PDF(Pubmed)
Abstract：
BACKGROUND: The PRIME-NL study prospectively evaluates a new integrated and personalized care model for people with parkinsonism, including Parkinson\'s disease, in a selected region (PRIME) in the Netherlands. We address the generalizability and sources of selection and confounding bias of the PRIME-NL study by examining baseline and 1-year compliance data.
METHODS: First, we assessed regional baseline differences between the PRIME and the usual care (UC) region using healthcare claims data of almost all people with Parkinson\'s disease in the Netherlands (the source population). Second, we compared our questionnaire sample to the source population to determine generalizability. Third, we investigated sources of bias by comparing the PRIME and UC questionnaire sample on baseline characteristics and 1-year compliance.
RESULTS: Baseline characteristics were similar in the PRIME (n = 1430) and UC (n = 26,250) source populations. The combined questionnaire sample (n = 920) was somewhat younger and had a slightly longer disease duration than the combined source population. Compared to the questionnaire sample in the PRIME region, the UC questionnaire sample was slightly younger, had better cognition, had a longer disease duration, had a higher educational attainment and consumed more alcohol. 1-year compliance of the questionnaire sample was higher in the UC region (96%) than in the PRIME region (92%).
CONCLUSIONS: The generalizability of the PRIME-NL study seems to be good, yet we found evidence of some selection bias. This selection bias necessitates the use of advanced statistical methods for the final evaluation of PRIME-NL, such as inverse probability weighting or propensity score matching. The PRIME-NL study provides a unique window into the validity of a large-scale care evaluation for people with a chronic disease, in this case parkinsonism.
METHODS: First, we assessed regional baseline differences between the PRIME and the usual care (UC) region using healthcare claims data of almost all people with Parkinson\'s disease in the Netherlands (the source population). Second, we compared our questionnaire sample to the source population to determine generalizability. Third, we investigated sources of bias by comparing the PRIME and UC questionnaire sample on baseline characteristics and 1-year compliance.
RESULTS: Baseline characteristics were similar in the PRIME (n = 1430) and UC (n = 26,250) source populations. The combined questionnaire sample (n = 920) was somewhat younger and had a slightly longer disease duration than the combined source population. Compared to the questionnaire sample in the PRIME region, the UC questionnaire sample was slightly younger, had better cognition, had a longer disease duration, had a higher educational attainment and consumed more alcohol. 1-year compliance of the questionnaire sample was higher in the UC region (96%) than in the PRIME region (92%).
CONCLUSIONS: The generalizability of the PRIME-NL study seems to be good, yet we found evidence of some selection bias. This selection bias necessitates the use of advanced statistical methods for the final evaluation of PRIME-NL, such as inverse probability weighting or propensity score matching. The PRIME-NL study provides a unique window into the validity of a large-scale care evaluation for people with a chronic disease, in this case parkinsonism.
摘要：
背景：PRIME-NL研究前瞻性地评估了一种针对帕金森病患者的新的综合和个性化护理模式，包括帕金森病,在荷兰的选定地区（PRIME）。我们通过检查基线和1年的依从性数据来解决PRIME-NL研究的选择和混淆偏差的普遍性和来源。
方法：首先，我们使用荷兰几乎所有帕金森病患者(来源人群)的医疗保健索赔数据评估了PRIME和常规治疗(UC)地区之间的区域基线差异.第二,我们将我们的问卷样本与来源人群进行了比较，以确定泛化性。第三，我们通过比较PRIME和UC问卷样本的基线特征和1年依从性,调查了偏倚来源.
结果：在PRIME（n=1430）和UC（n=26,250）来源人群中，基线特征相似。组合问卷样本（n=920）比组合来源人群更年轻，疾病持续时间稍长。与PRIME地区的问卷样本相比，UC问卷样本稍年轻，有更好的认知，疾病持续时间较长，有更高的教育程度和消耗更多的酒精。UC地区（96％）的问卷样本的1年依从性高于PRIME地区（92％）。
结论：PRIME-NL研究的普遍性似乎很好，然而我们发现了一些选择偏差的证据。这种选择偏差需要使用先进的统计方法对PRIME-NL进行最终评估，例如逆概率加权或倾向得分匹配。PRIME-NL研究为慢性疾病患者的大规模护理评估的有效性提供了一个独特的窗口，在这种情况下帕金森症。
方法：首先，我们使用荷兰几乎所有帕金森病患者(来源人群)的医疗保健索赔数据评估了PRIME和常规治疗(UC)地区之间的区域基线差异.第二,我们将我们的问卷样本与来源人群进行了比较，以确定泛化性。第三，我们通过比较PRIME和UC问卷样本的基线特征和1年依从性,调查了偏倚来源.
结果：在PRIME（n=1430）和UC（n=26,250）来源人群中，基线特征相似。组合问卷样本（n=920）比组合来源人群更年轻，疾病持续时间稍长。与PRIME地区的问卷样本相比，UC问卷样本稍年轻，有更好的认知，疾病持续时间较长，有更高的教育程度和消耗更多的酒精。UC地区（96％）的问卷样本的1年依从性高于PRIME地区（92％）。
结论：PRIME-NL研究的普遍性似乎很好，然而我们发现了一些选择偏差的证据。这种选择偏差需要使用先进的统计方法对PRIME-NL进行最终评估，例如逆概率加权或倾向得分匹配。PRIME-NL研究为慢性疾病患者的大规模护理评估的有效性提供了一个独特的窗口，在这种情况下帕金森症。
