|Year : 2016 | Volume
| Issue : 4 | Page : 235-246
Assessment of in-hand manipulation: Tool development
Kavitha Raja1, Payal Katyal2, Saumen Gupta3
1 JSS College of Physiotherapy, Mysore, Karnataka, India
2 Department of Physiotherapy, School of Allied Health Sciences, Manipal University, Karnataka, India
3 Sikkim Manipal College of Physiotherapy, Sikkim Manipal University, Sikkim, India
|Date of Web Publication||15-Nov-2016|
Dr. Kavitha Raja
JSS College of Physiotherapy, MG Road, Mysore, Karnataka
Source of Support: None, Conflict of Interest: None
Objective: The aim of this study is to develop an assessment tool for in-hand manipulation skills (IHMS) and establish its psychometric properties. Design: Items are pooled based on literature and expert opinion. Content validation was performed by ten rehabilitation professionals. The test was administered to 123 typically developing, and 15 children with hand dysfunction-cerebral palsy (3), developmental coordination disorder (5), and Down syndrome (7). The latter group was given intervention, specific to upper extremity for 15 days, and test was readministered. Rasch analysis for rating scale structure, fit statistics, and dimension analysis was done. Results: Content validation was analyzed qualitatively. Suggestions were incorporated which consisted of instructions for scoring and test administration. The four-level ordinal rating scale was appropriate according to Rasch analysis. Of fifty items, three misfit items from translation subscale were removed based on fit statistics and clinical decision. The final test has 47 items. The tool had excellent inter-tester reliability and test stability and was responsive to change. Conclusion: Assessment of in-hand manipulation is a robust tool for clinical use in assessment IHMS.
Keywords: Assessment, hand dysfunction, motor skills, tool development
|How to cite this article:|
Raja K, Katyal P, Gupta S. Assessment of in-hand manipulation: Tool development. Int J Health Allied Sci 2016;5:235-46
| Introduction|| |
Hand function is a continuum of activities that encompass gross grasp and release and a number of fine functions with the most sophisticated of these being in-hand manipulation (IHM). Exner defined IHMS as "adjustment of an object within the hand for optimal orientation after grasp" and classified the skill into translation, shift, and rotation., Translation is described as the linear movement of an object achieved by alternating movement between fingers and palm. Shift is the adjustment of an object achieved by alternating movement between finger and thumb pads. In rotation, an object is moved around one or more of its axes. All of these skills can be performed in isolation and while holding one or more objects within the palm. The latter skill is termed "with stabilization." Thus, all of the components of IHM, i.e., translation, shift, and rotation can be performed singly or with stabilization. Moreover, shift and rotation can be further classified into simple and complex.
IHM skills (IHMS) are components of fine motor skills that are routinely used by children and adults during performed of activities of daily living, recreation, and work. Many authors have proposed methods to assess IHMS. These include two tests and two assessment protocol.
IHMS has been unequivocally been established as a prerequisite to academic and leisure activities. Delay or deficiencies in IHMS can be one of the reasons for poor academic performance and apparent clumsiness in children.,, Many of these dysfunctions can be effectively managed with focused intervention aimed at increasing proficiency of hand use. Therefore, there is a need to have sensitive, easily available tool to assess this function.
There are two tests and three assessment protocols described in literature to assess IHMS. The two tests are IHM test - quality section (IMT-Q) and test of IHM (TIHM) which has been recently revised by Pont et al. and named as TIHM-R. The assessment protocols are unnamed protocol, and the Observation Protocol on IHM and Functional Skill Development. These methods differ from each other on the basis of type of task administered, scoring criteria, and the number of IHMS assessed. Difference in testing methods, scoring criteria, incomplete evaluation of IHMS and lack of psychometrics leads to inconclusive evidence on which tool to use as an outcome measurement for IHMS and a need to develop a standardized tool to assess IHMS.
Tool development can be broadly classified into tools developed according to the classical response theory and those developed according to item response theory. In child development research, a combination of the two approaches is ideal. The classical response theory recognizes that a test must have content validity, test-retest stability, inter-rater reliability and be sensitive enough to pick up change with intervention also known as responsiveness, the item response theory seeks to establish robustness of the tool within itself and as related to the person. These characteristics include hierarchy of items, difficulty level of items, the function of equal separation between scores, and ability to the test discriminate between levels of functioning. These functions are undertaken by Rasch analysis.
The aim of this study was to develop a tool following stringent methods of tool development. The objectives undertaken for this process are as follows:
- To develop an assessment on IHMS using a two-step Delphi process
- To evaluate content validity and test-retest, inter-rater reliability, and responsiveness of the IHMS
- To evaluate item hierarchy, item fit, item difficulty, and scoring of the IHM using Rasch modeling.
| Methods|| |
Ethical approval was obtained from the Institutional Ethical Board, and the study was conducted in two phase: Phase I was tool development and Phase II was evaluation of responsiveness.
Phase I: Tool development
Based on literature review, the construct was identified and defined. The construct was identified as IHMS. It was defined as adjustment of an object in the hand after grasp. Exner's classification of IHMS was followed to identify the subcategories. These were finger to palm translation, palm to finger translation, shift, simple rotation, and complex rotation each with and without stabilization.
Dimensionality of the construct
According to definitions found in literature, IHMS is a complex construct and can be further divided into subconstructs on the framework of types of IHMS. Hence, it was hypothesized that IHMS is a multidimensional construct and the dimensions identified were translation, shift, and rotation.
Item pool generation
Items were pooled on the basis of literature and experts opinion. The number of items for the test was decided on the basis of the table of specification for IMT-Q section. A total of fifty items were pooled [Table 1]. The test constructed in an activity format, and common activities were included. This was ensured by interviewing a group of parents (10) of children between 3 and 9 years of age to assess if the items chosen were familiar to their children. Only those items that were unanimously considered familiar and easily available to the target group were retained. Instructions to administer the test were formulated.
Type of measurement
The tool is meant to be a discriminative (performance based) tool. It was expected to help in identifying children with IHMS dysfunction when their performance was compared with age-matched norms. The tool is also expected to identify hand skill maturation in children. The quality of skill was considered in determining the child's score.
Based on the qualitative nature of the construct, a 4-point ordinal scale was formulated for scoring [Appendix 1].
IHMS is seen as early as 12-15 months of age, but the lower age limit of 3 years 6 months was decided for this study to ensure that the child was able to follow instructions. All components of IHMS develop by the age of 9-10 years, but speed and skill improve until 12 years of age. The scoring criteria for this test were confined to quality of skill; hence, the upper age limit was kept at 9 years 6 months.
Children were recruited from local schools. Children whose parents consented to their child's participation and who assented to the study were recruited. Teachers and parents were given a questionnaire and who assess the child's preferred hand for skilled activities. The questionnaire is given below. A standard handedness questionnaire cannot be used in India as handedness is preferred for cultural reasons. Hence, even left-handed individuals are expected to use their right hand for certain activities such as eating, writing, and offering objects to another person. Hence, preferred hand was considered as the hand the child chose for fine functions.
Children were recruited based on school reports to reports to ascertain that they were typically developing children younger than six were recruited if the pediatrician's report confirmed that they were typically developing.
The final test administered to 123 typically developing children. Children in the relevant age group of both genders, and those with ability to understand test instructions were included. Informed consent from teachers/parents and assent from the child were obtained. The test administered is described in Appendix 1.[Additional file 1]
The recorded activities of 12 children (1 girl and 1 boy in each age group) were scored by 2 of the authors blinded to each other's scoring. The test was administered to these 12 children after a gap of 2 weeks to assess test-retest stability.
Phase II: Responsiveness
Fifteen children with hand dysfunction were recruited for this phase. Participants included children with Down syndrome, and clumsiness are reported by teachers (7), children with developmental coordination disorder (5), and children with cerebral palsy classified as Manual Ability Classification System II (3). No effort was made to ensure a homogenous population as the aim of this study was to identify the metrics of the tool and not effects of intervention on a particular group of children. Children were excluded if they had any history of upper limb surgery, severe sensory loss (auditory or visual) or were unable to understand test instructions. Informed consent was taken from parents/caregivers and assent was taken from each child and test was administered. Fifteen days of intervention was given in the form of circuit training comprising proximal strengthening, grip strengthening, proprioceptive training, and task-specific activities for hand.
Duration of each training session was 25 min the test was readministered the day after the last intervention session. Proximal strengthening exercises were included as proximal joint stability is a prerequisite for IHM.
The [Figure 1] depicts the steps taken to develop the tool.
Data were analyzed as follows.
Content validity was analyzed as below:
- Items which had score of 8 or above on the Likert scale given by 80% of raters were retained
- Items which were scored below 6 by 80% of raters were discarded
- Other items were discussed in a consensus meeting and were retained/modified/discarded.
Test-retest stability and inter-tester reliability were determined by intraclass correlation (ICC) coefficients.
Factor analysis was done to determine internal consistency and factor loading followed by Rasch analysis partial credit scale model.,
Data were analyzed using SPSS 17 for Windows (SPSS Inc., Chicago, IL, USA) and WINSTEPS (version 184.108.40.206) for windows, Beaverton, Oregon, USA).
Step I - Rating scale analysis
The aptness of the four-level ordinal scale was analyzed according to the following guidelines:
- Each category of the scale should have minimum of ten responses
- Average measures should increase monotonically across the scale categories
- Outfit mean square (Mnsq) value for each category should be <2.0
- Step calibration should increase monotonically across the scale category.
The rating scale categories are considered disordered if they did not follow the above guidelines and reorganization of the scale categories would be deemed necessary.
Step II - Goodness of fit for person and items
Fit statistics was used to evaluate the person and item fit based on observed data and Rasch assumptions. The fit statistics is reported as two Chi-square ratios: infit and outfit statistics. Mnsq and Z standard distribution statistics (Zstd) are given for both infit and outfit. An item was considered as underfit when Mnsq was >1.3 and Zstd is >2.0 and is overfit when Mnsq is <0.75 and Zstd is <−2.0. The above values of infit and outfit Mnsq and Zstd and clinical decision were used to remove misfit items.
Based on the initial observations from factors analysis that IHMS is a multi-dimensional construct, fit analysis was done for each subscale separately and unfit items were removed. Fit analysis was rerun until all the items fit the model.
Step III - Dimensionality
The test is considered unidimensional when the eigenvalue of the first contrast was of 3.0 (<5% of variance). An eigenvalue above 3.0 indicates that the residual contrast has strength of three items and the test may be multidimensional. Principal components analysis (PCA) was done separately for each subscale and the total scale.
| Results|| |
Five therapists completed the content validation. None of the items were considered invalid by more than 1 responder. Hence, none of the items were removed. Suggestions were incorporated which consisted solely of instructions for scoring and test administration.
Factor analysis revealed that the scale loaded on three factors translation, shift, and rotation. The correlation of these subscales ranged from 0.7 to 0.9 for shift. A high correlation coefficient suggested that the items have too little dissimilarity between them.
Inter-tester reliability was acceptable at an ICC of 0.87. Test-retest stability was excellent at 0.95 for translation, 0.91 for shift with stabilization, 0.89 for simple rotation, and 0.82 for complex rotation.
Results of Rasch analysis are described as follows.
Rating scale analysis
Rasch analysis showed that the four-level ratings scale (0-3) used for the test had the ability to discriminate performance of a child across the score. The category frequency count, outfit values, average measures, and step calibration for the scoring scale are depicted in [Table 2]. The categories were redundant. The outfit Mnsq was <2 for all four categories. Both averages measure and step calibration increased monotonically from 0 to 3. Based on these results, the rating scale was accepted. The rating scale analysis was done separately for children with hand dysfunction as a part of the responsiveness study data. None of the categories were found redundant and it followed other assumptions of the Rasch model such as optimal number of responses in each category, hierarchical scoring, and adequate data in each category for stable estimate.
|Table 2: Rating scale analysis for four-level ordinal rating scale of assessment of in-hand manipulation |
Click here to view
Person response validity
Based on goodness-of-fit analysis, 26 (21%) children were found misfitting as more than 5% children were misfitting their data was removed. After eliminating misfitting children's data, further goodness-of-it analysis for test items was done with data of 97 (78.8%) children who demonstrated accepted fit.
Goodness-of-fit analysis for items
Fit analysis for items was done separately for each subconstruct of IHMS based on factor analysis results that IHMS is multidimensional. [Table 3], [Table 4], and [Table 5] depict fit statistics for all three subscales. Based on above-mentioned fit criteria and clinical decision, three items from translation subscale, i.e., translation 3, 4, and 10 were removed.
Hierarchy of in-hand manipulation skills
Hierarchical order of items of three subscales on the basis of measure is depicted in [Table 6]. Higher values of measure indicate higher difficulty levels. The results demonstrate that the easiest items were of translation subscale and the most difficult ones were from complex rotation. Most of the items in shift subscale had higher measure as compared to simple rotation. The person-item map [Figure 2] depicts this difficulty continuum of IHMS for typically developing children.
|Figure 2: Person-item map for 47 items of assessment of in-hand manipulation in typically developing children|
Click here to view
|Table 6: Hierarchical ordering and measures for the 47 items of 3 subscales of assessment of in-hand manipulation |
Click here to view
PCA was done after eliminating three misfitting items. The PCA of translation and shift showed unidimensional nature of these subconstructs. The first residual contrast of translation subscale had an eigenvalue of 2.6 suggestive of unidimensionality. The shift subscale also demonstrates unidimensionality based on its first residual contrast which had an eigenvalue of 1.7. The rotation subscale had an eigenvalue of 3.6 indicative of a hidden construct. Rotation subscale has two components simple and complex rotation. This was considered as the reason for eigenvalue of more than 3.
When the two subscales were analyzed separately, simple rotation had an eigenvalue of 2.1 and complex rotation had an eigenvalue of 2.8 on first contrast confirming the hypothesis.
All three subscales were combined and PCA was done for complete scale of IHMS. The complete scale had an eigenvalue of 8.6 on first residual contrast.
Responsiveness to change
The person-item map as shown in [Figure 3] and [Figure 4] showed that children with hand dysfunction were able to perform on difficult item postintervention as compared to preintervention. Most of them performed well on easier items of translation subscale, before intervention. On comparing rating scale analysis for pre- and post-intervention data, the frequency count increased for the highest category level and decreased for the lowest category level showing improvement in skill acquisition, thus establishing the ability of the scale to pick up change with intervention or responsiveness.
|Figure 3: Preintervention person-item map for children with hand dysfunction on Assessment of in-hand manipulation|
Click here to view
|Figure 4: Postintervention person-item map for children with hand dysfunction on assessment of in-hand manipulation|
Click here to view
| Discussion|| |
The assessment of IHM (AIHM) was developed to assess IHMS for children. The number of items pooled in each subscale was based on from literature. However, we ensured that the items were familiar to Indian children from all socioeconomic strata. The final comprises 47 items - 19 in translation, 5 in shift, and 24 in rotation subscale.
Three items from translation subscale, taking the lid off the jam jar, moving the lid to put on jam jar, and crumpling of the paper to make a ball of it, were removed based on clinical decision and fit statistics. The size of the lid and paper was not changed according to the size of child's hand, and hence these activities involved more proximal joint strength than IHMS. This may be one of the reasons for under-fit of these items and formed the clinical basis for removing the item. Two of the under-fit items based on manipulation clay were retained based on theoretical constructs as manipulating clay is undeniably a component of IHMS. One of the items, from shift subscale-adjusting pen so that finger and thumb, is closer to writing pad was under-fitting but it was retained as it is an important activity for children when they write. In complex rotation, one item-picking and rotating pen to write was under fitting. It was retained based on clinical decision and published literature.
The hierarchy of skill difficulty observed in this study as depicted by item measure was that items in translation subscales were easiest than those in complex rotation were most difficult. Some of the items from simple rotation were easier as compared to shift items. The reason for this can be attributed to the child's familiarity to the object used in the particular activity. Hand skill development is based not only on neuromaturation but also on cultural demands and previous exposure to the object can have significant effect on skill development. In shift, the most difficult items were - pushing pen top with one hand and shifting of key to the tip of the fingers. Children may not have manipulated a lock and key previously, and pushing single handed requires more strength then manipulation. Hence, we suggested that these items may be used with clinical judgment for younger children.
Those items in simple rotation which had lower measure, as compared to shift were items involving commonly used objects such as plastic top, nut and bolt, crayons, and pegs. Ease in manipulating these objects may be due to prior experience with these objects. The number of items in shift subscale was few (5) as compared to other subscales. This was decided based on Exner's specification table. However, there is a need to add more items to shift subscale before concluding that shift is more difficult than simple rotation.
Based on literature, we had initially hypothesized that IHMS is a multidimensional construct and translation, shift, and rotation are its subconstructs. Initial factor analysis conformed this hypothesis. According to PCA, translation and shift subscale met criteria, confirming the unidimensional nature of these constructs. Rotation subscale had an eigenvalue of 3.6 suggestive of hidden construct or dimension. The PCA showed that simple and complex rotation was loaded as two contrasts. Both types of IHMS require the child to rotate the object within the hand, but in simple rotation, the object is rotated through one-fourth or one-half around its axis, where as in complex rotation the object should be rotated 180-360° around its axis. Considering this difference in the skill, and results of PCA analysis, rotation subscale was divided into simple and complex rotation and dimension analysis was repeated, conforming the hypothesis. Thus, these two types of rotation may be considered as separate constructs. Therefore, from the results of our study, the components of IHMS are translation, shift, simple rotation, and complex rotation with and without stabilization.
The four-level ordinal scale used in this study was deemed appropriate. The test has 47 test items divided into 4 subscales translation, simple shift, and complex rotation. The test items were easy and children were interested in the test. Time required to administer this test is 15-20 min, and therapists may require more time to administer the test to younger children, due to shorter attention span and emerging skills. We suggest that the test may be administered in two sessions for younger children. The test should be administered with the child positioned comfortably on a chair, feet rested on floor and elbow flexed to 90°. The test items have to be handed to the child and quality of movement observed simultaneously, for accurate observation of skill, we recommend video analysis; however, analysis can be done with naked eye if an assistant is present to hand object to the child.
This study was done on children from a wide age group. This was done to allow for interpretation of the development of IHMS. As is evident from the item difficulty map, the continuum of skill development is in the order of translation, simple shift, simple rotation, complex shift, and finally complex rotation. Translation is completed by the age of 6 years and complex rotation is still developing at the upper age limit of the children included in this study. Thus, this study gives a map of the development of IHMS in childhood.
The current study has attempted to develop a tool to comprehensively evaluate IHMS in children taking into consideration the limitations of existing tools. This tool has a wider applicability in terms of skills and age compared to TIHM and TIHM-R. The current tool by describing the items and methodology of testing makes the assessment of IHMS easily available to the clinician. IMT-Q the other comprehensive tool available does not list the test items in sufficient details for reproduction. The clarity of test items, scoring, the research rigor, and analysis using both classical response and item response construct are the main strength of this study. AIHM can be used as an outcome measure in children with dysfunction including children with developmental coordination disorder, cerebral palsy, attention deficit disorder, children with sensory processing dysfunctions, and in children who have difficulty with hand skills not attributable to a medical reason. The AIHM is sensitive enough to show effect of intervention and hence can be used as an outcome measure to gauge the effectiveness of treatment protocol.
Limitation and future research
The time taken to complete the item is an important aspect of skill maturation. However, we did not consider this in our scoring criteria. Hand size of the child should be considered when designing a test for fine as size of object may affect the quality of skill. Future research is warranted on improving the quality of this test considering the hand size and speed of skill. Future research on children with fine motor dysfunction and developing age appropriate norms for IHMS are also future goals.
| Conclusion|| |
The AIHM is a robust tool to assess IHMS in children.
This tool can be used to identify and plan remediation of dysfunctions of hand skills related to writing, craft, life skills, and play activities. The specific subset of IHMS can be identified as the scale is divided into four constructs. This test can be easily administered by teachers, therapists, and other child-care experts.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Exner CE. Development of hand functions. In: Pratt PN, Allen AS, editors. Occupational Therapy for Children. St Louis, MO: Mosby-Year Book; 1989. p 235-59.
Exner CE. In-hand manipulation skills in normal young children: A pilot study. Occup Ther Pract 1990;1:63-72.
Pehoski C, Henderson A, Tickle-Degnen L. In-hand manipulation in young children: Translation movements supporting children in their work. Am J Occup Ther 1997;51:719-28.
Humphry R, Jewell K, Rosenberger RC. Development of in-hand manipulation and relationship with activities. Am J Occup Ther 1995;49:763-71.
Case-Smith J. The effects of tactile defensiveness and tactile discrimination on In-hand manipulation. Am J Occup Ther 1991;45:811-8.
Exner CE. The zone of proximal development in in-hand manipulation skills of non-dysfunctional 3 and 4-year-old children. Am J Occup Ther 1990;44:884-91.
Pehoski C, Henderson A, Tickle-Degnen L. In-hand manipulation in young children: Rotation of an object in the fingers. Am J Occup Ther 1997;51:544-52.
Pont K, Wallen M, Bundy A, Case-Smith J. Reliability and validity of the test of in-hand manipulation in children ages 5 to 6 years. Am J Occup Ther 2008;62:384-92.
Exner CE. Content validity of the in-hand manipulation test. Am J Occup Ther 1993;47:505-13.
Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide to their Development and Use. 4 th
ed. Oxford, NY: Oxford University Press; 2008.
Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. 3 rd
ed. Upper Saddle River, NJ: Pearson and Prentice Hall; 2009.
Bond TG, Fox CM. Applying the Rasch Model: Fundamental Measurement in the Human Sciences. 2 nd
ed. Mahwah, NJ: Lawrence Erlbaum Associates; 2007.
Linacre JM. Optimizing rating scale category effectiveness. J Appl Meas 2002;3:85-106.
Spector PE. Quantitative applications in the social sciences. Summated Rating Scale Construction: An Introduction. Vol. 82. Los Angeles: SAGE Publications Inc.; 1992.
Smith-Zuzovsky N, Exner CE. The effect of seated positioning quality on typical 6- and 7-year-old children′s object manipulation skills. Am J Occup Ther 2004;58:380-8.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]