Data mining and association rule mining of risk factors for hyperuricemia

doi:10.3877/cma.j.issn.1674-0785.2020.10.017

Abstract

Abstract:

Objective

To analyze the association and correlation strength of risk factors for hyperuricemia using data mining and association rule mining.

Methods

The clinical data of 3724 patients and community residents were taken as the data set. The Apriori algorithm of IBM SPSS Statistic19.0 was used to analyze the data set.

Results

Approximately 13.32% of the patients had hyperuricemia. Abnormal triglycerides, creatinine, urea nitrogen, and total bilirubin occurred in 25.76%, 7.90%, 13.91%, and 13.56% of the patients, respectively. The prevalence of hyperuricemia in patients with abnormal triglycerides, creatinine, urea nitrogen, and total bilirubin was 20.24%, 43.70%, 22.28%, and 20.72%, respectively. Association rule analysis showed that hypertension (χ²=34.009, P＜0.001), diabetes (χ²=9.075, P=0.003), triglycerides (χ²=20.388, P＜0.001), total bilirubin (χ²=57.032, P＜0.001), ALT (χ²=6.156, P=0.017), creatinine (χ²=102.707, P＜0.001), and urea nitrogen (χ²=72.818, P＜0.001) were significantly different, although there were no significant differences in blood glucose (χ²=0.369, P=0.584), total cholesterol (χ²=1.081, P=0.326), and AST (χ²=3.656, P=0.074).

Conclusion

Hypertension, Cr, TG, BUN, and total bilirubin are risk factors for hyperuricemia. Hyperuricemia is closely related to hypertension and is an independent risk factor and predictive factor for hypertension.

Key words: Association rule, Hyperuricemia, Data mining, Prevalence

Rurong Sun, Ting Xu, Wen Xie, Yingchun Ma, Peirong Zhang, Shaobo Yu, Zheng Xu, Lu Zhang, Shiliang Zhou, Wanlan Jiang, Junyan Tan, Min Wu. Data mining and association rule mining of risk factors for hyperuricemia[J]. Chinese Journal of Clinicians(Electronic Edition), 2020, 14(10): 835-839.

/ / Recommend

Add to citation manager EndNote|Ris|BibTeX

URL: https://zhlcyszz.cma-cmc.com.cn/EN/10.3877/cma.j.issn.1674-0785.2020.10.017

https://zhlcyszz.cma-cmc.com.cn/EN/Y2020/V14/I10/835

Figures/Tables 4

References 14

1	王华, 胡学钢. 医学数据挖掘中的数据预处理与Apriori算法改进 [J]. 计算机系统应用, 2009, 18(9): 94-97.
2	Boytcheva S, Angelova G, Angelov Z, et al. Text Mining and Big Data Analytics for Retrospective Analysis of Clinical Texts from Outpatient Care [J]. Cybernetics & Information Technologies, 2015, 15(4): 58-77.
3	Roughley MJ, Belcher J, Mallen CD, et al. Gout and risk of chronic kidney disease and nephrolithiasis: metaanalysis of observational studies [J]. Arthritis Res Ther, 2015, 1(17): 90.
4	Liu WC, Hung CC, Chen SC, et al. Association of hyperuricemia with renal outcomes, cardiovascular disease, and mortality [J]. Clin J Am Soc Nephrol, 2012, 7(4): 541-548.
5	Wei CY, Sun CC, Wei JC,et al. Association between hyperuricemia and metabolic syndrome: an epidemiological study of a labor force population in Taiwan [J]. Biomed Res Int, 2015, 2015: 369179.
6	Witten I H, Frank E. Data mining: practical machine learning tools and techniques [J]. Acm Sigmod Record, 2005, 31(1): 76-77.
7	Dhanaseelan R, Jeya SM. Diagnosis of coronary artery disease using an efficient hash table based closed frequent item sets mining [J]. Med Biol Eng Comput, 2018, 56(5): 749-759.
8	Pattanaprateep O, Mcevoy M, Attia J, et al. Evaluation of rational nonsteroidal anti-inflammatory drugs and gastro-protective agents use; association rule data mining using outpatient prescription patterns [J]. BMC Med Inform Decis Mak, 2017, 17(1): 96.
9	Fahrudin TM, Syarif I, Barakbah AR. Discovering patterns of NED-breast cancer based on association rules using apriori and FP-growth[C]//International Electronics Symposium on Knowledge Creation and Intelligent Computing, 2017: 132-139.
10	Kim S, An YJ, Lee KH, et al. Using association analysis to find diseases related to childhood obesity [C]// Ninth International Conference on Ubiquitous and Future Networks. IEEE, 2017: 847-850.
11	韩楠, 乔少杰, 宫兴伟, 等. 面向正负关联规则的方剂配伍规律挖掘算法 [J]. 小型微型计算机系统, 2017, 38(7): 1538-1542.
12	郜峦, 王键, 李锋刚, 等. 基于关联规则的肺系疾病症药关系研究 [J]. 中医杂志, 2013, 54(8): 697-700.
13	邢泽欣. 数据挖掘在方剂功效分析中的应用[D]. 石家庄: 河北地质大学, 2016.
14	杨余垒. 改进的关联规则算法在慢性病数据挖掘中的研究[D]. 杭州: 浙江理工大学, 2017.

变量	赋值
尿酸值是否高于420 μmol/L	是=1，否=0
是否高血压	是=1，否=0
是否糖尿病	是=1，否=0
血糖（mmol/L）	≥6为1，＜6为0
总胆固醇（mmol/L）	≥5.2为1，＜5.2为0
三酰甘油（mmol/L）	≥1.88为1，＜1.88为0
总胆红素（μmol/L）	≥17.1为1，＜17.1为0
谷丙草转氨酶（U/L）	≥35为1，＜35为0
谷草转氨酶（U/L）	≥35为1，＜35为0
肌酐（μmol/L）	≥97.5为1，＜97.5为0
尿素氮（mmol/L）	≥7.14为1，＜7.14为0

变量	赋值
尿酸值是否高于420 μmol/L	是=1，否=0
是否高血压	是=1，否=0
是否糖尿病	是=1，否=0
血糖（mmol/L）	≥6为1，＜6为0
总胆固醇（mmol/L）	≥5.2为1，＜5.2为0
三酰甘油（mmol/L）	≥1.88为1，＜1.88为0
总胆红素（μmol/L）	≥17.1为1，＜17.1为0
谷丙草转氨酶（U/L）	≥35为1，＜35为0
谷草转氨酶（U/L）	≥35为1，＜35为0
肌酐（μmol/L）	≥97.5为1，＜97.5为0
尿素氮（mmol/L）	≥7.14为1，＜7.14为0

后项	前项	支持度＞5%	信赖度＞20%	规则置信度	增益
高尿酸血症	三酰甘油	25.76	20.24	5.219	1.502
高尿酸血症	三酰甘油	18.82	21.62	4.609	1.598
	高血压	18.82	21.62	4.609	1.598
高尿酸血症	肌酐	7.90	43.70	3.450	3.229
高尿酸血症	肌酐	5.72	47.51	2.742	3.510
	高血压	5.72	47.51	2.742	3.510
高尿酸血症	尿素氮	13.91	22.28	3.118	1.646
高尿酸血症	尿素氮	9.62	25.06	2.410	1.851
	高血压	9.62	25.06	2.410	1.851
高尿酸血症	尿素氮	5.75	21.15	1.215	1.563
	高血糖	5.75	21.15	1.215	1.563
高尿酸血症	总胆红素	13.56	20.72	2.808	1.531
高尿酸血症	总胆红素	8.98	23.15	2.079	1.711
	高血压	8.98	23.15	2.079	1.711

后项	前项	支持度＞5%	信赖度＞20%	规则置信度	增益
高尿酸血症	三酰甘油	25.76	20.24	5.219	1.502
高尿酸血症	三酰甘油	18.82	21.62	4.609	1.598
	高血压	18.82	21.62	4.609	1.598
高尿酸血症	肌酐	7.90	43.70	3.450	3.229
高尿酸血症	肌酐	5.72	47.51	2.742	3.510
	高血压	5.72	47.51	2.742	3.510
高尿酸血症	尿素氮	13.91	22.28	3.118	1.646
高尿酸血症	尿素氮	9.62	25.06	2.410	1.851
	高血压	9.62	25.06	2.410	1.851
高尿酸血症	尿素氮	5.75	21.15	1.215	1.563
	高血糖	5.75	21.15	1.215	1.563
高尿酸血症	总胆红素	13.56	20.72	2.808	1.531
高尿酸血症	总胆红素	8.98	23.15	2.079	1.711
	高血压	8.98	23.15	2.079	1.711

因素		病例组	对照组	P值
高血压	1	465	370	＜0.001
	0	147	242
糖尿病	1	75	113	0.003
	0	537	499
血糖	1	207	197	0.584
	0	405	415
总胆固醇	1	253	271	0.326
	0	359	341
三酰甘油	1	236	162	＜0.001
	0	376	450
总胆红素	1	127	37	＜0.001
	0	485	575
谷丙转氨酶	1	58	35	0.017
	0	554	577
谷草转氨酶	1	40	25	0.074
	0	572	587
肌酐	1	156	29	＜0.001
	0	456	583
尿素氮	1	141	36	＜0.001
	0	471	576

Please choose a citation manager

Content to export