Modulation of medical condition likelihood by patient history similarity

Jonathan Turner; Dympna O'Sullivan; Jon Bird

doi:10.3233/SHTI200176

Modulation of medical condition likelihood by patient history similarity

Jonathan Turner, Dympna O'Sullivan, Jon Bird

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Introduction: We describe an analysis that modulates the simple population prevalence derived likelihood of a particular condition occurring in an individual by matching the individual with other individuals with similar clinical histories and determining the prevalence of the condition within the matched group. Methods: We have taken clinical event codes and dates from anonymised longitudinal primary care records for 25,979 patients with 749,053 recorded clinical events. Using a nearest neighbour approach, for each patient, the likelihood of a condition occurring was adjusted from the population prevalence to the prevalence of the condition within those patients with the closest matching clinical history. Results: For conditions investigated, the nearest method performed well in comparison with standard logistic regression. Conclusions: Results indicate that it may be possible to use histories to identify 'similar' patients and thus to modulate future likelihoods of a condition occurring.

Original language	English
Title of host publication	Digital Personalized Health and Medicine - Proceedings of MIE 2020
Editors	Louise B. Pape-Haugaard, Christian Lovis, Inge Cort Madsen, Patrick Weber, Per Hostrup Nielsen, Philip Scott
Publisher	IOS Press
Pages	327-331
Number of pages	5
Volume	16
Edition	270
ISBN (Electronic)	9781643680828
DOIs	https://doi.org/10.3233/SHTI200176
Publication status	Published - 16 Jun 2020
Event	30th Medical Informatics Europe Conference, MIE 2020 - Geneva, Switzerland Duration: 28 Apr 2020 → 1 May 2020

Publication series

Name	Studies in Health Technology and Informatics
Volume	270
ISSN (Print)	0926-9630
ISSN (Electronic)	1879-8365

Conference

Conference	30th Medical Informatics Europe Conference, MIE 2020
Country/Territory	Switzerland
City	Geneva
Period	28/04/20 → 1/05/20

Keywords

Clinical terminologies
EHR
Machine learning

Access to Document

10.3233/SHTI200176Licence: CC BY-NC

https://arrow.tudublin.ie/scschcomart/138Licence: CC BY-NC

Cite this

Turner, J., O'Sullivan, D., & Bird, J. (2020). Modulation of medical condition likelihood by patient history similarity. In L. B. Pape-Haugaard, C. Lovis, I. C. Madsen, P. Weber, P. H. Nielsen, & P. Scott (Eds.), Digital Personalized Health and Medicine - Proceedings of MIE 2020 (270 ed., Vol. 16, pp. 327-331). (Studies in Health Technology and Informatics; Vol. 270). IOS Press. https://doi.org/10.3233/SHTI200176

Turner, Jonathan ; O'Sullivan, Dympna ; Bird, Jon. / Modulation of medical condition likelihood by patient history similarity. Digital Personalized Health and Medicine - Proceedings of MIE 2020. editor / Louise B. Pape-Haugaard ; Christian Lovis ; Inge Cort Madsen ; Patrick Weber ; Per Hostrup Nielsen ; Philip Scott. Vol. 16 270. ed. IOS Press, 2020. pp. 327-331 (Studies in Health Technology and Informatics).

@inproceedings{a33d949a63084f7480bc597c3b3f17e0,

title = "Modulation of medical condition likelihood by patient history similarity",

abstract = "Introduction: We describe an analysis that modulates the simple population prevalence derived likelihood of a particular condition occurring in an individual by matching the individual with other individuals with similar clinical histories and determining the prevalence of the condition within the matched group. Methods: We have taken clinical event codes and dates from anonymised longitudinal primary care records for 25,979 patients with 749,053 recorded clinical events. Using a nearest neighbour approach, for each patient, the likelihood of a condition occurring was adjusted from the population prevalence to the prevalence of the condition within those patients with the closest matching clinical history. Results: For conditions investigated, the nearest method performed well in comparison with standard logistic regression. Conclusions: Results indicate that it may be possible to use histories to identify 'similar' patients and thus to modulate future likelihoods of a condition occurring.",

keywords = "Clinical terminologies, EHR, Machine learning",

author = "Jonathan Turner and Dympna O'Sullivan and Jon Bird",

note = "Publisher Copyright: {\textcopyright} 2020 European Federation for Medical Informatics (EFMI) and IOS Press.; 30th Medical Informatics Europe Conference, MIE 2020 ; Conference date: 28-04-2020 Through 01-05-2020",

year = "2020",

month = jun,

day = "16",

doi = "10.3233/SHTI200176",

language = "English",

volume = "16",

series = "Studies in Health Technology and Informatics",

publisher = "IOS Press",

pages = "327--331",

editor = "Pape-Haugaard, \{Louise B.\} and Christian Lovis and Madsen, \{Inge Cort\} and Patrick Weber and Nielsen, \{Per Hostrup\} and Philip Scott",

booktitle = "Digital Personalized Health and Medicine - Proceedings of MIE 2020",

edition = "270",

}

Turner, J, O'Sullivan, D & Bird, J 2020, Modulation of medical condition likelihood by patient history similarity. in LB Pape-Haugaard, C Lovis, IC Madsen, P Weber, PH Nielsen & P Scott (eds), Digital Personalized Health and Medicine - Proceedings of MIE 2020. 270 edn, vol. 16, Studies in Health Technology and Informatics, vol. 270, IOS Press, pp. 327-331, 30th Medical Informatics Europe Conference, MIE 2020, Geneva, Switzerland, 28/04/20. https://doi.org/10.3233/SHTI200176

Modulation of medical condition likelihood by patient history similarity. / Turner, Jonathan; O'Sullivan, Dympna; Bird, Jon.
Digital Personalized Health and Medicine - Proceedings of MIE 2020. ed. / Louise B. Pape-Haugaard; Christian Lovis; Inge Cort Madsen; Patrick Weber; Per Hostrup Nielsen; Philip Scott. Vol. 16 270. ed. IOS Press, 2020. p. 327-331 (Studies in Health Technology and Informatics; Vol. 270).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Modulation of medical condition likelihood by patient history similarity

AU - Turner, Jonathan

AU - O'Sullivan, Dympna

AU - Bird, Jon

PY - 2020/6/16

Y1 - 2020/6/16

N2 - Introduction: We describe an analysis that modulates the simple population prevalence derived likelihood of a particular condition occurring in an individual by matching the individual with other individuals with similar clinical histories and determining the prevalence of the condition within the matched group. Methods: We have taken clinical event codes and dates from anonymised longitudinal primary care records for 25,979 patients with 749,053 recorded clinical events. Using a nearest neighbour approach, for each patient, the likelihood of a condition occurring was adjusted from the population prevalence to the prevalence of the condition within those patients with the closest matching clinical history. Results: For conditions investigated, the nearest method performed well in comparison with standard logistic regression. Conclusions: Results indicate that it may be possible to use histories to identify 'similar' patients and thus to modulate future likelihoods of a condition occurring.

AB - Introduction: We describe an analysis that modulates the simple population prevalence derived likelihood of a particular condition occurring in an individual by matching the individual with other individuals with similar clinical histories and determining the prevalence of the condition within the matched group. Methods: We have taken clinical event codes and dates from anonymised longitudinal primary care records for 25,979 patients with 749,053 recorded clinical events. Using a nearest neighbour approach, for each patient, the likelihood of a condition occurring was adjusted from the population prevalence to the prevalence of the condition within those patients with the closest matching clinical history. Results: For conditions investigated, the nearest method performed well in comparison with standard logistic regression. Conclusions: Results indicate that it may be possible to use histories to identify 'similar' patients and thus to modulate future likelihoods of a condition occurring.

KW - Clinical terminologies

KW - EHR

KW - Machine learning

U2 - 10.3233/SHTI200176

DO - 10.3233/SHTI200176

M3 - Conference contribution

C2 - 32570400

AN - SCOPUS:85086930323

VL - 16

T3 - Studies in Health Technology and Informatics

SP - 327

EP - 331

BT - Digital Personalized Health and Medicine - Proceedings of MIE 2020

A2 - Pape-Haugaard, Louise B.

A2 - Lovis, Christian

A2 - Madsen, Inge Cort

A2 - Weber, Patrick

A2 - Nielsen, Per Hostrup

A2 - Scott, Philip

PB - IOS Press

T2 - 30th Medical Informatics Europe Conference, MIE 2020

Y2 - 28 April 2020 through 1 May 2020

ER -

Turner J, O'Sullivan D, Bird J. Modulation of medical condition likelihood by patient history similarity. In Pape-Haugaard LB, Lovis C, Madsen IC, Weber P, Nielsen PH, Scott P, editors, Digital Personalized Health and Medicine - Proceedings of MIE 2020. 270 ed. Vol. 16. IOS Press. 2020. p. 327-331. (Studies in Health Technology and Informatics). doi: 10.3233/SHTI200176

Modulation of medical condition likelihood by patient history similarity

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this