Epidemiology and Population Health Unit, British Columbia Centre for Excellence in HIV/AIDS, Vancouver, British Columbia, Canada.
Vancouver Coastal Health Authority, Vancouver, British Columbia, Canada.
BMJ Open. 2019 Jul 11;9(7):e030530. doi: 10.1136/bmjopen-2019-030530.
Administrative data are increasingly being used for surveillance and monitoring of mental health and substance use disorders (MHSUD) across Canada. However, the validity of the diagnostic codes specific to MHSUD is unknown in emergency departments (EDs). Our objective was to determine the concordance, and individual-level and hospital-level factors associated with concordance, between diagnosis codes assigned in ED and at discharge from hospital for MHSUD-related conditions.
Population-based retrospective cohort study.
EDs and hospitals within Vancouver Coastal Health Authority (VCH), British Columbia, Canada.
16 926 individuals who were admitted into a VCH hospital following an ED visit from 1 April 2009 to 31 March 2017, contributing to 48 116 pairs of ED and hospital discharge diagnoses.
We examined concordance in identifying MHSUD between the primary discharge diagnosis codes based on the International Statistical Classification of Diseases, 9th and 10th Revisions (Canada) assigned in the ED and those assigned in the hospital among all ED visits resulting in a hospital admission. We calculated the percent overall agreement, positive agreement, negative agreement and Cohen's kappa coefficient. We performed multiple regression analyses to identify factors independently associated with discordance.
We found a high level of concordance for broad categories of MH conditions (overall agreement=0.89, positive agreement=0.74 and kappa=0.67), and a fair level of concordance for SUDs (overall agreement=0.89, positive agreement=0.31 and kappa=0.27). SUDs were less likely to be indicated as the primary cause in ED as opposed to in hospital (3.8% vs 11.7%). In multiple regression analyses, ED visits occurring during holidays, weekends and overnight (21:00-8:59 hours) were associated with increased odds of discordance in identifying MH conditions (adjusted OR 1.47, 95% CI 1.11 to 1.93; 1.27, 95% CI 1.16 to 1.40; 1.30, 95% CI 1.19 to 1.42, respectively).
ED data could be used to improve surveillance and monitoring of MHSUD. Future efforts are needed to improve screening for individuals with MHSUD and subsequently connect them to treatment and follow-up care.
行政数据越来越多地被用于加拿大各地的精神健康和物质使用障碍（MHSUD）的监测和监控。然而，在急诊部门（ED）中，针对 MHSUD 的特定诊断代码的有效性尚不清楚。我们的目的是确定 ED 中分配的诊断代码与医院出院时与 MHSUD 相关疾病的诊断代码之间的一致性，以及与一致性相关的个体水平和医院水平因素。
基于人群的回顾性队列研究。
加拿大不列颠哥伦比亚省温哥华沿海卫生局（VCH）的 ED 和医院。
2009 年 4 月 1 日至 2017 年 3 月 31 日期间在 ED 就诊后被收治入 VCH 医院的 16926 人，共产生了 48116 对 ED 和医院出院诊断。
我们检查了 ED 中基于国际疾病分类（第 9 版和第 10 版修订版）的主要出院诊断代码与所有导致住院的 ED 就诊中在医院分配的 MHSUD 诊断代码之间在识别 MHSUD 方面的一致性。我们计算了总体一致性、阳性一致性、阴性一致性和 Cohen's kappa 系数。我们进行了多项回归分析，以确定与不匹配相关的独立因素。
我们发现，对于 MH 疾病的广泛类别（总体一致性=0.89，阳性一致性=0.74，kappa=0.67），一致性很高，对于 SUD 一致性适中（总体一致性=0.89，阳性一致性=0.31，kappa=0.27）。ED 中 SUD 作为主要病因的指示可能性低于医院（3.8%比 11.7%）。在多项回归分析中，节假日、周末和夜间（21:00-8:59 小时）的 ED 就诊与识别 MH 疾病时出现不一致的可能性增加有关（调整后的比值比 1.47，95%置信区间 1.11 至 1.93；1.27，95%置信区间 1.16 至 1.40；1.30，95%置信区间 1.19 至 1.42）。
ED 数据可用于改善 MHSUD 的监测和监控。未来需要努力改善对 MHSUD 患者的筛查，随后将他们与治疗和随访护理联系起来。
