Professional Documents
Culture Documents
Abstract— Acute Kidney Injury (AKI) is common and inaccurate or delayed [2]. These explain why early detection of
harmful disorder in hospitalized patients. It is associated with AKI and risk identification are often failed.
poor outcomes such as a decrease chance of survival, longer
hospital stays and an increase progression of chronic kidney The clinical practice guidelines for defining AKI have been
disease. To diagnosis AKI, the KDIGO clinical practice guideline published in various definitions such as RIFLE, AKIN (Acute
has been published for providing standardized criteria of AKI Kidney Injury Network) and KDIGO (Kidney Disease
definition and the recommendation in medical pathway. Improving Global Outcomes). However, the incidence rate of
Moreover, early detection of AKI in patient at risk can also AKI is still increasing because it is not a single disease. The
improve the outcomes. This paper presents an approach to assist progression of AKI comes from multiple clinical conditions
the doctor in diagnosis and decision making process. First, the that make the diagnostic of this disease become more complex.
risk factors of AKI were identified using data mining approach Furthermore, the implementation of these guidelines is not
based on Decision Tree classification technique. Simple Cart and successful as it should be. This is because of the difference in
J48 were selected as the algorithms for this process. Second, a viewpoints from multidisciplinary stakeholders, barriers in
concept of tool requirements and design named “AKIHelper” is knowledge and a lack of understanding how to apply these
presented. This tool is created based on KDIGO guideline which guidelines in clinical practice [3].
is expected to use for diagnosis and staging severity of AKI.
This paper proposes an approach to identify the risk factors
Keywords— Acute Kidney Injury; AKI; KDIGO; E-Health; of AKI using data mining techniques. The result will be used
Healthcare; Risk Factors; Diagnostic Tool; Data Mining; Decision for providing information in case of monitoring and caring of
Tree; Classification focused patients. Furthermore, the paper presents an idea to
develop tool for staging AKI severity based on KDIGO
I. INTRODUCTION guideline named “AKIHelper” that AKI risk factors retrieved
from data mining process is also planned to be a part of it. This
Acute Kidney Injury (AKI) is a common disorder with a
tool is hoped to systematically identify stage of disease, assist
rapid reduction of kidney function over hours, days or weeks.
the clinical decision making processes, improve the outcome or
The result of the treatment is unsatisfied and the cause is
clinical care practice and reduce hospital mortality rate as well
unclear, this is accounted for more than 10% of all hospitalized
as medical expenses.
patients and more than 2/3 of all Intensive Care Unit (ICU)
admissions [1]. Besides this, AKI leads to a consequence of The rest of the paper is structured as follows. In Section II,
other acute illness, such as Fluid Overload, Electrolyte background of KDIGO guideline and data mining in healthcare
Abnormalities, Impaired Innate Immunity and Progression of are presented. Section III briefly reviews related works of AKI.
Chronic Kidney Disease (CKD). These can also cause the Section IV describes the proposed approach to identify the risk
extension of hospital stay, increasing mortality rate and higher factors of AKI and presents a concept to develop tool based on
cost of care. KDIGO guideline, including evaluation. Section V discusses
the limitation of the approach. Finally, section VI provides
To date, serum creatinine and urine output criteria have
conclusion and plan for future work.
been developed and used to define the development of AKI.
Level rise of serum creatinine compared with the baseline and
duration of oliguria can be used to define the stage of disease. II. BACKGROUND
Early detection of AKI in patients at risk can be helpful in
improving outcomes of the treatment. Unfortunately, the A. KDIGO Guideline for AKI
diagnosis and monitoring are often neglected or missed during KDIGO (Kidney Disease Improving Global Outcomes)
hospital/ICU admission. Moreover, the measurement and was established in 2003 as non-profit global organization
documentation of creatinine level or urine output are either managed by the National Kidney Foundation. The objectives
of KDIGO are to develop and implement evidence-based
Fig. 4. Result of sample patient dataset applied with Simple Cart [<Admission time] [> Admission time]
Estimate Weight
The result from this section will be used for monitoring by Height
Choose Lower Value Male = Height - 100,
patients at risk to increase the outcome and be applied to the Use colAE as
Female = Height - 105
“True baseline” (colAE VS (colBI))
tool for providing information purpose.
B. Developing Diagnostic Tool
To develop diagnostic tool “AKIHelper”, understanding the Obtain Reference Divide Urine Output per
KDIGO guideline is required because it will be used as the Creatinine (colBG) Day (colCK) by Weight
main structure of the tool, for example, definition, severity
staging and recommendations. The approach for developing Compare Creatinine Day Obtain Urine Output
the tool consists of 2 steps, as in (1) AKI Severity Staging 1 – 7, 14, 21, 28 with colBG per Kilogram
Model and (2) Tool Development Approach. Each of step is
presented seperately in the following subsections.
1) AKI Severity Staging Model: The model was created Each Time Determine Status of AKI
based on KDIGO guideline which provide the definition and (0 – 3 Using the Guideline)
the staging of AKI [4] as described in Table 2. and Table 3.
respectively.
Fig. 5. Model for staging AKI severity using KDIGO guideline
TABLE II. DEFINITION OF AKI (NOT GRADED)
Definition Period 2) Tool Development Approach: This approach is begin
Increase in SCr by ı0.3 mg/dl Within 48 hours with tool fuctions design. AKIHelper is expected to have 4
(ı26.5 lmol/l) main functions as follows:
Increase in SCr to ı1.5 times Known or presumed to have a) Patient Data: This function contains a record of
baseline occurred within the prior 7 days
Urine volume <0.5 ml/kg/h For 6 hours
patient which can be filled in by user or import from database.
The patient data such as date admission, creatinine level and
TABLE III. STAGING OF AKI urine output will be used for diagnosis AKI stage.
Stage Serum Creatinine (SCr) Urine Output
b) AKI Severity Staging: This function is based on AKI
1 1.5–1.9 times baseline < 0.5 ml/kg/h for 6–12 hours severity staging model. If any stage of AKI is detected, an
OR alert will be promptly displayed to notice the user.
0.3 mg/dl (26.5 μmol/l)
increase within 48 hours c) AKI Recommendation: This function is for providing
2 2.0–2.9 times baseline < 0.5 ml/kg/h for 12 hours KDIGO information where the user can look up for reading
3 3 3.0 times baseline < 0.3 ml/kg/h for 24 hours guideline or considering appropriate medical advices.
OR OR
Increase in serum creatinine Anuria for 12 hour
d) Risk Factors: The risk factors are obtained from the
to 4.0 mg/dl (353.6 μmol/l) result retrieved from data mining process and will be stored in
OR the tool for risk factors viewing. This part can be updated
Initiation of renal replacement depending on location setting or amount of dataset.
therapy
OR, In patients < 18 years, Use case diagram and user interface of the designed
decrease in eGFR to < 35 system are shown as in Fig. 6 and Fig. 7 respectively. The
ml/min per 1.73 m2 source of patient data come from 2 channels, Medical Staff
fill-in data directly to the system and retrieve from external proportion of negatives that are correctly identified. Accuracy
system for example, Electronic Medical Record (EMR). This value is used to show the precision of the system.
approach will be then used for developing AKIHelper tool.
V. LIMITATIONS
AKIHelper System
Although the approach is proposed for assisting doctor in
various expectations. One limitation of this tool is that it can
<< actor >> only be used in the part of diagnosis process for staging and
Record Data Medical Record
provide guideline information only, it is not able to handle and
System
manage all of the clinical processes especially decision
making, for example, the severity of AKI that AKIHelper can
Stage Severity detect may be missed if there are multiple serum creatinine
of AKI
results within a short time period. In this case, using such tool
is not effective. Doctor’s consideration and judgement on the
Doctor patient treatment will be required.
View
Recommendation In addition, the risk factors may various across the areas
due to source of patient data, location setting, differences in
Medical Staff clinical practice, process of care or residual confounding from
View Risk unmeasured factors. Therefore, using different dataset before
Factors applying data mining techniques would be effective to identify
the risk factors.
VI. CONCLUSION
Fig. 6. Use case diagram of AKIHelper
This paper presents the proposed approach to identify risk
factors using Decision Tree classification of data mining
technique. Simple Cart and J48 were selected as the algorithms
for this process. The findings from the study will be used to
improve monitoring and treatment pathway in patients at risk.
Next, an idea for developing diagnostic tool named AKIhelper
was presented. The model for building tool was created based
on KDIGO guideline for staging the severity of AKI. The
completed AKIhelper is expected to assist the doctor in
decision making processes and improve quality of medical
outcome which may result in a reduction in mortality and
Fig. 7. Example user interface of AKIHelper
medical expenses.
A further study is needed to provide better understanding of
C. Evaluation
the association between undisclosed factors and the increased
To evaluate the proposed approach, 2x2 contingency table risk of AKI and to develop tool to be plugged in with the
is planned to use as depicted in Table 4. electronic medical record system of the hospital or clinic.
TABLE IV. 2 X 2 CONTINGENCY TABLE
ACKNOWLEDGMENT
Actual Result This paper was supported by Dr. Nattachai S., Excellent
Disease Non-Disease Center for Critical Care Nephrology, Department of Medicine,
a b Chulalongkorn University, Thailand and Department of
Positive
True Positive False Positive Critical Care Medicine, University of Pittsburgh School of
Test Result
c d Medicine, USA and Dr. Wirichada P., Mathematical and
Negative
False Negative True Negative Economic Modelling group, Mahidol-Oxford Tropical
Medicine Research Unit, Mahidol University for advices and
With this table, sensitivity, specificity and accuracy value assistance with data collection.
can be measured as equations below.
REFERENCES
Sensitivity = a / (a + c) (1) [1] Dealmeida, D.; Al-Jaghbeer, M.; Abdelhak, M.; Kellum, J., "A Study to
Specificity = d / (b + d) (2) Evaluate the Effectiveness of the Currently Utilized Acute Kidney
Accuracy = (a + d) / (a + b + c + d) (3) Injury (AKI) Alert: A Use Case Example for a Learning Health
System," in System Sciences (HICSS), 2015 48th Hawaii International
Conference on , vol., no., pp.3125-3131, 5-8 Jan. 2015.
Sensitivity value is used to measure the proportion of
[2] K. Kashani and V. Herasevich, "Sniffing out acute kidney injury in the
positives that are correctly identified (rule out the disease). On ICU: do we have the tools?," Curr Opin Crit Care, vol. 19, pp. 531-6,
the contrary, specificity value is used to measures the Dec 2013.
[3] M. James, E. Dixon, D. Roberts, R. Barry, C. Balint, A. Bharwani, et al., Injury (AKI) Alert: A Use Case Example for a Learning Health
"Improving prevention, early recognition and management of acute System," in System Sciences (HICSS), 2015 48th Hawaii International
kidney injury after major surgery: results of a planning meeting with Conference on , vol., no., pp.3125-3131, 5-8 Jan. 2015
multidisciplinary stakeholders," Canadian Journal of Kidney Health and [10] F. P. Wilson, M. Shashaty, J. Testani, I. Aqeel, Y. Borovskiy, S. S.
Disease, vol. 1, p. 20, 2014. Ellenberg, et al., "Automated, electronic alerts for acute kidney injury: a
[4] Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney single-blind, parallel-group, randomised controlled trial," The Lancet,
Injury Work Group. KDIGO Clinical Practice Guideline for Acute vol. 385, pp. 1966-1974.
Kidney Injury. Kidney inter., Suppl. 2012; 2: 1–138. [11] Hussein, A.S.; Omar, W.M.; Xue Li; Ati, M., "Efficient Chronic Disease
[5] B. Milovic and M. Milovic, "Prediction and decision making in Health Diagnosis prediction and recommendation system," in Biomedical
Care using Data Mining," Kuwait Chapter of the Arabian Journal of Engineering and Sciences (IECBES), 2012 IEEE EMBS Conference on ,
Business and Management Review, vol. 1, p. 126, 2012. vol., no., pp.209-214, 17-19 Dec. 2012
[6] Muhamad, H.; Muhamad, A.; Husain, W.; Rashid, N., “Data Mining for [12] Huang, Feixiang; Wang, Shengyong; Chan, Chien-Chung, "Predicting
Medical Systems: A Review,” in Proc. of the International Conference disease by using data mining based on healthcare information system,"
on Advances in Computer and Information Technology - ACIT 2012, in Granular Computing (GrC), 2012 IEEE International Conference on
pp.17-22 , vol., no., pp.191-194, 11-13 Aug. 2012
[7] N. Srisawat, F. E. Sileanu, R. Murugan, R. Bellomod, P. Calzavacca, R. [13] Neeraj et al., International Journal of Advanced Research in Computer
Cartin-Ceba, et al., "Variation in risk and mortality of acute kidney Science and Software Engineering 3(6), June - 2013, pp. 1114-1119
injury in critically ill patients: a multicenter study," Am J Nephrol, vol. [14] Kalmegh, Sushilkumar. "Analysis of WeKA Data Mining Algorithm
41, pp. 81-8, 2015. REPTree, Simple Cart and Random Tree for Classification of Indian
[8] Singbartl, Kai, and John A. Kellum. "AKI in the ICU: definition, News." International Journal of Innovative Science, Engineering &
epidemiology, risk stratification, and outcomes." Kidney international Technology (IJISET) 2.2 (2015): 438-446.
81.9 (2012): 819-825.
[9] Dealmeida, D.; Al-Jaghbeer, M.; Abdelhak, M.; Kellum, J., "A Study to
Evaluate the Effectiveness of the Currently Utilized Acute Kidney