Quantile regression model-based analysis of influencing factors on quality of life of inpatients with type 2 diabetes
-
摘要:
目的 了解银川市2型糖尿病((type 2 diabetes mellitus, T2DM)住院患者健康相关生命质量(health-related quality of life, HRQOL)现状, 分析影响T2DM住院患者HRQOL的因素; 探讨分位数回归分析HRQOL影响因素的应用价值。 方法 使用横断面调查的方法获得480例T2DM住院患者的资料, 采用中文版糖尿病相关生命质量(Chinese normal audit of diabetes-dependent quality of Life, CN-ADDQoL)量表评估患者的HRQOL状况, 比较应用传统线性回归和分位数回归模型分析T2DM住院患者HRQOL影响因素的结果。 结果 T2DM住院患者平均加权影响得分(average weight impact, AWI)为-2.7(-3.6, -1.9), 得分最低的条目是“工作生涯”(AWI, -4(-6, -2)), “吃东西随意性”(AWI, -4(-6, -2))和“喝东西随意性”(AWI, -4(-6, -2))。线性回归结果显示, 年龄18~59岁(β=0.465, P=0.001)、有肾脏(β=-0.375, P=0.012)和循环系统并发症(β=-0.287, 0.036)是T2DM住院患者HRQOL的危险因素。分位数回归进一步发现, 生命质量越好, 年龄对其影响越弱(β1=0.931, P1=0.001;β2=0.699, P2=0.001;β3=0.370, P3=0.012;β4=0.313, P4=0.035);农村居民(β5=-0.421, P5 < 0.001)、接受胰岛素治疗(β3=-0.325, P3=0.024)、眼部(β1=-0.546, P1=0.008;β5=-0.352, P5=0.008)、肾脏(β5=-0.358, P5=0.025)及循环系统并发症(β1=-0.803, P1 < 0.001;β5=-0.302, P5=0.011)在个别百分位点上对HRQOL有影响。 结论 年龄、城乡居住地、是否接受胰岛素治疗、合并并发症是T2DM住院患者HRQOL的影响因素, 分位数回归模型可以分析在不同百分位点上影响患者生命质量的因素, 为不同HRQOL患者生命质量的提高提供有针对性的合理建议。 -
关键词:
- T2DM住院患者 /
- 健康相关生命质量 /
- 中文版糖尿病相关生命质量量表 /
- 分位数回归 /
- 影响因素
Abstract:Objective This study aimed to understand the status of health-related quality of life(HRQOL) among the hospitalized patients with type 2 diabetes(T2 DM) in Yinchuan, so as to analyze the factors associated with HRQOL of hospitalized T2 DM patients. The applied value of quantile regression in analysis of HRQOL was explored. Methods A cross-sectional study was conducted to obtain data of 480 hospitalized T2 DM patients. The Chinese Normal Audit of Diabetes-Dependent Quality of Life(CN-ADDQoL) scale was used to assess the HRQOL status of patients. Traditional linear regression and quantile regression were used to analyze the influencing factors of HRQOL in hospitalized T2 DM patients. Results The average weight impact(AWI) of hospitalized T2 DM patients was-2.7(-3.6, -1.9), and the items with lowest score were in the dimension of "work life"(AWI, -4(-6, -2)), "eat" AWI, -4(-6, -2)) and "Anything of Drinking"(AWI, -4(-6, -2)). Linear regression results showed that 18-59 years old or the renal and circulatory complications were risk factors for HRQOL in T2 DM hospitalized patients. Quantile regression further found that the better the quality of life, the weaker the effect on age(β1=0.931, P1=0.001; β2=0.699, P2=0.001; β3=0.370, P3=0.012; β4=0.313, P4=0.035), the rural residents(β5=-0.421, P5 < 0.001), insulin treatment(β3=-0.325; P3=0.024), the ocular(β1=-0.546, P1=0.008; β5=-0.352, P5=0.008), renal(β5=-0.358, P5=0.025) and circulatory complications(β1=-0.803, P1 < 0.001; β5=-0.302, P5=0.011) had effect on HRQOL at different quantiles. Conclusions Age, urban and rural residence, whether receive insulin therapy and complications are the influencing factors of HRQOL in hospitalized T2 DM patients. The quantile regression model can show different factors affecting the quality of life of patients at different quintiles. The results from quantile regression can provide the targeted and reasonable recommendations for improvement of HRQOL of T2 DM patients. -
表 1 T2DM住院患者ADDQoL各条目加权得分情况
Table 1. Weighted scores of ADDQoL entries in hospitalized patients with T2DM
条目 加权得分 P25 P50 P75 休闲活动 -4 -2 -1 工作生涯 -6 -4 -2 短/长途旅程 -4 -2 0 度假 -4 -2 -1 体力活动 -4 -2 -1 家庭生活 -6 -3 0 朋友/社交生活 -3 -1 0 最亲近的个人关系 -4 -2 0 性生活 -4 -2 0 身体外表 -4.0 -2.0 -0.8 自信心 -4.0 -2.5 -1.0 动力 -4 -3 -1 人们对我的反应 -2 0 0 对未来的感觉 -4 -4 -2 经济状况 -4 -4 -2 不愿但不得不依赖他人 -4 -4 -2 生活状况/条件 -4 -2 -1 吃东西随意性 -6 -4 -2 喝东西随意性 -6 -4 -2 注:P25:第25百分位数; P50中位数; P75第75百分位数。 表 2 不同基本特征T2DM住院患者AWI得分
Table 2. AWI scores of T2DM hospitalized patients with different basic characteristics
变量 人数[n(%)] AWI [M(P25, P75)] W/H值 P值 性别a 0.409 0.523 男 268(55.8) -2.8(-3.8, -1.9) 女 212(44.2) -2.7(-3.5, -1.9) 年龄a 3.838 0.050 18~ 240(50.0) -3.0(-4.1, -1.9) ≥60 240(50.0) -2.6(-3.4, -1.9) 文化程度a 0.195 0.659 初中及以下 314(65.4) -2.7(-3.6, -2.0) 高中及以上 166(34.6) -2.6(-3.9, -1.8) 职业b 2.510 0.474 无工作 71(14.8) -2.7(-3.8, -2.0) 农民 89(18.5) -2.7(-3.4, -2.2) 其他 142(29.6) -3.0(-4.1, -1.9) 退休 178(37.1) -2.6(-3.5, -1.9) 城乡a 0.068 0.068 城市 311(64.8) -2.7(-3.6, -1.9) 农村 169(35.2) -2.9(-3.9, -2.1) 现居住地a 0.300 0.584 银川市内 211(44.0) -2.7(-3.6, -2.0) 银川市外 269(56.0) -2.8(-3.8, -1.9) 个人月收入(元)b 3.211 0.201 0~ 181(37.7) -2.7(-3.4, -2.1) 2 000~ 170(35.4) -3.0(-3.8, -2.0) ≥4 000 129(26.9) -2.5(-3.6, -1.8) 注:a采用Wilcoxon rank sum test进行统计学检验; b采用Kruskal-Wallis H test进行统计学检验。 表 3 不同健康相关行为的T2DM住院患者AWI得分
Table 3. AWI scores of T2DM hospitalized patients with different health-related behaviors
变量 人数[n(%)] AWI [M(P25, P75)] H/W值 P值 吸烟a 0.959 0.619 否 322(67.1) -2.7(-3.5, -1.9) 是 106(22.1) -2.9(-3.8, -2.2) 以前吸 52(10.8) -2.7(-4.1, -1.9) 饮酒a 0.431 0.806 否 358(74.6) -2.7(-3.6, -2.0) 是 84(17.5) -2.6(-3.8, -1.9) 已戒酒 38(7.9) -2.8(-4.0, -1.9) 日常体育锻炼a 1.923 0.382 基本不锻炼 111(23.1) -2.9(-3.6, -2.1) 偶尔锻炼 104(21.7) -2.7(-3.7, -2.1) 经常锻炼 265(55.2) -2.6(-3.6, -1.9) 进食高胆固醇或高脂肪食物b 26 582 0.680 否 165(34.4) -2.7(-3.6, -1.9) 是 315(65.6) -2.7(-3.8, -2.0) 进食五谷杂粮b 27 438 0.573 否 272(56.7) -2.8(-3.6, -1.9) 是 208(43.3) -2.6(-3.7, -1.9) 进食高糖食物b 17 264 0.195 否 99(20.6) -2.8(-3.8, -2.2) 是 381(79.4) -2.7(-3.6, -1.9) 注:a采用Kruskal-Wallis test进行统计学检验; b采用Wilcoxon rank sum test进行统计学检验。 表 4 不同临床特征T2DM患者AWI得分情况
Table 4. AWI scores of T2DM hospitalized patients withdifferent clinical features
变量 [n(%)]/[M(P25, P75)] AWI [M(P25, P75)] P值 T2DM患病时间(年)a 9.0(5, 15) - 0.205 糖化血红蛋白水平(%)a 8.9(7.5, 10.4) - 0.305 胰岛素治疗b 0.014 否 210(43.8) -2.6(-3.4, -1.8) 是 270(56.6) -3.0(-3.8, -2.1) 近一年门诊就诊次数a 11(3, 12) - 0.144 近一年T2DM住院次数a 1(1, 1) - 0.052 并发症数目c 0.018 无 76(15.8) -2.5(-3.2, -1.6) 一种 94(19.6) -2.7(-3.6, -1.8) 多种 310(64.6) -2.9(-3.8, -2.1) 眼部病变b 0.005 否 293(61.0) -2.6(-3.5, -1.8) 是 187(39.0) -2.9(-4.1, -2.2) 神经病变b 0.277 否 215(44.8) -2.7(-3.5, -1.9) 是 265(55.2) -2.8(-3.8, -2.1) 糖尿病性肾病b 0.002 否 343(71.5) -2.7(-3.5, -1.8) 是 137(28.5) -3.1(-4.2, -2.2) 皮肤病变b 0.413 否 377(78.5) -2.7(-3.6, -1.9) 是 103(21.5) -2.9(-3.8, -2.0) 循环系统并发症b 0.014 否 209(43.5) -2.6(-3.5, -1.8) 是 271(56.5) -2.9(-3.8, -2.1) 注:a采用Spearman’s rank correlation进行统计学检验; b采用Wilcoxon rank sum test进行统计学检验; c采用Kruskal-Wallis H test进行统计学检验。 表 5 T2DM住院患者线性回归与多分位数回归分析结果比较
Table 5. Comparison of linear regression and multi-quantile regression analysis in hospitalized patients with T2DM
变量 LR QR-10 QR-25 QR-50 QR-75 QR-90 β (95% CI)值 P值 β (95% CI)值 P值 β (95% CI)值 P值 β (95% CI)值 P值 β (95% CI)值 P值 β (95% CI)值 P值 截距 -2.604(-2.868~-2.340) < 0.001 -4.409(-5.867~-3.740) < 0.001 -2.931(-3.544~-2.687) < 0.001 -2.337(-2.626~-1.991) < 0.001 -1.469(-1.840~-1.103) < 0.001 -0.708(-1.004~-0.627) < 0.001 年龄(岁) 0.001 0.001 0.001 0.012 0.035 0.074 18~ 1.000 1.000 1.000 1.000 1.000 1.000 ≥60 0.465(0.203~ 0.727) 0.931(0.521~1.492) 0.699(0.376~1.036) 0.370(0.169~0.689) 0.313(0.056~0.625) 0.148(-0.088~0.309) 居住地 0.112 0.103 0.231 0.296 0.176 < 0.001 农村 1.000 1.000 1.000 1.000 1.000 1.000 城市 -0.214(-0.477~0.050) -0.327(-1.064~0.232) -0.272(-0.608~0.088) -0.162(-0.462~0.138) -0.182(-0.390~0.103) -0.421(-0.515~-0.258) 胰岛素治疗 0.061 0.690 0.129 0.024 0.711 0.059 是 1.000 1.000 1.000 1.000 1.000 1.000 否 -0.254(-0.520~0.012) 0.097(-0.479~0.522) -0.337(-0.670~0.002) -0.325(-0.576~-0.115) -0.052(-0.360~0.138) -0.131(-0.342~0.069) 眼部并发症 0.071 0.008 0.583 0.120 0.130 0.008 有 1.000 1.000 1.000 1.000 1.000 1.000 无 -0.248(-0.516~0.021) -0.546(-0.940~-0.015) -0.127(-0.710~0.140) -0.263(-0.452~0.052) -0.238(-0.430~-0.036) -0.352(-0.527~-0.094) 糖尿病肾病 0.012 0.062 0.055 0.080 0.173 0.025 有 1.000 1.000 1.000 1.000 1.000 1.000 无 -0.375(-0.668~-0.082) -0.702(-1.625~0.051) -0.453(-0.720~-0.089) -0.333(-0.511~-0.001) -0.218(-0.461~0.008) -0.358(-0.594~-0.131) 循环系统并发症 0.036 < 0.001 0.172 0.117 0.233 0.011 有 1.000 1.000 1.000 1.000 1.000 1.000 无 -0.287(-0.555~-0.019) -0.803(-1.189~-0.108) -0.313(-0.797~0.108) -0.296(-0.479~0.140) -0.193(-0.411~0.108) -0.302(-0.710~-0.105) 注:LR: linear regression; QR: quantile regression; β: regression coefficient。 -
[1] 杨昆, 刘鸿奇, 贺培凤.糖尿病患者生命质量评价量表国内研究现状[J].世界最新医学信息文摘, 2017, 17(9): 65-66. DOI: 10.3969/j.issn.1671-3141.2017.009.031.Yang K, Liu HQ, He PF. Research status of diabetes patients' quality of life scale in China[J]. World Latest Medicine Information(Electronic Version), 2017, 17(9): 65-66. DOI: 10.3969/j.issn.1671-3141.2017.009.031. [2] 郭月玲, 李春波.分位数回归理论及其应用[J].吉首大学学报(自然科学版), 2014, 35(5): 26-28. DOI: 10.3969/j.issn.1007-2985.2014.05.007.Guo YL, Li CB. Introduction to quantile regression and its application[J]. J Jishou Univ(Natural Science Edition), 2014, 35(5): 26-28. DOI: 10.3969/j.issn.1007-2985.2014.05.007. [3] 丁元林, 倪宗瓒, 张菊英, 等.修订的糖尿病生命质量量表(A-DQOL)信度与效度初探[J].中国慢性病预防与控制, 2000, (4): 160-161, 176. DOI: 10.3969/j.issn.1004-6194.2000.04.007.Ding YL, Ni ZZ, Zhang JY, et al. The assessment on reliability and validity of adjusted diabetes quality of lLife(A-DQOL)scale[J]. Chin J Prev Contr Chron Non-commun Dis, 2000, 8(4): 160-161, 176. DOI: 10.3969/j.issn.1004-6194.2000.04.007. [4] 孔丹莉, 张广恩, 潘海燕, 等.糖尿病特异性生存质量量表的信度与效度初探[J].中国慢性病预防与控制, 2007, 15(3): 202-204. DOI: 10.3969/j.issn.1004-6194.2007.03.004.Kong DL, Zhang GE, Pan HY, et al. The reliability and validity evaluation on the chinese version of the audit of diabetes dependent quality of life scale[J]. Chin J Prev Contr Chron Non-commun Dis, 2007, 15(3): 202-204. DOI: 10.3969/j.issn.1004-6194.2007.03.004. [5] 于文林, 夏萍.糖尿病特异性生存质量量表ADDQoL的研究进展[J].中国慢性病预防与控制, 2018, 26(9): 710-712. DOI: 10.16386/j.cjpccd.issn.1004-6194.2018.09.020.Yu WL, Xia P. Progress in the study of diabetes-specific quality of life scale ADDQoL[J]. Chin J Prev Contr Chron Non-commun Dis, 2018, 26(9): 710-712. DOI: 10.16386/j.cjpccd.issn.1004-6194.2018.09.020. [6] Arditi C, Zanchi A, Peytremann-Bridevaux I. Health status and quality of life in patients with diabetes in Switzerland[J]. Primary care diabetes, 2019, 13(3): 233-241. DOI:https://doi.org/ 10.1016/j.pcd.2018.11.016. [7] Kim D, Kim S, Cho D, et al. Health-related quality of life in patients with type-2 diabetes: Korean audit of diabetes dependent quality-of-life[J]. Value in Health, 2011, 14(3): A100. DOI:https://doi.org/ 10.1016/j.jval.2011.02.557. [8] Bradley C, Speight J. Patient perceptions of diabetes and diabetes therapy: assessing quality of life[J]. Diabetes Metab Res Rev, 2002, 18(S3): S64-S69. DOI: 10.1002/dmrr.279. [9] 胡明, 孙振球.生活质量测评在糖尿病患者疗效评价中的应用[J].中南大学学报(医学版), 2004, 29(1): 99-101. DOI: 10.3321/j.issn:1672-7347.2004.01.029.Hu M, Sun ZQ. Application of quality of life assessment in evaluation of curative effect of diabetes patients[J]. J Cent South Univ(Med Sci), 2004, 29(1): 99-101. DOI: 10.3321/j.issn:1672-7347.2004.01.029. [10] 金岳龙, 丁伶灵, 汪全海, 等.皖江南北地区老年糖尿病患者生命质量及影响因素分析[J].南方医科大学学报, 2014, 34(2): 283-285. DOI: 10.3969/j.issn.1673-4254.2014.02.29.Jin YL, Ding LL, Wang QM, et al. Factors affecting the quality of life of elderly diabetic patients: survey in north and south Wanjian[J]. J South Med Univ, 2014, 34(2): 283-285. DOI: 10.3969/j.issn.1673-4254.2014.02.29. [11] 晋溶辰, 黄金, 杨玲凤.糖尿病患者生活质量的研究现状[J].医学综述, 2010, 16(2): 272-274. DOI: 10.3969/j.issn.1006-2084.2010.02.039.Jin RC, Huang J, Yang LF. Research progress on the quality of life in patients with diabetes mellitus[J]. Medical Recapitulate, 2010, 16(2): 272-274. DOI: 10.3969/j.issn.1006-2084.2010.02.039. [12] 王乐三, 孙振球.糖尿病患者生存质量研究及进展[J].实用预防医学, 2003, 10(1): 120-123. DOI: 10.3969/j.issn.1006-3110.2003.01.071.Wang LS, Sun ZQ. Research and progress on quality of life in diabetic patients[J]. Pract Prev Med, 2003, 10(1): 120-123. DOI: 10.3969/j.issn.1006-3110.2003.01.071. [13] 侯艳, 杨秋萍.糖尿病患者健康相关生命质量研究现状[J].中华临床医师杂志(电子版), 2016, 10(3): 433-436. DOI: 10.3877/cma.j.issn.1674-0785.2016.03.027.Hou Y, Yang QP. Current research status of health-related quality of life in patients with diabetes mellitus[J]. Chin J Clinicians(Electronic Edition), 2016, 10(3): 433-436. DOI: 10.3877/cma.j.issn.1674-0785.2016.03.027. [14] 张妮娅, 唐伟, 刘超.糖尿病患者健康相关生命质量研究进展[J].医学综述, 2009, 15(19): 3002-3004. DOI: 10.3969/j.issn.1006-2084.2009.19.041.Zhang NY, Tang W, Liu C. Health related quality of life in diabetes patient[J]. Medical Recapitulate, 2009, 15(19): 3002-3004. DOI: 10.3969/j.issn.1006-2084.2009.19.041.