MCID and MDC Calculator: SEM and Reliable Change

How to Use the MCID and MDC Calculator

Enter the measurement error in one of three ways: a standard deviation and a test retest reliability coefficient from the same study, the SD of test retest differences, or a published standard error of measurement (SEM).
Choose the confidence level for the minimal detectable change. Both 95% and 90% appear in the literature, for example 90% in studies of physical performance tests[1]. Decide before you look at the data and report which one you used.
Add the baseline SD of your own sample if you want distribution based benchmarks such as half a standard deviation.
Add one person's baseline and follow up scores to see whether their change is larger than measurement error, using the reliable change index[2].
Look up or enter an anchor based MCID to compare it with the MDC. The lookup shows each published estimate with its population and anchor, because an MCID only applies to a population like the one it came from.

The calculator runs in your browser and shows its working. Our clinical instruments directory covers scoring and interpretation for more than 350 questionnaires and clinical assessments, a good place to start when you need a reliability study for your instrument. If you collect the questionnaires in a study, the WeGuide ePRO form builder scores them automatically.

What MCID, MDC and SEM Mean

Minimal clinically important difference (MCID)

Jaeschke and colleagues defined the MCID in 1989 as the smallest difference in score that patients perceive as beneficial and that would mandate a change in the patient's management, in the absence of troublesome side effects and excessive cost[3][4]. Related terms include minimal important difference (MID) and minimal important change (MIC), the term used in COSMIN publications[5].

Standard error of measurement (SEM)

The typical size of the measurement error in one person's score, expressed in the units of the instrument[6]. Unlike a reliability coefficient, it depends much less on how varied the study sample was[7].

Minimal detectable change (MDC)

The smallest change in one person's score that is larger than measurement error with a chosen confidence, also called the smallest detectable change (SDC) or smallest real difference[6][8]. It answers whether a change is real, not whether it matters.

Reliable change index (RCI)

One person's change divided by the standard error of a difference. Jacobson and Truax proposed it to judge whether an individual changed beyond what measurement error alone would produce[2][9].

The Formulas Used Here

The SEM comes from one of three routes. With an SD and a reliability coefficient r, the SD should be the pooled SD of test and retest from the same study that reported r[10]:

SEM = SD × √(1 − r)
SEM = SD of test retest differences ÷ √2

The MDC multiplies the SEM by √2, because a change score carries the error of two measurements, and by the two sided z value for the chosen confidence (1.645 for 90%, 1.960 for 95%)[6]:

MDC = z × √2 × SEM

The reliable change index uses the same standard error of a difference. A change is reliable when the absolute RCI is larger than z, which is the same test as asking whether the change is larger than the MDC[2][9][11]:

RCI = (follow up − baseline) ÷ (√2 × SEM)

When repeated testing improves scores on its own, the practice adjusted version subtracts the mean change seen in a stable reference group before dividing[9][12]. For a mean change in a group of n people, the smallest detectable change shrinks by the square root of the sample size[13][14]:

Group SDC = MDC ÷ √n

Interpreting Change in People and Groups

The MDC and the MCID answer different questions. COSMIN rates measurement error as sufficient only when the smallest detectable change is smaller than the minimal important change[5].

One person

Use the MDC or the reliable change index to decide whether an individual changed beyond measurement error. Only then ask whether the change is large enough to matter to the patient.

Groups

Measurement error shrinks with the square root of the sample size for a group mean, so a mean change smaller than the individual MDC can still be detected[14][13].

When the MDC is larger than the MCID

In 164 shoulder patients, every questionnaire studied had a smallest detectable change slightly larger than its minimal important change, so an individual change of MCID size could not be separated from error[14].

Why this is common

By our calculation from the formulas above, MDC95 is larger than half a standard deviation whenever reliability is below about 0.97. Most instruments are below that, so individual change needs careful interpretation.

Anchor Based or Distribution Based?

Anchor based methods link change on the instrument to an external criterion of important change, most often a global rating of change. Distribution based methods use statistics of the sample, such as the SD or the SEM. Recommended practice treats anchor based estimates as primary and distribution based estimates as supporting evidence[15].

The FDA has said distribution based methods are not appropriate as the sole basis for a responder definition[16], and that they do not directly consider the patient voice[17]. Half a standard deviation is a useful benchmark: across 38 studies, minimal important differences clustered close to 0.5 SD[18]. One SEM approximated patient based MCIDs in heart failure and chronic lung disease questionnaires[19][20]. Neither replaces an anchor based estimate.

Anchors vary in quality. A credibility instrument for anchor based MIDs asks, among other things, whether the anchor is patient rated and whether it correlates with the instrument at 0.5 or more[21].

Why an MCID Depends on Context

An MID is not a fixed property of an instrument and may vary by population and context[15]. The published values in the lookup show this:

Population. For 6 minute walk distance, the anchor based estimate was 25 m in people with COPD in pulmonary rehabilitation[22] and 17.8 m in moderately frail Asian older adults[23]. For a 0 to 10 pain scale it was about 2 points in chronic pain trials[24] and 1.3 points in mechanical neck pain[25].
Baseline severity. In a UK primary care cohort, people with milder depressive symptoms needed larger percentage reductions on the PHQ-9 to report feeling better[26]. Testing this by splitting people on their baseline score can create a false impression of baseline dependency, so the method matters[27].
Direction. The same cohort showed smaller increases in those feeling worse than reductions in those feeling better[26]. The FDA lists equal thresholds for improvement and deterioration as an assumption to check[17].

Use an estimate from a population close to yours, report the range when several exist, and prespecify the threshold.

Using the MCID in Trials

Sample size. DELTA2 guidance favours anchor and opinion based methods for choosing a target difference, advises against the distribution method, and treats a standardised effect size alone as a last resort[28].

Responder analysis. A within patient threshold can define responders, and the FDA asks sponsors to prespecify a single threshold for a responder endpoint[17]. A threshold below the individual MDC will count some people as responders because of measurement error, so report the MDC next to it.

Reliability evidence. FDA guidance recommends intraclass correlation coefficients to estimate test retest reliability in most cases, and says internal consistency alone may not be enough[29]. COSMIN rates reliability as sufficient at an ICC of 0.70 or more[5]. For test retest studies, Koo and Li recommend the absolute agreement form of the ICC[30]. When a questionnaire moves from paper to an app, our guide to eCOA validation covers the equivalence testing that regulators expect.

Common Mistakes to Avoid

Using Cronbach's alpha for the SEM. Alpha reflects a single occasion, so it misses variation between occasions[10][6].
Using an ICC from another population with your own SD. Mokkink and colleagues describe this as a common misuse[10].
Calling 0.5 SD or 1 SEM the MCID. The widely quoted 5 point MCID for the PHQ-9 was estimated as 2 SEMs, a distribution based figure[31].
Treating the MDC as the MCID. The MDC tells you about noise, not about what patients value, and it should not replace the minimal important difference[32].
Applying a group threshold to individuals, or a within patient threshold to a difference between group means[17].
Forgetting √2. A change score carries the error of two measurements[6].
Using Pearson's r as the reliability, which stays high even when everyone's score shifts by the same amount[9].

Collecting PROs over time?

WeGuide schedules ePRO questionnaires, scores them automatically and keeps participants completing their follow ups, so change can be tracked for every person in a study or registry.

Book a demo

Frequently Asked Questions

Answers about minimal clinically important difference, minimal detectable change and how to use them in research.

Have another question?

What is the difference between the MCID and the MDC?

The MDC is the smallest change in one person's score that is larger than measurement error, 1.96 × √2 × SEM at 95% confidence. The MCID is the smallest change that patients regard as important, found with an external anchor such as a global rating of change. One is about noise, the other is about value, so report both.

Which reliability coefficient should I use?

Use a test retest intraclass correlation coefficient from stable patients, in its absolute agreement form for a single measurement, and ideally from the same sample as the standard deviation. FDA guidance recommends ICCs for test retest reliability in most cases, and COSMIN rates reliability as sufficient at 0.70 or more.

Can I use Cronbach's alpha to calculate the SEM?

You can, but the result will usually be too small. Alpha comes from a single occasion, so it leaves out the variation between occasions that matters for change scores. FDA guidance says internal consistency alone may not be sufficient evidence of reliability for clinical trials. Use a test retest ICC if one has been published.

Is half a standard deviation a valid MCID?

It is a useful benchmark, not an MCID. Across 38 studies, minimal important differences clustered near 0.5 SD, but the figure is not anchored in what patients value. The FDA does not accept distribution based values as the sole basis for a responder definition, so label them as distribution based when you report them.

Why is the MDC bigger than the MCID for my instrument?

A change score combines the error of two measurements, so unless reliability is close to 1 the individual MDC exceeds typical important differences. By our calculation, MDC95 is larger than half a standard deviation whenever reliability is below about 0.97. Interpret individual change with care and analyse group means where you can.

How do I calculate the reliable change index?

Divide the change from baseline to follow up by the standard error of a difference, √2 × SEM. An absolute value above 1.96 means the change is unlikely to be measurement error alone at 95% confidence. If repeated testing improves scores on its own, subtract the mean practice effect from the change first.

Should I report MDC90 or MDC95?

Both appear in the literature: MDC90 is common in studies of physical performance tests and the 95% version in many patient reported outcome studies. Choose the level before you look at the data and state which one you used, because the 90% threshold is about 16% smaller than the 95% one.

Can I use a published MCID in my study population?

Only with care. MCIDs vary with the population, the anchor, the baseline severity and the direction of change. Prefer an estimate from a similar population with a credible, patient rated anchor, report the range when several estimates exist, and prespecify the threshold in your protocol or analysis plan.

How do I use the MCID in a sample size calculation?

Use an anchor based or opinion based important difference as the target difference. DELTA2 guidance advises against distribution based targets and treats a standardised effect size alone as a last resort. A difference between group means is not the same as a within patient threshold, so check which one your estimate describes.

Is the MCID the same for getting better and getting worse?

Not necessarily. In one UK primary care cohort, changes among people feeling worse were smaller than among those feeling better, while a review of many studies found no consistent pattern. FDA draft guidance asks sponsors to check this assumption. This calculator applies the same threshold in both directions and says so.

Free tools for planning and running clinical research, built on published methods.

References

Ortega-Pérez de Villar L, Martínez-Olmos FJ, Junqué-Jiménez A, et al. Test-retest reliability and minimal detectable change scores for the short physical performance battery, one-legged standing test and timed up and go test in patients undergoing hemodialysis. PLoS One. 2018;13(8):e0201035. PubMed 30133445
Jacobson NS, Truax P. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. J Consult Clin Psychol. 1991;59(1):12-19. PubMed 2002127
Jaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference. Control Clin Trials. 1989;10(4):407-415. PubMed 2691207
Mouelhi Y, Jouve E, Castelli C, et al. How is the minimal clinically important difference established in health-related quality of life instruments? Review of anchors and methods. Health Qual Life Outcomes. 2020;18(1):136. PubMed 32398083
Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Qual Life Res. 2018;27(5):1147-1157. PubMed 29435801
de Vet HC, Terwee CB, Ostelo RW, et al. Minimal changes in health status questionnaires: distinction between minimally detectable change and minimally important change. Health Qual Life Outcomes. 2006;4:54. PubMed 16925807
de Vet HC, Terwee CB, Knol DL, et al. When to use agreement versus reliability measures. J Clin Epidemiol. 2006;59(10):1033-1039. PubMed 16980142
Beckerman H, Roebroeck ME, Lankhorst GJ, et al. Smallest real difference, a link between reproducibility and responsiveness. Qual Life Res. 2001;10(7):571-578. PubMed 11822790
Duff K. Evidence-based indicators of neuropsychological change in the individual patient: relevant concepts and methods. Arch Clin Neuropsychol. 2012;27(3):248-261. PubMed 22382384
Mokkink LB, Eekhout I, Boers M, et al. Studies on reliability and measurement error of measurements in medicine: from design to statistics explained for medical researchers. Patient Relat Outcome Meas. 2023;14:193-212. PubMed 37448975
Gronchi G, Scoccianti S, Peru A. Cognitive decline after cranial irradiation: hoping for a systematic application of the reliable change index. Front Psychol. 2020;11:573919. PubMed 33123053
Chelune GJ, Naugle RI, Lüders H, Sedlak J, Awad IA. Individual change after epilepsy surgery: practice effects and base-rate information. Neuropsychology. 1993;7(1):41-52. doi:10.1037/0894-4105.7.1.41
Khadka J, Milte R, Hutchinson C, et al. Reliability of the quality of life-aged care consumers (QOL-ACC) and EQ-5D-5L among older people using aged care services at home. Health Qual Life Outcomes. 2024;22(1):40. PubMed 38816717
van Kampen DA, Willems WJ, van Beers LW, et al. Determination and comparison of the smallest detectable change (SDC) and the minimal important change (MIC) of four-shoulder patient-reported outcome measures (PROMs). J Orthop Surg Res. 2013;8:40. PubMed 24225254
Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol. 2008;61(2):102-109. PubMed 18177782
U.S. Food and Drug Administration. Guidance for Industry. Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims. December 2009. fda.gov
U.S. Food and Drug Administration. Patient-Focused Drug Development: Incorporating Clinical Outcome Assessments Into Endpoints for Regulatory Decision-Making. Draft guidance. April 2023. fda.gov
Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582-592. PubMed 12719681
Wyrwich KW, Nienaber NA, Tierney WM, Wolinsky FD. Linking clinical relevance and statistical significance in evaluating intra-individual changes in health-related quality of life. Med Care. 1999;37(5):469-478. PubMed 10335749
Wyrwich KW, Tierney WM, Wolinsky FD. Further evidence supporting an SEM-based criterion for identifying meaningful intra-individual changes in health-related quality of life. J Clin Epidemiol. 1999;52(9):861-873. PubMed 10529027
Devji T, Carrasco-Labra A, Qasim A, et al. Evaluating the credibility of anchor based estimates of minimal important differences for patient reported outcomes: instrument development and reliability study. BMJ. 2020;369:m1714. PubMed 32499297
Holland AE, Hill CJ, Rasekaba T, et al. Updating the minimal important difference for six-minute walk distance in patients with chronic obstructive pulmonary disease. Arch Phys Med Rehabil. 2010;91(2):221-225. PubMed 20159125
Kwok BC, Pua YH, Mamun K, Wong WP. The minimal clinically important difference of six-minute walk in Asian older adults. BMC Geriatr. 2013;13:23. PubMed 23510291
Farrar JT, Young JP Jr, LaMoreaux L, et al. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain. 2001;94(2):149-158. PubMed 11690728
Cleland JA, Childs JD, Whitman JM. Psychometric properties of the Neck Disability Index and Numeric Pain Rating Scale in patients with mechanical neck pain. Arch Phys Med Rehabil. 2008;89(1):69-74. PubMed 18164333
Kounali D, Button KS, Lewis G, et al. How much change is enough? Evidence from a longitudinal study on depression in UK primary care. Psychol Med. 2022;52(10):1875-1882. PubMed 33138872
Terluin B, Roos EM, Terwee CB, et al. Assessing baseline dependency of anchor-based minimal important change (MIC): don't stratify on the baseline score! Qual Life Res. 2021;30(10):2773-2782. PubMed 34041680
Cook JA, Julious SA, Sones W, et al. DELTA2 guidance on choosing the target difference and undertaking and reporting the sample size calculation for a randomised controlled trial. BMJ. 2018;363:k3750. PubMed 30560792
U.S. Food and Drug Administration. Patient-Focused Drug Development: Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments. Guidance. October 2025. fda.gov
Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155-163. PubMed 27330520
Löwe B, Unützer J, Callahan CM, et al. Monitoring depression treatment outcomes with the patient health questionnaire-9. Med Care. 2004;42(12):1194-1201. PubMed 15550799
de Vet HC, Terwee CB. The minimal detectable change should not replace the minimal important difference. J Clin Epidemiol. 2010;63(7):804-805. PubMed 20399609

Formulas and test values checked against the published sources and 59 reference calculations on 20 October 2026 by the WeGuide editorial team. This tool is for research planning and education. It is not a medical device and does not give medical advice.