peer-reviewed articles. For peer-reviewed articles, please visit the Journal of ISAKOS at jisakos.com.