AIM This study aimed at improving the utility of the fine motor subscale of the comprehensive developmental inventory for infants and toddlers (CDIIT) by developing a computerized adaptive test of fine motor skills. METHODS We built an item bank for the computerized adaptive test of fine motor skills using the fine motor subscale of the CDIIT items fitting the Rasch model. We also examined the psychometric properties and efficiency of the computerized adaptive test of fine motor skills with simulated computerized adaptive tests. RESULTS AND CONCLUSIONS Data from 1742 children with suspected developmental delays were retrieved. The mean scores of the fine motor subscale of the CDIIT increased along with age groups (mean scores = 1.36-36.97). The computerized adaptive test of fine motor skills contains 31 items meeting the Rasch model's assumptions (infit mean square = 0.57-1.21, outfit mean square = 0.11-1.17). For children of 6-71 months, the computerized adaptive test of fine motor skills had high Rasch person reliability (average reliability >0.90), high concurrent validity (rs = 0.67-0.99), adequate to excellent diagnostic accuracy (area under receiver operating characteristic = 0.71-1.00), and large responsiveness (effect size = 1.05-3.93). The computerized adaptive test of fine motor skills used 48-84% fewer items than the fine motor subscale of the CDIIT. The computerized adaptive test of fine motor skills used fewer items for assessment but was as reliable and valid as the fine motor subscale of the CDIIT. Implications for Rehabilitation We developed a computerized adaptive test based on the comprehensive developmental inventory for infants and toddlers (CDIIT) for assessing fine motor skills. The computerized adaptive test has been shown to be efficient because it uses fewer items than the original measure and automatically presents the results right after the test is completed. The computerized adaptive test is as reliable and valid as the CDIIT.

中文翻译：

---

改善婴幼儿全面发展清单中精细运动技能子量表的实用性：计算机自适应测试。

目的本研究旨在通过开发计算机化的精细运动技能适应性测试，来改善婴幼儿综合发展清单（CDIIT）的精细运动子量表的实用性。方法我们使用适合Rasch模型的CDIIT项目的精细运动子量表，建立了一个用于精细适应运动技能的计算机自适应测试的项目库。我们还通过模拟的计算机自适应测验检查了精细运动技能的计算机自适应测验的心理测量特性和效率。结果与结论检索了1742例可疑发育迟缓儿童的数据。CDIIT的精细运动分量表的平均得分随着年龄组的增加而增加（平均得分= 1.36-36.97）。精细运动技能的计算机自适应测试包含符合Rasch模型的31个项目 s假设（适应度均方= 0.57-1.21，适应度均方= 0.11-1.17）。对于6-71个月的儿童，精细运动技能的计算机自适应测试具有较高的Rasch人信度（平均信度> 0.90），较高的并发有效性（rs = 0.67-0.99），足以达到出色的诊断准确性（在接收器工作特性下的区域） = 0.71-1.00）和较大的响应度（效果大小= 1.05-3.93）。计算机化的精细运动技能适应性测试所使用的项目比CDIIT的精细运动子量表少48-84％。精细运动技能的计算机自适应测试使用的评估项目较少，但与CDIIT精细运动子量表一样可靠和有效。康复的意义我们根据婴幼儿综合发展清单（CDIIT）开发了一种计算机化的适应性测试，用于评估精细运动技能。事实证明，计算机化的自适应测试是有效的，因为它使用的项目比原始测量少，并且在测试完成后立即自动显示结果。计算机化的自适应测试与CDIIT一样可靠和有效。