STATISTICAL SOFTWARE
MedCalc Statistical Analysis Service
Diagnostic Accuracy You Can Defend
Our MedCalc statistical analysis service carries out ROC, diagnostic accuracy and method comparison analysis for clinical research in MedCalc, working from your datasets, with the threshold rule and 95% confidence intervals reported and the data and output provided where agreed.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Is MedCalc Used For in Research?
MedCalc is a commercial statistical package for biomedical research, built around diagnostic accuracy, ROC curves, method comparison and agreement, with the usual group comparisons, regression and survival analysis alongside. Researchers choose it because procedures such as ROC curve comparison, Bland-Altman agreement and Passing-Bablok or Deming regression sit in one data-sheet workflow, and because laboratory and clinical journals expect those analyses. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, provides a MedCalc statistical analysis service in which an in-house PhD-level statistician carries out the agreed analysis in MedCalc, working from your datasets and documenting the threshold and the estimates reported.
WHEN TO ENGAGE
Four Situations in Which Clinical Teams Bring Us a MedCalc Project
Requests for MedCalc biostatistics usually come from a diagnostic accuracy study, a device or assay comparison, a cutoff to justify or results that need a second look.
Accuracy Study to Analyze A test, score or biomarker has been measured against a reference standard, and you need sensitivity, specificity and ROC results estimated with confidence intervals under a stated threshold rule.
Two Methods to Compare A MedCalc method comparison sets a new device or assay against an established one on the same samples, and the question is bias and agreement across the range, not how well the two correlate.
Cutoff That Needs a Rationale A threshold was picked from the same data that now evaluate it, and its choice has to be justified, labeled as exploratory or checked in a separate sample.
Results to Verify MedCalc output already exists and needs a second look, from the prevalence behind the predictive values to the intervals around every accuracy estimate.
ANALYSES IN MEDCALC
Analyses Carried Out in MedCalc for Clinical and Laboratory Data
Each analysis follows the design of the accuracy or agreement study, and its card links to the method page where the choice of measure is explained.
ROC Curve Analysis
MedCalc ROC analysis summarizes how well a continuous marker separates cases from non-cases, with the area under the curve and its 95% interval, and compares correlated curves from the same patients with a paired method.
Sensitivity & Specificity (Diagnostic Accuracy)
MedCalc diagnostic accuracy results at the chosen threshold include sensitivity, specificity and likelihood ratios with intervals, and predictive values restated for the prevalence of your setting.
Inter-Rater Reliability (Cohen's Kappa & ICC)
Agreement between raters uses kappa for categories and the intraclass correlation for measurements, and agreement between two measurement methods is shown with Bland-Altman limits.
Relative Risk & Odds Ratios
Two-by-two tables give risk differences, relative risks and odds ratios with confidence intervals, and the measure is matched to a cohort, trial or case-control design.
Logistic Regression
A binary outcome is modeled with adjusted odds ratios, and a combined marker from the model can be evaluated with its own ROC curve, ideally in data not used to fit it.
Kaplan-Meier Analysis
Survival curves keep censored observations and are compared with the log-rank test, with Cox regression added for adjusted hazard ratios where the design supports them.
COMMON MEDCALC ISSUES
Where Diagnostic Analyses Go Wrong Before the Estimate
The most frequent errors in diagnostic projects are conceptual and survive any software, so each check below is made before an accuracy figure is reported.
Prevalence Set Before Predictive Values MedCalc reports predictive values only when a prevalence is set, and one option takes it from the ratio of cases in the sample. That is wrong for a case-control design, so the prevalence of the intended setting is entered and named in the table.
Passing-Bablok Applied Within Its Limits Passing-Bablok regression assumes a linear relation between two closely related methods. Its cusum test only checks that assumption and says nothing about agreement, so the regression is read beside the Bland-Altman results.
Cutoff Chosen From the Same Data MedCalc's ROC output names the criterion with the highest Youden index, which invites using it as the cutoff. Chosen and evaluated in one sample, that point overstates accuracy, so the threshold rule is fixed before the output is read or the cutoff is labeled exploratory.
Correlation Is Not Agreement Two methods can correlate strongly while one reads consistently higher. Paired readings get limits of agreement judged against a clinically acceptable difference, and regression methods for proportional bias.
Several Readings per Patient Multiple lesions or repeated readings from one patient are not independent. A clustered analysis or one reading per patient is used, because the intervals are otherwise too narrow.
Small Groups and Wide Intervals Accuracy in a small subgroup is imprecise, so each estimate is reported with its interval and the number of cases and non-cases behind it, instead of a bare percentage.
DELIVERABLES
What You Receive From a MedCalc Analysis
The files follow a documented diagnostic workflow, agreed in the scope and provided where applicable, so that every accuracy figure can be traced to its source.
Dataset and Output The data sheet and the MedCalc output for each analysis, kept together so that every reported estimate traces to its source.
Threshold Rule A written rule for how the cutoff was chosen, whether it was prespecified, and the sensitivity and specificity it yields.
Accuracy Estimates The area under the curve, sensitivity, specificity and likelihood ratios with 95% confidence intervals, plus predictive values at the stated prevalence.
Agreement Results For a method comparison, the bias, the limits of agreement and the plot, or the regression-based results where proportional bias is present.
Tables and Figures Publication-ready tables, ROC curves and agreement plots in Word or Excel, organized to the STARD items where the study is a diagnostic accuracy study.
Statistical Interpretation Each estimate is explained within what the design supports, with its limitations stated plainly.
OTHER PLATFORMS
When Another Platform May Be More Appropriate
MedCalc suits many diagnostic and agreement studies, but the software follows the design, and some projects need another platform.
ILLUSTRATIVE EXAMPLE
How a MedCalc Analysis May Look in Practice
An illustration of how a device comparison moves from a correlation to a defensible agreement and threshold analysis.
ILLUSTRATIVE EXAMPLE · NEONATAL BILIRUBIN SCREENING RESEARCH · UNITED STATES
A hospital newborn unit wanted to know whether a transcutaneous bilirubin meter could replace some blood draws, by comparing it with serum measurement and by testing a referral threshold.
The unit had first reported a high correlation between the two methods and considered the question settled. The analysis was reframed as agreement: bias and limits of agreement were estimated and judged against a difference the clinicians considered acceptable, and a check for proportional bias was added. Some infants contributed several paired readings, so those readings were handled as clustered, not as independent pairs. The referral threshold was prespecified from clinical guidance and then evaluated for sensitivity and specificity, while predictive values were restated for the prevalence of the unit’s own population, which differed from that of the study sample.
Outcome: The unit received the agreement and accuracy estimates with their 95% confidence intervals, the written threshold rule, the data and MedCalc output files, and tables and plots organized to the STARD items.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
RESEARCH DISCIPLINES
Research Fields Where MedCalc Is Common
MedCalc is most often used in clinical and laboratory medicine, and the same accuracy and agreement questions arise in any field that evaluates a test, a score or a measuring device.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCHER VOICE
What Researchers Say
Feedback from a clinical researcher on predictive values, prevalence and a reproducible workflow built from the files already in hand.
“Our issue with MedCalc was not the interface itself; it was confusing sensitivity and specificity with predictive values when prevalence changed. They worked directly with the files we already had and left us with a much cleaner, reproducible workflow rather than just screenshots of output.”
Statistical Support Within Your Research OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, including the MedCalc files described in the scope. The researcher or research team retains the scientific question, substantive interpretation, conclusions and overall responsibility for the research.
FREQUENTLY ASKED QUESTIONS
Questions About MedCalc Analysis for Research
Short answers to what researchers ask before sharing a MedCalc project, from the analyses we carry out and existing output to agreement, platform choice and confidentiality.
Which analyses do you carry out in MedCalc?
MedCalc analysis at OnlineSPSS covers diagnostic accuracy and ROC analysis, method comparison and agreement, reliability, risk measures from two-by-two tables, logistic regression and survival analysis. The choice follows your study design and reference standard, and it is confirmed during the assessment.
Can you check MedCalc results we have already produced?
Yes, from the data file and the procedure settings, not from the printed report alone. The coding of the classification variable, the prevalence entered for predictive values and the criterion reported as the cutoff are read first, and a completed analysis is checked in full through Independent Statistical Review.
Can you compare two measurement methods in MedCalc?
Yes, with paired measurements on the same samples. The comparison reports bias and limits of agreement, and uses Passing-Bablok or Deming regression where the difference changes across the range. A correlation coefficient alone is not accepted as evidence of agreement.
Is MedCalc enough for my diagnostic accuracy study, or do I need R or Stata?
MedCalc is enough for a single test against a reference standard, for comparing ROC curves measured on the same patients and for method comparison. R or Stata is needed when accuracy must be modeled with several readings per patient, adjusted for covariates or pooled across studies with a bivariate model, or when the study’s other analyses already run there.
How do you protect my research data and confidentiality?
Your datasets and MedCalc output are handled confidentially, and project information is not published. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the dataset and any protocol, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Discuss Your
MedCalc Analysis With Us
Share your research question, your data, the reference standard and any output you already have. We check the design and confirm the MedCalc analysis before any work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014