Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts

Antonio Miguel-Cruz, Pedro Antonio Aya-Parra, William Ricardo Rodríguez-Dueñas, Andres Felipe Camelo-Ocampo, Viena Sofia Plata-Guao, Hector H. Correal O., Nidia Patricia Córdoba-Hernández, Angelmiro Nuñez-Cruz, Jefferson S. Sarmiento-Rojas, Daniel Alejandro Quiroga-Torres

Research output: Contribution to journalConference article

1 Citation (Scopus)

Abstract

The purpose of this study was to determine whether the maintenance of medical equipment should be outsourced (or not). For this, we used data mining techniques called decision trees. We (1) collected 2364 maintenance works orders from 62 medical devices installed in a 900-bed hospital; (2) then we randomly selected 90% of the maintenance works orders to train 8 different decision tree schemas (J48 (pruned and unpruned), Naive Bayes tree, random tree, alternating decision tree, logistic model tree, decision stump, REP tree); (3) next, the remaining 10% of the works orders were used to test the decision tree schemas. The relative absolute error was used to evaluate what the tested decision tree schemas had learned; finally (4), we chose the decision tree schema with the lowest relative absolute error. Overall, the decision tree schemas performed well. 62.5% (5/8) of the decision tree schemas had less than 20% relative absolute error. 87.5% (7/8) of the decision tree schemas had more than 90% in the correct classification (whether to outsource maintenance tasks or not). The different tested decision tree schemas showed that the most important variables when making the decision whether to outsource maintenance tasks or not were: medical device, risk class (I, IIA, IIB, III), complexity, obsolescence, maintenance frequency, service time and outsourcing. The best decision tree schema was the logistic model tree (LMT) with 14.6628% relative absolute error and 94.7034% in the correct classification.

Original languageEnglish (US)
Pages (from-to)295-298
Number of pages4
JournalIFMBE Proceedings
Volume68
Issue number3
DOIs
StatePublished - Jan 1 2019
EventWorld Congress on Medical Physics and Biomedical Engineering, WC 2018 - Prague, Czech Republic
Duration: Jun 3 2018Jun 8 2018

All Science Journal Classification (ASJC) codes

  • Bioengineering
  • Biomedical Engineering

Cite this

Miguel-Cruz, Antonio ; Aya-Parra, Pedro Antonio ; Rodríguez-Dueñas, William Ricardo ; Camelo-Ocampo, Andres Felipe ; Plata-Guao, Viena Sofia ; Correal O., Hector H. ; Córdoba-Hernández, Nidia Patricia ; Nuñez-Cruz, Angelmiro ; Sarmiento-Rojas, Jefferson S. ; Quiroga-Torres, Daniel Alejandro. / Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts. In: IFMBE Proceedings. 2019 ; Vol. 68, No. 3. pp. 295-298.
@article{5fbefa8c29454c608a5c7bed0f8f9fdd,
title = "Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts",
abstract = "The purpose of this study was to determine whether the maintenance of medical equipment should be outsourced (or not). For this, we used data mining techniques called decision trees. We (1) collected 2364 maintenance works orders from 62 medical devices installed in a 900-bed hospital; (2) then we randomly selected 90{\%} of the maintenance works orders to train 8 different decision tree schemas (J48 (pruned and unpruned), Naive Bayes tree, random tree, alternating decision tree, logistic model tree, decision stump, REP tree); (3) next, the remaining 10{\%} of the works orders were used to test the decision tree schemas. The relative absolute error was used to evaluate what the tested decision tree schemas had learned; finally (4), we chose the decision tree schema with the lowest relative absolute error. Overall, the decision tree schemas performed well. 62.5{\%} (5/8) of the decision tree schemas had less than 20{\%} relative absolute error. 87.5{\%} (7/8) of the decision tree schemas had more than 90{\%} in the correct classification (whether to outsource maintenance tasks or not). The different tested decision tree schemas showed that the most important variables when making the decision whether to outsource maintenance tasks or not were: medical device, risk class (I, IIA, IIB, III), complexity, obsolescence, maintenance frequency, service time and outsourcing. The best decision tree schema was the logistic model tree (LMT) with 14.6628{\%} relative absolute error and 94.7034{\%} in the correct classification.",
author = "Antonio Miguel-Cruz and Aya-Parra, {Pedro Antonio} and Rodr{\'i}guez-Due{\~n}as, {William Ricardo} and Camelo-Ocampo, {Andres Felipe} and Plata-Guao, {Viena Sofia} and {Correal O.}, {Hector H.} and C{\'o}rdoba-Hern{\'a}ndez, {Nidia Patricia} and Angelmiro Nu{\~n}ez-Cruz and Sarmiento-Rojas, {Jefferson S.} and Quiroga-Torres, {Daniel Alejandro}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-981-10-9023-3_52",
language = "English (US)",
volume = "68",
pages = "295--298",
journal = "IFMBE Proceedings",
issn = "1680-0737",
publisher = "Springer",
number = "3",

}

Miguel-Cruz, A, Aya-Parra, PA, Rodríguez-Dueñas, WR, Camelo-Ocampo, AF, Plata-Guao, VS, Correal O., HH, Córdoba-Hernández, NP, Nuñez-Cruz, A, Sarmiento-Rojas, JS & Quiroga-Torres, DA 2019, 'Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts', IFMBE Proceedings, vol. 68, no. 3, pp. 295-298. https://doi.org/10.1007/978-981-10-9023-3_52

Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts. / Miguel-Cruz, Antonio; Aya-Parra, Pedro Antonio; Rodríguez-Dueñas, William Ricardo; Camelo-Ocampo, Andres Felipe; Plata-Guao, Viena Sofia; Correal O., Hector H.; Córdoba-Hernández, Nidia Patricia; Nuñez-Cruz, Angelmiro; Sarmiento-Rojas, Jefferson S.; Quiroga-Torres, Daniel Alejandro.

In: IFMBE Proceedings, Vol. 68, No. 3, 01.01.2019, p. 295-298.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Using data mining techniques to determine whether to outsource medical equipment maintenance tasks in real contexts

AU - Miguel-Cruz, Antonio

AU - Aya-Parra, Pedro Antonio

AU - Rodríguez-Dueñas, William Ricardo

AU - Camelo-Ocampo, Andres Felipe

AU - Plata-Guao, Viena Sofia

AU - Correal O., Hector H.

AU - Córdoba-Hernández, Nidia Patricia

AU - Nuñez-Cruz, Angelmiro

AU - Sarmiento-Rojas, Jefferson S.

AU - Quiroga-Torres, Daniel Alejandro

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The purpose of this study was to determine whether the maintenance of medical equipment should be outsourced (or not). For this, we used data mining techniques called decision trees. We (1) collected 2364 maintenance works orders from 62 medical devices installed in a 900-bed hospital; (2) then we randomly selected 90% of the maintenance works orders to train 8 different decision tree schemas (J48 (pruned and unpruned), Naive Bayes tree, random tree, alternating decision tree, logistic model tree, decision stump, REP tree); (3) next, the remaining 10% of the works orders were used to test the decision tree schemas. The relative absolute error was used to evaluate what the tested decision tree schemas had learned; finally (4), we chose the decision tree schema with the lowest relative absolute error. Overall, the decision tree schemas performed well. 62.5% (5/8) of the decision tree schemas had less than 20% relative absolute error. 87.5% (7/8) of the decision tree schemas had more than 90% in the correct classification (whether to outsource maintenance tasks or not). The different tested decision tree schemas showed that the most important variables when making the decision whether to outsource maintenance tasks or not were: medical device, risk class (I, IIA, IIB, III), complexity, obsolescence, maintenance frequency, service time and outsourcing. The best decision tree schema was the logistic model tree (LMT) with 14.6628% relative absolute error and 94.7034% in the correct classification.

AB - The purpose of this study was to determine whether the maintenance of medical equipment should be outsourced (or not). For this, we used data mining techniques called decision trees. We (1) collected 2364 maintenance works orders from 62 medical devices installed in a 900-bed hospital; (2) then we randomly selected 90% of the maintenance works orders to train 8 different decision tree schemas (J48 (pruned and unpruned), Naive Bayes tree, random tree, alternating decision tree, logistic model tree, decision stump, REP tree); (3) next, the remaining 10% of the works orders were used to test the decision tree schemas. The relative absolute error was used to evaluate what the tested decision tree schemas had learned; finally (4), we chose the decision tree schema with the lowest relative absolute error. Overall, the decision tree schemas performed well. 62.5% (5/8) of the decision tree schemas had less than 20% relative absolute error. 87.5% (7/8) of the decision tree schemas had more than 90% in the correct classification (whether to outsource maintenance tasks or not). The different tested decision tree schemas showed that the most important variables when making the decision whether to outsource maintenance tasks or not were: medical device, risk class (I, IIA, IIB, III), complexity, obsolescence, maintenance frequency, service time and outsourcing. The best decision tree schema was the logistic model tree (LMT) with 14.6628% relative absolute error and 94.7034% in the correct classification.

UR - http://www.scopus.com/inward/record.url?scp=85048276324&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048276324&partnerID=8YFLogxK

U2 - 10.1007/978-981-10-9023-3_52

DO - 10.1007/978-981-10-9023-3_52

M3 - Conference article

AN - SCOPUS:85048276324

VL - 68

SP - 295

EP - 298

JO - IFMBE Proceedings

JF - IFMBE Proceedings

SN - 1680-0737

IS - 3

ER -