Rodriguez Anthony, Zeng Chengbo, Hays Ron D, Herman Patricia M, Edelen Maria O
RAND Corporation, Behavioral and Policy Sciences, 20 Park Plaza #910, Boston, MA, USA.
Patient Reported Outcomes, Value and Experience (PROVE) Center, Department of Surgery, Brigham and Women's Hospital, Boston, MA, USA.
Qual Life Res. 2025 Jan;34(1):35-42. doi: 10.1007/s11136-024-03826-6. Epub 2024 Nov 6.
This longitudinal study evaluates whether the Patient-Reported Outcomes Measurement and Information System (PROMIS)-16 domains capture average change over time comparable to the PROMIS-29 + 2 and have similar associations with change in overall health rating and two disability indices.
Data were collected using Amazon's Mechanical Turk at baseline, 3 months, and 6 months among individuals reporting chronic low back pain. The analytic sample includes respondents who completed baseline and at least one follow-up assessment (N = 1137). We estimated latent growth models for eight PROMIS domains and compared growth parameters between the PROMIS-16 and PROMIS 29 + 2 with a z-test. Additionally, for each domain, random intercept and slope scores for individuals were computed for the PROMIS-29 + 2 and PROMIS-16 and correlated to estimate concordance. Using growth parameters for physical function and pain interference, we predicted average change in the Oswestry Disability Index (ODI), Roland Morris Disability Questionnaire (RMDQ), the overall health rating, and compared regression coefficients between the PROMIS-16 and PROMIS 29 + 2.
All growth models fit the data well. Intercept and slope parameters were statistically comparable (p's > 0.05) in magnitude across all domains between the PROMIS-16 and PROMIS-29 + 2. Correlations between random intercept and slope scores for individuals across domains were high. Additionally, the regression coefficients between slopes for pain interference and physical function and ODI, RMDQ, and overall health rating were statistically comparable (p's > 0.05) between the PROMIS-16 and PROMIS 29 + 2.
Results provide between-level support for the longitudinal and predictive validity of the PROMIS-16. Similar average baseline scores and changes over time were observed between PROMIS-16 and PROMIS-29 + 2. Further, average change estimates comparably predicted average change in distal outcomes. This work provides evidence supporting the utility of the PROMIS-16 as a viable, short-profile option for use in clinical and research settings.
这项纵向研究评估患者报告结局测量与信息系统（PROMIS）-16领域能否捕捉与PROMIS-29 + 2相当的随时间的平均变化，以及与总体健康评级变化和两个残疾指数是否具有相似的关联。
在报告慢性下腰痛的个体中，于基线、3个月和6个月时使用亚马逊的Mechanical Turk收集数据。分析样本包括完成基线和至少一次随访评估的受访者（N = 1137）。我们估计了八个PROMIS领域的潜在增长模型，并通过z检验比较了PROMIS-16和PROMIS 29 + 2之间的增长参数。此外，对于每个领域，计算了PROMIS-29 + 2和PROMIS-16个体的随机截距和斜率得分，并进行相关性分析以估计一致性。使用身体功能和疼痛干扰的增长参数，我们预测了Oswestry残疾指数（ODI）、罗兰·莫里斯残疾问卷（RMDQ）、总体健康评级的平均变化，并比较了PROMIS-16和PROMIS 29 + 2之间的回归系数。
所有增长模型均很好地拟合了数据。PROMIS-16和PROMIS-29 + 2之间所有领域的截距和斜率参数在大小上具有统计学可比性（p值> 0.05）。各领域个体的随机截距和斜率得分之间的相关性很高。此外，PROMIS-16和PROMIS 29 + 2之间疼痛干扰和身体功能斜率与ODI、RMDQ和总体健康评级之间的回归系数具有统计学可比性（p值> 0.05）。
结果为PROMIS-16的纵向和预测效度提供了层面间的支持。PROMIS-16和PROMIS-29 + 2之间观察到相似的平均基线得分和随时间的变化。此外，平均变化估计值能相当程度地预测远端结局的平均变化。这项工作提供了证据，支持PROMIS-16作为一种可行的、简短形式的选项在临床和研究环境中的应用。
