Mazumder Shrabanti, Khan Md Tareq Ferdous, Bhuiyan Mohammad Alfrad Nobel, Kiser Humayun
Independent Researcher, Cincinnati, OH, USA.
Division of Biostatistics and Bioinformatics, University of Cincinnati, Cincinnati, OH, USA; Department of Statistics, Jahangirnagar University, Dhaka, Bangladesh.
Addict Behav. 2020 Oct;109:106478. doi: 10.1016/j.addbeh.2020.106478. Epub 2020 May 19.
The purposes of the study include (i) demonstrating the US national level historical trends of the number of admitted patients due to substance abuse and those reported the selected substances at the time of admission, and more importantly, (ii) identifying the significant covariates in the association of using each of the substances along with the dynamics of likelihood over the different levels of the covariates. The trend of total admitted patients shows an increasing pattern from 1992 to 2008 and later exhibits a decreasing pattern before experiencing a significant upturn again in the last two consecutive years. During the study period, the highest growth rate of around 1088% is evident for methamphetamine followed by heroin (192%) and marijuana or hashish (45%), while both cocaine or crack (-33%) and alcohol (-29%) show negative growth rates. The estimated logistic regression models show that every covariate, including age, education, employment, gender, living status, race, and ethnicity, has a significant effect on the status of using each of the five selected substances. In parallel, the dynamics of likelihood over the levels of each covariate on every substance unearth even more information. In conclusion, the findings on trend analysis suggest the immediate attention to the growth in admissions for substance abuse treatment, and in response to taking appropriate policy measures, the likelihood dynamics revealed for every substance would undoubtedly play a vital role in identifying the target group as per priority.
（i）展示美国全国范围内因药物滥用而入院的患者数量以及入院时报告使用所选药物的患者数量的历史趋势，更重要的是，（ii）确定使用每种药物与不同协变量水平下可能性动态之间关联中的显著协变量。入院患者总数的趋势在1992年至2008年呈上升模式，随后呈下降模式，直到最近连续两年再次显著上升。在研究期间，甲基苯丙胺的最高增长率约为1088%，其次是海洛因（192%）和大麻或哈希什（45%），而可卡因或快克（-33%）和酒精（-29%）均呈现负增长率。估计的逻辑回归模型表明，每个协变量，包括年龄、教育程度、就业情况、性别、生活状况、种族和族裔，对使用所选五种药物中的每一种的状况都有显著影响。同时，每种药物在每个协变量水平上的可能性动态揭示了更多信息。总之，趋势分析的结果表明应立即关注药物滥用治疗入院人数的增长，并且为了采取适当的政策措施，每种药物所揭示的可能性动态无疑将在按优先级确定目标群体方面发挥至关重要的作用。
