Recursive partitioning

Recursive partitioning is a statistical method for multivariable analysis.cite book |author=Breiman, Leo |title=Classification and Regression Trees |publisher=Chapman & Hall/CRC |location=Boca Raton |year=1984 |pages= |isbn=0-412-04841-8 |oclc= |doi=] . Recursive partitioning creates a decision tree that strives to correctly classify members of the population based on several dichotomous dependent variables.

This article focuses on recursive partioning for medical diagnostic tests,but the technique has far wider applications.See regression trees.

As compared to regression analyses that creates a formula that health care providers can use to calculate the probability that a patient has a disease, recursive partition creates a rule such as 'If a patient has finding x, y, or z they probably have disease q.

A variation is 'Cox linear recursive partitioning'.

Advantages and disadvantages

Compared to other multivariable methods, recursive partitioning has advantages and disadvantages.
*Advantages are:
**Generates clinically more intuitive models that do not require the user to perform calculations.cite journal |author=James KE, White RF, Kraemer HC |title=Repeated split sample validation to assess logistic regression and recursive partitioning: an application to the prediction of cognitive impairment |journal=Statistics in medicine |volume=24 |issue=19 |pages=3019–35 |year=2005 |pmid=16149128 |doi=10.1002/sim.2154]
**Allows varying prioritizing of missclassifications in order to create a decision rule hat has more sensitivity or specificity.cite journal |author=Cook EF, Goldman L |title=Empiric comparison of multivariate analytic techniques: advantages and disadvantages of recursive partitioning analysis |journal=Journal of chronic diseases |volume=37 |issue=9-10 |pages=721–31 |year=1984 |pmid=6501544 |doi=]
**May be more accurate.cite journal |author=Kattan MW, Hess KR, Beck JR |title=Experiments to determine whether recursive partitioning (CART) or an artificial neural network overcomes theoretical limitations of Cox proportional hazards regression |journal=Comput. Biomed. Res. |volume=31 |issue=5 |pages=363–73 |year=1998 |pmid=9790741 |doi=]
*Disadvantages are:
** Does not work well for continuous variablescite journal |author=Lee JW, Um SH, Lee JB, Mun J, Cho H |title=Scoring and staging systems using cox linear regression modeling and recursive partitioning |journal=Methods of information in medicine |volume=45 |issue=1 |pages=37–43 |year=2006 |pmid=16482368 |doi=]
** May overfit data.

Examples

Examples are available of using recursive partitioning in research of diagnostic tests.cite journal |author=Fonarow GC, Adams KF, Abraham WT, Yancy CW, Boscardin WJ |title=Risk stratification for in-hospital mortality in acutely decompensated heart failure: classification and regression tree analysis |journal=JAMA |volume=293 |issue=5 |pages=572–80 |year=2005 |pmid=15687312 |doi=10.1001/jama.293.5.572] cite journal |author=Stiell IG, Wells GA, Vandemheen KL, "et al" |title=The Canadian C-spine rule for radiography in alert and stable trauma patients |journal=JAMA |volume=286 |issue=15 |pages=1841–8 |year=2001 |pmid=11597285 |doi=] cite journal |author=Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM |title=Indications for computed tomography in patients with minor head injury |journal=N. Engl. J. Med. |volume=343 |issue=2 |pages=100–5 |year=2000 |pmid=10891517 |doi=] cite journal |author=Edworthy SM, Zatarain E, McShane DJ, Bloch DA |title=Analysis of the 1982 ARA lupus criteria data set by recursive partitioning methodology: new insights into the relative merit of individual criteria |journal=J. Rheumatol. |volume=15 |issue=10 |pages=1493–8 |year=1988 |pmid=3060613 |doi=] cite journal |author=Stiell IG, Greenberg GH, Wells GA, "et al" |title=Prospective validation of a decision rule for the use of radiography in acute knee injuries |journal=JAMA |volume=275 |issue=8 |pages=611–5 |year=1996 |pmid=8594242 |doi=] cite journal |author=Goldman L, Weinberg M, Weisberg M, "et al" |title=A computer-derived protocol to aid in the diagnosis of emergency room patients with acute chest pain |journal=N. Engl. J. Med. |volume=307 |issue=10 |pages=588–96 |year=1982 |pmid=7110205 |doi=] Goldman used recursive partitioning to prioritize sensitivity in the diagnosis of myocardial infarction among patients with chest pain in the emergency room.

References