Dartmouth Medical School, Hanover, NH, 03755, USA.
3 M Health Information Systems Inc, Salt Lake City, UT, USA.
Qual Life Res. 2018 Jan;27(1):51-58. doi: 10.1007/s11136-017-1573-x. Epub 2017 Apr 11.
Targeting resources for a designated higher-risk subgroup is a strategy for chronic care management. However, risk-designation has several limitations: it is inaccurate, seldom helpful for care guidance, and potentially misallocates care away from many patients.
To address limitations of risk designation, we tested a "what matters index" (WMI) in 19,593 adult patients with chronic conditions. The WMI contains five binary measures: insufficient confidence to manage health problems, level of pain, emotional problems, polypharmacy, and adverse medication effects. We examined its sum for association with patient-reported quality of life and prior emergency or hospital use. We compared its accuracy to a prototypic risk-designation model.
The WMI was a good indicator for quality of life and in three diverse test populations it was strongly associated with the use of hospital and emergency services. For example, a sum of WMI ≥2 was associated with twice as many average uses as none; for ≥3, uses were three times higher. However, since relatively few patients use costly care, both the WMI and a prototypic risk-designation model had comparably low-positive predictive values. The WMI uses the patient voice to identify needs strongly associated with quality of life. Akin to risk designation models, the WMI can be used to place patients into groups associated with levels of costly services, but neither is likely to forecast costly service use for individuals. However, unlike risk-designation models, the WMI is based on measures that will immediately guide care for every patient.
为指定的高风险亚组分配资源是慢性病管理的一种策略。然而，风险指定存在几个局限性：它不准确，对护理指导帮助不大，并且可能会将护理资源从许多患者身上转移走。
为了解决风险指定的局限性，我们在 19593 名患有慢性病的成年患者中测试了一个“重要指标”（WMI）。WMI 包含五个二进制测量值：管理健康问题的信心不足、疼痛程度、情绪问题、多种药物治疗和药物不良反应。我们检查了其总和与患者报告的生活质量和先前急诊或住院使用情况的关联。我们将其准确性与原型风险指定模型进行了比较。
WMI 是生活质量的良好指标，在三个不同的测试人群中，它与医院和急诊服务的使用强烈相关。例如，WMI 得分≥2 与平均使用次数增加一倍相关；得分≥3 与使用次数增加两倍相关。然而，由于相对较少的患者使用昂贵的护理，WMI 和原型风险指定模型的阳性预测值都比较低。WMI 利用患者的声音来识别与生活质量密切相关的需求。与风险指定模型类似，WMI 可用于将患者分组为与昂贵服务水平相关的组，但两者都不太可能预测个人的昂贵服务使用情况。然而，与风险指定模型不同，WMI 基于的测量值将立即为每位患者提供护理指导。
