In Sook Cho; Eunja Chung

doi:10.4040/jkan.2011.41.3.423

Articles

Page Path: HOME > J Korean Acad Nurs > Volume 41(3); 2011 > Article

Original Article
Predictive Bayesian Network Model Using Electronic Patient Records for Prevention of Hospital-Acquired Pressure Ulcers: In Sook Cho, Eunja Chung; Journal of Korean Academy of Nursing 2011;41(3):423-431.
DOI: https://doi.org/10.4040/jkan.2011.41.3.423
Published online: June 13, 2011

¹Associated Professor, Department of Nursing, Inha University, Incheon, Korea.

²Director, Department of Nursing, Seoul National University Bundang Hospital, Seongnam, Korea.

Address reprint requests to: Cho, In Sook. Department of Nursing, Inha University, Yonghyeon 4-dong, Nam-gu, Incheon 402-751, Korea. Tel: +82-32-860-8201, Fax: +82-32-874-5880, insook.cho@inha.ac.kr

• Received: July 16, 2010 • Accepted: June 7, 2011

21 Views
0 Download
8 Scopus

Full Article

Download PDF

Abstract
NOTES
REFERENCES

Abstract

Purpose
The study was designed to determine the discriminating ability of a Bayesian network (BN) for predicting risk for pressure ulcers.
Methods
Analysis was done using a retrospective cohort, nursing records representing 21,114 hospital days, 3,348 patients at risk for ulcers, admitted to the intensive care unit of a tertiary teaching hospital between January 2004 and January 2007. A BN model and two logistic regression (LR) versions, model-I and -II, were compared, varying the nature, number and quality of input variables. Classification competence and case coverage of the models were tested and compared using a threefold cross validation method.
Results
Average incidence of ulcers was 6.12%. Of the two LR models, model-I demonstrated better indexes of statistical model fits. The BN model had a sensitivity of 81.95%, specificity of 75.63%, positive and negative predictive values of 35.62% and 96.22% respectively. The area under the receiver operating characteristic (AUROC) was 85.01% implying moderate to good overall performance, which was similar to LR model-I. However, regarding case coverage, the BN model was 100% compared to 15.88% of LR.
Conclusion
Discriminating ability of the BN model was found to be acceptable and case coverage proved to be excellent for clinical use.
Keywords: Pressure ulcer; Bayesian prediction; Logistic models; Risk assessment; Data mining

This work was supported by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MOST) (No. 2009-0053032).

REFERENCES

1. Baumgarten M, Margolis DJ, Localio AR, Kagan SH, Lowe RA, Konsian B, et al. Extrinsic risk factors for pressure ulcers early in the hospital stay: A nested case-control study. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences. 2008;63:408–413.Article PubMed PMC
2. Beeckman D, Schoonhoven L, Fletcher J, Furtado K, Gunningberg L, Heyman H, et al. EPUAP classification system for pressure ulcers: European reliability study. Journal of Advanced Nursing. 2007;60:682–691. doi:10.1111/j.1365-2648.2007.04474.x.Article PubMed
3. Bergstrom N, Braden B, Kemp M, Champagne M, Ruby E. Predicting pressure ulcer risk: A multisite study of the predictive validity of the Braden Scale. Nursing Research. 1998;47:261–269. doi:10.1097/00006199-199809000-00005.Article PubMed
4. Brown SJ. The Braden Scale: A review of the research evidence. Orthopaedic Nursing. 2004;23:30–38. doi:10.1097/00006416-200401000-00010.Article PubMed
5. Burnside ES, Rubin DL, Shachter RD. Using a Bayesian network to predict the probability and type of breast cancer represented by microcalcifications on mammography. Studies in Health Technology And Informatics. 2004;107(Pt 1):13–17.PubMed
6. Cho I, Noh M. Braden Scale: Evaluation of clinical usefulness in an intensive care unit. Journal of Advanced Nursing. 2010;66:293–302. doi:10.1111/j.1365-2648.2009.05153.x.Article PubMed
7. Cho I, Yoon HY, Park SI, Lee HS. Availability of nursing data in an electronic nursing record system for a development of a risk assessment. Journal of Korean Society of Medical Informatics. 2008;14:161–168.
8. Cios KJ, Moore GW. Uniqueness of medical data mining. Artificial Intelligence in Medicine. 2002;26:1–24. doi:10.1016/S0933-3657(02)00049-0.Article PubMed
9. Cover TM, Thomas JA. Elements of information theory. 1991;New York, John Wiley & Sons.
10. De Laat EH, Pickkers P, Schoonhoven L, Verbeek AL, Feuth T, van Achterberg T. Guideline implementation results in a decrease of pressure ulcer incidence in critically ill patients. Critical Care Medicine. 2007;35:815–820. doi:10.1097/01.CCM.0000257072.10313.56.Article PubMed
11. Defloor T, Grypdonck MF. Pressure ulcers: Validation of two risk assessment scales. Journal of Clinical Nursing. 2005;14:373–382. doi:10.1111/j.1365-2702.2004.01058.x.Article PubMed
12. Fernandez G. Statistical data mining using SAS applications. 2010;2nd ed. Minnesota, CRC Press.
13. Fischer JE, Bachmann LM, Jaeschke R. A readers' guide to the interpretation of diagnostic test properties: Clinical example of sepsis. Intensive Care Medicine. 2003;29:1043–1051. doi:10.1007/s00134-003-1761-8.Article PubMed PDF
14. Fogerty MD, Abumrad NN, Nanney L, Arbogast PG, Poulose B, Barbul A. Risk factors for pressure ulcers in acute care hospitals. Wound Repair and Regeneration. 2008;16:11–18. doi:10.1111/j.1524-475X.2007.00327.x.Article PubMed
15. Harrison JH. Introduction to the mining of clinical data. Clinics in Laboratory Medicine. 2008;28:1–7. doi:10.1016/j.cll.2007.10.001.Article PubMed
16. Hatanaka N, Yamamoto Y, Ichihara K, Mastuo S, Nakamura Y, Watanabe M, et al. A new predictive indicator for development of pressure ulcers in bedridden patients based on common laboratory tests results. Journal of Clinical Pathology. 2008;61:514–518. doi:10.1136/jcp.2007.050195.Article PubMed
17. Jones KR, Fennie K, Lenihan A. Evidence-based management of chronic wounds. Advances in Skin and Wound Care. 2007;20:591–600. doi:10.1097/01.ASW.0000284936.32707.8d.Article PubMed
18. Korb KB, Nicholson AE. Bayesian artificial intelligence. 2004;London, Chapman & Hall/CRC press.
19. Lee KO, Shin HJ, Park HA, Jeong HM, Lee MH, Choi EH, et al. Patient severity classification in a medical ICU using APACHE III and patient severity classification tool. Journal of Korean Academy of Nursing. 2000;30:1243–1253.Article PDF
20. Lee SM, Abbott P, Johantgen M. Logistic regression and Bayesian networks to study outcomes using large data sets. Nursing Research. 2005;54:133–138. doi:10.1097/00006199-200503000-00009.PubMed
21. Lee YH, Jeong IS, Jeon SS. A comparative study on the predictive validity among presusre ulcer risk assessment scales. Journal of Korean Academy of Nursing. 2003;33:162–169.PubMed
22. Lucas PJF, van der Gaag LC, Abu-Hanna A. Bayesian network in biomedicine and health-care. Artificial Intelligence in Medicine. 2004;30:201–214. doi:10.1016/j.artmed.2003.11.001.Article PubMed
23. Park HS, Park KY, Yu SM. Factors influencing the development of pressure ulcers in surgical patients. Journal of Korean Academy of Nursing. 2005;35:125–134.Article PubMed PDF
24. Pesut D. Weaver C, Delaney C, Weber P, Carr R. 21st century nursing knowledge work: Reasoning into the future. In: Nursing and informatics for the 21st century: Cases, practice, and the future. 2006;Chicago, HIMSS Press. 14–17.
25. Russo CA, Steiner C, Spector W. Rep. No. HCUP Statistical Brief #64. Hospitalizations related to pressure ulcers, 2006. 2008;Rockville, MD, Agency for Healthcare Research and Quality.
26. Schoonhoven L, Haalboom JR, Bousema MT, Algra A, Grobbee DE, Grypdonck MH, et al. Prospective cohort study of routine use of risk assessment scales for prediction of pressure ulcers. British Medical Journal. 2002;325:797–800. doi:10.1136/bmj.325.7368.797.Article PubMed PMC
27. Thede LQ, Sewell JP. Informatics and nursing: Competencies and applications. 2010;3rd ed. Philadelphia, Wolters Kluwer Health | Lippincott Williams & Wilkins.

Figure 1

Procedure for extracting data from the clinical data repository (CDR) of the research hospital. N_pt and Nhospital-day are the numbers of patients and hospital-day, respectively.

Figure 2

A summary of analytic procedures, N_hospital-day is the number of hospital-day, BN stands for Bayesian Network.

Table 1

Group Comparison of Demographic Characteristics between Ulcer Group and Risk Group

Table 2

Multiple Logistic Regression Models Examining Risks of Pressure Ulcers

^*Confidence intervals (95% CI) not containing the null value (1.00) are statistically significant at p<.05; ^†p<.001 for χ² tests; The Akaike information criterion (AIC), a measure of statistical model fit, was used to compare the amount of information explained across the logistic regression models. A lower AIC value indicates a model is a better fit for the observed data.

ER=emergency room; OR=odds ratio; CI=confidence interval.

Table 3

Results of the Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value

PPV=positive predictive value; NPV=negative predictive value.

Table 4

Test Characteristics of the Predictive Models

Table 5

Case Coverage by the Predictive Models

Figure & Data

REFERENCES

Citations

Citations to this article as recorded by

Cite

CITE: export Copy Download Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Predictive Bayesian Network Model Using Electronic Patient Records for Prevention of Hospital-Acquired Pressure Ulcers

J Korean Acad Nurs. 2011;41(3):423-431. Published online June 13, 2011

DOI: https://doi.org/10.4040/jkan.2011.41.3.423

XML Download

Figure

We recommend

Predictive Bayesian Network Model Using Electronic Patient Records for Prevention of Hospital-Acquired Pressure Ulcers

Figure 1 Procedure for extracting data from the clinical data repository (CDR) of the research hospital. Npt and Nhospital-day are the numbers of patients and hospital-day, respectively.

Figure 2 A summary of analytic procedures, Nhospital-day is the number of hospital-day, BN stands for Bayesian Network.

Figure 1

Figure 2

Predictive Bayesian Network Model Using Electronic Patient Records for Prevention of Hospital-Acquired Pressure Ulcers

Table 1

Group Comparison of Demographic Characteristics between Ulcer Group and Risk Group

Table 2

Multiple Logistic Regression Models Examining Risks of Pressure Ulcers

ER=emergency room; OR=odds ratio; CI=confidence interval.

Table 3

Results of the Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value

PPV=positive predictive value; NPV=negative predictive value.

Table 4

Test Characteristics of the Predictive Models

Table 5

Case Coverage by the Predictive Models

Table 1 Group Comparison of Demographic Characteristics between Ulcer Group and Risk Group

Table 2 Multiple Logistic Regression Models Examining Risks of Pressure Ulcers

ER=emergency room; OR=odds ratio; CI=confidence interval.

Table 3 Results of the Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value