Supplementary material: Healthcare resource utilization among nursing home residents with Parkinson’s disease psychosis: an analysis of Medicare beneficiaries treated with pimavanserin or other-atypical antipsychotics

These are peer-reviewed supplementary materials for the article 'Healthcare resource utilization among nursing home residents with Parkinson’s disease psychosis: an analysis of Medicare beneficiaries treated with pimavanserin or other-atypical antipsychotics' published in the Journal of Comparative Effectiveness Research.Supplementary Table 1: Diagnostic Code List Used in Patient SelectionSupplementary Table 2: STROBE Statement—Checklist of Items that Should be Included in Reports of Observational StudiesAim: Real-world healthcare resource use (HCRU) burden among patients with Parkinson’s disease psychosis (PDP) treated with pimavanserin (PIM) versus other atypical antipsychotics (other-AAPs) including quetiapine (QUE) in long term care (LTC) and nursing home (NH) settings are lacking. This analysis examines HCRU differences among residents in LTC/NH settings who initiate PIM versus QUE or other-AAPs. Methods: A retrospective analysis of LTC/NH residents with PDP from the 100% Medicare claims between 1 April 2015 and 31 December 2021 was conducted. Treatment-naive residents who initiated ≥6 months continuous monotherapy with PIM or QUE or other-AAPs between 04/01/16 and 06/30/2021 were propensity score matched (PSM) 1:1 using 31 variables (age, sex, race, region and 27 Elixhauser comorbidity characteristics). Post-index (i.e., 6 months) HCRU outcomes included: proportion of residents with ≥1 all-cause inpatient (IP) hospitalizations and emergency room (ER) visits. HCRU differences were assessed via log binomial regression and reported as relative risk ratios (RR) and 95% confidence intervals after controlling for dementia, insomnia and index year. Results: From a total of PIM (n = 1827), QUE (n = 7770) or other-AAPs (n = 9557), 1:1 matched sample (n = 1827) in each cohort were selected. All-cause IP hospitalizations (PIM [29.8%]) versus QUE [36.7%]) and ER visits (PIM [47.3%] versus QUE [55.8%]), respectively, were significantly lower for PIM. PIM versus QUE cohort also had significantly lower RR for all-cause IP hospitalizations and ER visits, respectively, (IP hospitalizations RR: 0.82 [0.75. 0.9]; ER visits RR: 0.85 [0.8. 0.9]). PIM versus other-AAPs also had lower likelihood of HCRU outcomes. Conclusion: In this analysis, LTC/NH residents on PIM monotherapy (versus QUE) had a lower likelihood of all-cause hospitalizations (18%) and ER (15%) visits. In this setting, PIM also had lower likelihood of all-cause HCRU versus other-AAPs.

本数据集为《比较疗效研究杂志》上发表的论文《患有帕金森病精神症状的养老院居民医疗资源利用：使用匹莫van森或其它非典型抗精神病药物治疗的医疗保险受益者分析》的同行评审补充材料。补充表1：用于患者选择的诊断代码清单；补充表2：STROBE声明——应包括在观察性研究报告中的项目清单。研究目的：分析在长期护理（LTC）和养老院（NH）环境中接受匹莫van森（PIM）与其它非典型抗精神病药物（其它-AAPs）包括喹硫平（QUE）治疗的帕金森病精神症状（PDP）患者群体在现实世界中的医疗资源利用（HCRU）负担。目前关于PDP患者接受PIM或其它-AAPs治疗的HCRU差异研究尚不充分。本研究旨在考察LTC/NH环境中启动PIM治疗与启动QUE或其它-AAPs治疗的患者之间的HCRU差异。研究方法：对2015年4月1日至2021年12月31日的100%医疗保险索赔数据进行回顾性分析，选取了2016年4月1日至2021年6月30日期间启动≥6个月持续单药治疗的PIM、QUE或其它-AAPs的无治疗史患者，采用31个变量（年龄、性别、种族、地区及27项Elixhauser合并症特征）进行倾向评分匹配（PSM）1:1。观察期后（即6个月）的HCRU结局包括：≥1次全因住院（IP）和急诊室（ER）就诊的患者比例。通过对痴呆、失眠和基准年进行控制，采用对数二元回归方法评估HCRU差异，并以相对风险比（RR）和95%置信区间进行报告。研究结果：在总计匹莫van森（n = 1827）、喹硫平（n = 7770）或其它-AAPs（n = 9557）的患者中，每个队列均选择了1:1匹配样本（n = 1827）。PIM组与QUE组相比，全因IP住院率（PIM [29.8%]）和ER就诊率（PIM [47.3%]）显著较低。PIM组与QUE组的全因IP住院率和ER就诊率的相对风险比也显著较低（IP住院率RR：0.82 [0.75-0.9]；ER就诊率RR：0.85 [0.8-0.9]）。PIM组与其它-AAPs组的HCRU结局发生可能性也较低。研究结论：在本分析中，接受PIM单药治疗的LTC/NH居民与接受QUE治疗的居民相比，全因住院（18%）和ER（15%）就诊的可能性较低。在此环境下，PIM相对于其它-AAPs具有较低的全因HCRU可能性。