Prescription drug monitoring programs and opioid-related harm

 描述（由申请人提供）：在过去三十年中，与药物使用有关的致命过量增加了近600%，现在是该国伤害死亡的主要原因。这一增长主要是由处方阿片类药物（PO）过量导致的，其死亡人数超过海洛因，可卡因和兴奋剂的总和。处方药监测计划（PDMP），国家级数据库，药房配药员必须报告处方信息时，某些药物被分配，已被先进的工具，以减少PO相关的伤害。要确定PDMP是否能减少PO相关的伤害，必须回答四个关键问题：（1）PDMP对健康结果的影响是什么？虽然一些研究表明，PDMP使用最佳实践是有效地减少“医生购物”，大多数现有的研究没有检查PDMP在减少PO相关的伤害的作用，和少数，目前不一致的结果。(2)PAPs特征的变化会影响结果吗？现有的研究在很大程度上将PSTs的存在视为二元变量，而没有考虑专家推荐的PSTs操作特性的状态变化作为“最佳实践”。(3)谁从PDMP中受益最多？PDMP的好处可能集中在对PO有医疗需求的群体和较富裕地区的居民中。这两个群体更有可能通过其医疗提供者获得PO，并在PO相关伤害的情况下接受循证治疗的转诊。(4)PDMP会产生意想不到的负面结果吗？如果不作为减少阿片类药物相关危害的综合战略的一部分实施，与PDMP相关的处方减少可能导致滥用PO的人过渡到海洛因使用。过渡到海洛因使用可能是一个特别关注的不太富裕的地区，PO相关的伤害率较高，并获得基于证据的治疗PO依赖性较低。这项研究有两个目的：（1）检验实施“最佳实践”PSTs特征与因PO过量（POD）和海洛因过量（HOD）住院率变化之间的关系;（2）检验实施PSTs“最佳实践”特征与POD和HOD之间的关系是否因人群的医疗需求和社会经济特征而不同。为了实现这些目标，将制定一个Pestival特征的类型学，包括报告的药物时间表数量、数据报告频率、向授权用户主动提供数据、用户培训要求、注册和数据访问以及州际数据共享。将从医疗保健成本和利用项目中获得按邮政编码级别进行地理编码的医院住院患者用药过量数据。我们将测试PSTAP特征的变化对美国18个州因POD和HOD住院率的影响，这些州在PSTAP手术当年提供了PSTAP前后的住院数据和异质性，并在1993-2014年期间居住在这些州内的13，512个邮政编码区的人群中进行。

 DESCRIPTION (provided by applicant): Fatal overdoses related to drug use have increased nearly 600% in the past three decades, and are now the country's leading cause of injury death. This rise has been largely driven by prescription opioid (PO) overdoses, which account for more deaths than heroin, cocaine and stimulants combined. Prescription drug monitoring programs (PDMPs), state-level databases to which pharmacy dispensers must report prescription information when certain medications are dispensed, have been advanced as tools to reduce PO-related harm. Four critical questions must be answered to determine whether PDMPs reduce PO-related harm: (1) What is the impact of PDMPs on health outcomes? While some studies suggest that PDMPs using best practices are effective in reducing "doctor shopping", most existing research has not examined the role of PDMPs in reducing PO-related harm, and the few that have, present inconsistent findings. (2) Do variations in PDMP characteristics affect outcomes? Existing research largely treats presence of a PDMP as a binary variable, without considering state variation in PDMP operational characteristics that have been recommended by experts as "best practices". (3) Who benefits the most from PDMPs? The benefit of PDMPs is likely concentrated among groups with a medical need for POs, and residents of more affluent areas. Both of these groups are more likely to access POs through their medical providers, and to receive referrals to evidence- based treatment in the case of PO-related harm. (4) Can PDMPs have unintended negative outcomes? If not implemented as part of an integrated strategy to reduce opioid-related harm, the reduction in prescriptions associated with PDMPs could potentially lead to transition to heroin use among those who abuse POs. Transitions to heroin use may be a particular concern in less affluent areas, where rates of PO-related harm are higher, and access to evidence-based treatment for PO dependence is lower. This study has two aims: (1) to test the relation between implementation of"best practice" PDMP features and change in the rate of hospitalizations due to PO overdose (POD) and heroin overdose (HOD); and (2) to test whether the relationship between implementation of PDMP "best practice" characteristics and POD and HOD differed by medical need and socioeconomic characteristics of population groups. To address these aims, a typology of PDMP characteristics will be developed, including number of drug schedules reported, frequency of data reporting, proactive provision of data to authorized users, requirements for user training, registration, and data accessing, and interstate data sharing. Hospital inpatient overdose data geocoded to the zip code level will be obtained from the Healthcare Cost and Utilization Project. We will test the impact of variations in PDMP characteristics on rates of hospitalizations due to POD and HOD across 18 U.S. states with available pre- and post-PDMP hospitalization data and heterogeneity in the year of PDMP operation, and among persons living in 13,512 zip code areas within those states across 1993-2014.