● ResFlow · Methodology & references
Methodology & references
Every statistical engine in ResFlow is based on open, traceable standard methods. This page collects the methods and key references behind each tool, so you can cite them directly when writing methods or replying to reviewers, and check that the implementation is on track. Each row links to the matching online tool.
How to use this page
① When writing your methods, find the method and reference for each analysis you used and cite it properly; ② to check a tool's implementation, use each tool's "Export R code / R cross-check" to re-run and compare in R (see User guide · validation & cross-check). Use the original articles for the definitive version and page numbers.
Regression & prediction modelling
| Method | Key reference | Tool |
|---|---|---|
| Logistic regression | Hosmer DW, Lemeshow S. Applied Logistic Regression. Wiley. | Logistic regression → |
| Restricted cubic splines (RCS) | Harrell FE. Regression Modeling Strategies, 2nd ed., Springer 2015; Durrleman S, Simon R. Stat Med 1989. | RCS splines → |
| Nomogram | Iasonos A, et al. J Clin Oncol 2008;26:1364. | Nomogram → |
| Multicollinearity (VIF) | Marquardt DW. Technometrics 1970; standard regression diagnostics. | Project workbench · diagnostics → |
| Multiple linear regression (OLS, continuous outcome) | Standard least squares; Draper NR, Smith H. Applied Regression Analysis, 3rd ed., Wiley 1998. | Project workbench · linear regression → |
| Nested model comparison (likelihood-ratio test / AIC) | Akaike H. IEEE Trans Automat Contr 1974;19:716; likelihood-ratio test for nested models. | Project workbench · model comparison → |
Survival analysis & competing risks
| Method | Key reference | Tool |
|---|---|---|
| Cox proportional-hazards model | Cox DR. JRSS-B 1972;34:187; ties via Breslow NE. Biometrics 1974. | Cox regression → |
| Proportional-hazards (PH) test | Schoenfeld D. Biometrika 1982; Grambsch PM, Therneau TM. Biometrika 1994;81:515. | Project workbench · diagnostics → |
| Harrell C-index (concordance) | Harrell FE, et al. Stat Med 1996;15:361. | Cox regression → |
| Time-dependent ROC / AUC (IPCW) | Uno H, et al. JASA 2007;102:527; Blanche P, et al. Stat Med 2013 (timeROC). | Project workbench · time-dependent AUC → |
| Fine-Gray competing risks (subdistribution hazard) | Fine JP, Gray RJ. JASA 1999;94:496. | Fine-Gray regression → |
| Cumulative incidence function CIF (Aalen-Johansen) | Aalen OO, Johansen S. Scand J Stat 1978;5:141. | Competing risks CIF → |
| Restricted mean survival time (RMST) | Royston P, Parmar MKB. BMC Med Res Methodol 2013;13:152. | RMST → |
| Kaplan-Meier survival estimate | Kaplan EL, Meier P. JASA 1958;53:457. | Project workbench · KM curves → |
| Log-rank test (between-group survival) | Mantel N. Cancer Chemother Rep 1966;50:163; Peto R, Peto J. JRSS-A 1972;135:185. | Project workbench · log-rank → |
Diagnostic tests & discrimination
| Method | Key reference | Tool |
|---|---|---|
| ROC curve & AUC | Hanley JA, McNeil BJ. Radiology 1982;143:29. | ROC curve → |
| Comparing two ROC curves (DeLong test) | DeLong ER, et al. Biometrics 1988;44:837. | DeLong test → |
| Calibration & goodness-of-fit (Hosmer-Lemeshow) | Hosmer DW, Lemeshow S. Commun Stat 1980;9:1043. | Calibration curve → |
| Decision curve analysis (DCA) | Vickers AJ, Elkin EB. Med Decis Making 2006;26:565. | Decision curve → |
| Likelihood ratios & post-test probability (Fagan nomogram) | Fagan TJ. N Engl J Med 1975;293:257. | Fagan nomogram → |
| NRI / IDI (reclassification improvement) | Pencina MJ, et al. Stat Med 2008;27:157. | NRI / IDI → |
Model validation & calibration
| Method | Key reference | Tool |
|---|---|---|
| Bootstrap internal validation (optimism correction) | Harrell FE, et al. Stat Med 1996;15:361; Steyerberg EW, et al. J Clin Epidemiol 2001;54:774; Efron B, Tibshirani R. An Introduction to the Bootstrap, 1993. | Project workbench · internal validation → |
| External validation (discrimination & calibration) | Steyerberg EW, Vergouwe Y. Eur Heart J 2014;35:1925; Collins GS, et al. Ann Intern Med 2015;162:55. | Project workbench · external validation → |
| Calibration slope & calibration curve | Van Calster B, et al. J Clin Epidemiol 2016;74:167; Crowson CS, et al. Stat Methods Med Res 2018;27:1885. | Project workbench · calibration → |
| Events per variable (EPV) | Peduzzi P, et al. J Clin Epidemiol 1996;49:1373 (Logistic); Vittinghoff E, McCulloch CE. Am J Epidemiol 2007;165:710 (Cox). | Project workbench · internal validation → |
Agreement & reliability
| Method | Key reference | Tool |
|---|---|---|
| Bland-Altman agreement | Bland JM, Altman DG. Lancet 1986;1:307. | Bland-Altman → |
| Intraclass correlation coefficient (ICC) | Shrout PE, Fleiss JL. Psychol Bull 1979;86:420. | ICC → |
| Kappa agreement (two / multiple raters) | Cohen J. Educ Psychol Meas 1960; Fleiss JL. Psychol Bull 1971. | Kappa → |
Evidence synthesis & causal
| Method | Key reference | Tool |
|---|---|---|
| Meta-analysis (random effects) | DerSimonian R, Laird N. Control Clin Trials 1986;7:177. | Meta-analysis → |
| Publication bias (Egger test / funnel plot) | Egger M, et al. BMJ 1997;315:629. | Funnel plot → |
| Propensity score (PSM / IPTW / SMD) | Rosenbaum PR, Rubin DB. Biometrika 1983;70:41; Austin PC. Stat Med 2009. | Propensity score → |
| Trend test (Cochran-Armitage) | Cochran WG. Biometrics 1954; Armitage P. Biometrics 1955. | Trend test → |
Reporting standards
| Method | Key reference | Tool |
|---|---|---|
| Prediction-model reporting (TRIPOD) | Collins GS, et al. Ann Intern Med 2015;162:55. | Project workbench · manuscript material → |
| Observational-study reporting (STROBE) | von Elm E, et al. Lancet 2007;370:1453. | — |
| Randomised-trial reporting (CONSORT) | Schulz KF, et al. BMJ 2010;340:c332. | — |
These are the main primary sources for each method, for proper citation and traceability; use the original articles for exact volume/issue/pages and the latest version. The numerical implementations are self-checked (e.g. by reducing to known methods — see User guide · validation & cross-check), but internal performance is optimistic, so key conclusions should be cross-checked against external data and independent implementations such as R before reporting.