目的:我们旨在评估已发表文献中阿片类药物使用障碍(OUD)成本效益模型中发现的关键健康状态的健康状态效用值(HSUVs)。
方法:数据来自6项试验,代表1,777名OUD患者。我们实施了映射算法,以协调来自不同生活质量度量的数据(SF-12版本1和2以及EQ-5D-3L)。我们进行了回归分析,以量化HSUV与以下变量之间的关系:过去30天内使用阿片类药物的天数,注射行为,用药物治疗OUD,艾滋病毒状况,和年龄。次要分析探讨了阿片类药物戒断症状的影响。
结果:海洛因或其他阿片类药物每天额外使用与医疗外阿片类药物相关的HSUV有统计学意义的显着减少(-0.002(95%CI[-0.003,-0.0001])至-0.003(95%CI[-0.005,-0.002]),分别),药物注射与不注射相比(-0.043(95%CI[-0.079,-0.006])),HIV阳性诊断与未诊断相比(-0.074(95%CI[-0.143,-0.005])),和年龄(每年-0.001(95%CI[-0.003,-0.0002]))。与OUD治疗药物相关的参数在控制药物外阿片类药物使用后没有统计学意义(0.0131(95%CI[-0.0479,0.0769]),与之前的研究一致。次要分析表明,戒断症状是HSUV的基本驱动因素,预测为0.817(95%CI[0.768,0.858]),0.705(95%CI[0.607,0.786]),中度和0.367(95%CI[0.180,0.575]),严重,最严重的症状,分别。
结论:我们观察到OUD的HSUVs高于以前的研究,这些研究是在没有患者输入的情况下进行的。
到目前为止,美国阿片类药物使用障碍患者与健康相关的生活质量估计有限,而且重要的是,它们不是从患有这种疾病的人的研究中产生的。这项研究从六项临床试验中提取了数据,这些临床试验提供了1777名阿片类药物使用障碍患者的数据。由美国国立卫生研究院公开提供,产生与健康相关的生活质量的估计。我们的研究发现,与以前的研究相比,健康相关的生活质量估计更高,药物对阿片类药物使用障碍的适度影响以及戒断症状对该结果的强烈影响。阿片类药物使用障碍患者的这些较高的值可能反映了普通人群(其中这些估计是先前产生的)对这种情况的非常负面的看法。然而,这些相对较高的估计也可以反映对疾病的适应或缺乏对依赖背景下相关健康损害的认识。提供阿片类药物使用障碍的药物数据的观察数量很少,导致与健康相关的生活质量相关估计的不确定性很高。但我们的发现也可以反映患者在没有非药物阿片类药物的积极作用的情况下的真实经历,在临床实践中值得更多关注。我们的研究表明,系统地测量戒断症状并在健康经济模型中表示这些症状可能会更准确地表示阿片类药物使用障碍患者与健康相关的生活质量,从而提供干预措施的影响和成本效益。
OBJECTIVE: We aimed to estimate health state utility values (HSUVs) for the key health states found in opioid use disorder (OUD) cost-effectiveness models in the published literature.
METHODS: Data obtained from six trials representing 1,777 individuals with OUD. We implemented mapping algorithms to harmonize data from different measures of quality of life (the SF-12 Versions 1 and 2 and the EQ-5D-3 L). We performed a regression analysis to quantify the relationship between HSUVs and the following variables: days of extra-medical opioid use in the past 30 days, injecting behaviors, treatment with medications for OUD, HIV status, and age. A secondary analysis explored the impact of opioid withdrawal symptoms.
RESULTS: There were statistically significant reductions in HSUVs associated with extra-medical opioid use (-0.002 (95% CI [-0.003,-0.0001]) to -0.003 (95% CI [-0.005,-0.002]) per additional day of heroin or other opiate use, respectively), drug injecting compared to not injecting (-0.043 (95% CI [-0.079,-0.006])), HIV-positive diagnosis compared to no diagnosis (-0.074 (95% CI [-0.143,-0.005])), and age (-0.001 per year (95% CI [-0.003,-0.0002])). Parameters associated with medications for OUD treatment were not statistically significant after controlling for extra-medical opioid use (0.0131 (95% CI [-0.0479,0.0769])), in line with prior studies. The secondary analysis revealed that withdrawal symptoms are a fundamental driver of HSUVs, with predictions of 0.817 (95% CI [0.768, 0.858]), 0.705 (95% CI [0.607, 0.786]), and 0.367 (95% CI [0.180, 0.575]) for moderate, severe, and worst level of symptoms, respectively.
CONCLUSIONS: We observed HSUVs for OUD that were higher than those from previous studies that had been conducted without input from people living with the condition.
Thus far, health-related quality of life estimates for patients with opioid use disorder in the United States are limited, and importantly, they were not generated from studies among people living with the condition. This study extracted data from six clinical trials providing data among 1,777 people with opioid use disorder, made publicly available by the National Institutes of Health, to produce estimates of health-related quality of life. Our study found higher health-related quality of life estimates as compared to previous studies, modest impact of medications for opioid use disorder and strong impact of withdrawal symptoms on this outcome. These higher values among people with opioid use disorder might reflect the very negative perception of this condition among members of the general population (among whom these estimates have been generated previously). However, these relatively high estimates could also reflect an adaptation to the condition or a lack of awareness of associated-health damage in the context of dependence. The low number of observations providing data on medications for opioid use disorder led to high uncertainty around related estimates of health-related quality of life, but our findings could also reflect real experiences by patients in the absence of the positive effects of non-medication opioids, which deserve more attention in clinical practice. Our study suggests that systematically measuring withdrawal symptoms and representing these in health economic models might provide a more accurate representation of health-related quality of life among people with opioid use disorder and therefore of the impact and cost-effectiveness of interventions.