1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References

Aim  The aim of this study was to describe how the usefulness of the hemiplegic hand develops in children with unilateral cerebral palsy (CP) aged between 18 months and 8 years.

Method  A prospective longitudinal study of 43 children (22 males, 21 females) with unilateral CP was conducted. Inclusion age was 18 months to 5 years 4 months (mean 2y 8mo [SD 1y 1mo]). Children were assessed with the Assisting Hand Assessment (AHA) 3 to 11 times per child over a mean period of 4 years 6 months. Two models were used for grouping children: by AHA score at 18 months and by Manual Ability Classification System (MACS) levels. Estimated average motor development curves were fitted with a nonlinear mixed-effects model.

Results  Children with a high AHA score (high ability level) at 18 months reached a significantly higher ability level and at a higher progression rate than children with a low 18-month AHA score. Limits of development differed between the three MACS levels.

Interpretation  Results indicate that the AHA score at 18 months can be used to discuss future development of affected hand use in bimanual tasks in children with unilateral CP.

List of abbreviations

Assisting Hand Assessment


Manual Ability Classification System

Development of hand skills in typically developing children is characterized by a rapidly increasing ability during early years followed by a refinement of these skills until adolescence.1–3 The course of typical development has been thoroughly investigated and there are several norm-referenced tests.4,5 Characteristically, children with unilateral cerebral palsy (CP) have one well-functioning hand and one impaired hand.6 Common features of the hemiplegic hand are slowness, abnormal muscle tone, decreased strength, and coordination difficulties, which occur to a varying extent in children independently of age.7 Many children also have impaired sensibility and mirror movements.8,9 Perhaps most importantly, children with unilateral CP have varying degrees of limitation in their ability to handle objects in daily life. This limitation is most obvious in situations where two hands are needed.10 However, knowledge about bimanual ability in children with unilateral CP and its development over time is scarce at present. Information on this development would be useful when discussing treatment planning and prognosis for development with the families.

Development is best investigated in longitudinal studies in which the same individuals are measured repeatedly over time. Longitudinal studies are time-consuming, but they provide a unique way to follow the course of change in individuals and explore differences in development between individuals.11 An already classic example of motor development curves based on longitudinal observations of children with CP was given in 2002 by Rosenbaum et al.12 in the area of gross motor function. They found distinctly different development of gross motor function for children in all five levels of the Gross Motor Function Classification System (GMFCS). This knowledge can be used to predict the rate and limitation of development in young children at different GMFCS levels. So far there has been only one study describing motor development curves of hand function for children with CP.13 In that study diverse trends were seen depending on the subtype of CP and the impairment level, and also depending on the measurement tool used. On the Peabody Fine Motor Scale, which measures general upper extremity skills, a steady increase in average ability was found during preschool years in children with unilateral CP. This increase was more marked in the mildly impaired group. Yet for the same group of children a different development pattern appeared when the Quality of Upper Extremity Skills Test (QUEST) was used, a test that evaluates quality of movements. For QUEST there was a peak at about age 4 years and a decline that was more marked in children with lowest ability. Fedrizzi et al.14 specifically studied children with unilateral CP and investigated both grip pattern and spontaneous hand use by means of nonstandardized assessments. They found no change over time in spontaneous hand use, but a slight improvement in the quality of grip at ages 4 to 7 years, which declined after 11 years of age in a subgroup with low ability. Eliasson et al.15 reported improvement at older ages in a 13-year follow-up from ages 6 to 8 years. In 10 people with different subtypes of CP they found a 45% mean time decrease in the Jebsen–Taylor test of hand function and a 22% time decrease for an experimental grip-lift task. The above studies describe development of hand function for children with different types of CP, and have different perspectives on development depending on the measurement used.

From previous studies it is clear that the choice of measure is crucial as an expression of the aspect of hand function described, and it also needs to be suitable for the subtype of CP studied. For children with unilateral CP, we believe that the most critical issue is how to use the affected hand in collaboration with the well-functioning, dominant hand. When investigating development of hand function for this group it is important to choose an assessment that captures this specific characteristic of hand function. In recent years the Assisting Hand Assessment (AHA) has been developed, which seeks to measure and describe how effectively children with unilateral disability use their affected hand to perform bimanual tasks.16 The aim of this study was to describe how the usefulness of the hemiplegic hand develops in children with unilateral CP between 18 months and 8 years of age.


  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References

This was a prospective longitudinal study with the aim of following change over time. In this design children were included and measured at different ages (unbalanced data), which was accommodated for in a nonlinear mixed-models analysis.11


A convenience sample of 43 children with unilateral CP were recruited through community habilitation centres located mainly in Stockholm, but also in a few other regions of Sweden. Occupational therapists passed on information about the study to families with children fitting the inclusion criteria, namely a diagnosis of unilateral CP and age 18 months to 5 years. No record was taken of families who were asked but declined. The mean age for entering the study was 2 years 8 months (SD 1y 1mo), ranging from 18 months to 5 years 4 months. There was an even distribution of sexes (22 males, 21 females) and slightly more children with right-sided unilateral CP (n=25). Fourteen children had been born preterm (<37wks gestational age). These characteristics correspond well to the population of children with unilateral CP.7 All children had access to local habilitation services, which include regular check-ups and intervention according to their needs by multiprofessional teams. Informed consent was given by all caregivers to participation in the study and to the publication of results. This study was approved by the Ethics Board at the Karolinska Hospital in Stockholm, Sweden.

Data collection

Measurements with the AHA were made every 6 months for children who were under the age of 3 years; after 3 years of age, data were collected once a year. For practical reasons this protocol was not always followed strictly. Each child was measured 3 to 11 times (mean 6.5 times per child) over a mean period of 4 years 6 months (range 2y 2mo–7y 2mo). In total 280 AHA assessments were included in the study. At the time of the final AHA assessment (at mean age 7y 3mo), children were classified with the Manual Ability Classification System (MACS) by parental interview.


The AHA is a test designed for children with congenital unilateral disability in the arm and hand. The AHA measures how effectively children use their affected hand in bimanual activities.16 In this study two versions of the AHA were used: the Small Kids AHA (18mo–5y) and School Kids AHA (6–12y). These versions have been shown to measure the same construct and give equivalent results.17 The AHA is scored on 22 items on a 4-point rating scale and yields both a sum score and a logit measure. Logit measures are interval level data, obtained in a Rasch analysis,18 ranging from –10.18 to 8.70 logits. In this study logit measures were used. The AHA has demonstrated excellent psychometric properties including rater and retest reliability, giving a smallest detectable difference of just under 1 logit.17,19 Videotaped AHA test sessions were performed by researchers or by local therapists. To minimize rater error all sessions were scored at the end of the study, and all videos for each child were rated by one of three experienced raters.

The MACS classifies how children with different types of CP use their hands to handle objects in their daily lives. The aim of the MACS is to report the involvement of both hands in manual activities. It has been validated for children aged between 4 and 18 years. The MACS has five levels, but children with unilateral CP are commonly found on levels I to III.10,20 In brief, a child at level I handles objects easily and successfully; at level II the child handles most objects but with somewhat reduced quality and/or speed of achievement, and alternative ways of handling objects may be used; at levels I and II manual abilities do not restrict independence in daily activities; at level III a child handles objects with difficulty and commonly needs help to prepare and/or modify activities. Distribution between MACS levels in this sample was as follows: level I, n=7; level II, n=25; level III, n=11.

Statistical analysis

Nonlinear mixed-effects models were used to evaluate the maximum level of AHA (limit) and the rate to achieve this maximum. The limit and rate were estimated by assuming a negative exponential function of development:11

  • image

The choice of a nonlinear mixed model with a negative exponential shape was based on an assumption that assisting hand use would increase with time in a decreasing rate, a pattern similar to that seen in typical development of manual abilities, and an assumption that the rate of increase and limit level would vary between individuals. These assumptions were confirmed when inspecting the observed individual development curves. The level of AHA at birth was assumed to be the same for all individuals and the minimum logit scale value of −10.18 was used as a known value for start before the analyses. Limit and rate were used as possible random effects in the analyses and were, therefore, estimated for every child. Two models for grouping children were used. In the first model, the AHA level at 18 months (observed or, if not available, predicted) was used to divide individuals into two groups named ‘high 18-month AHA score’ and ‘low 18-month AHA score’. In the second model, MACS levels at age 7 years were used to group individuals, and average curves for each MACS level were fitted. The individual limits and rates were used to estimate a mean limit and mean rate for making conclusions for each of the groups in the two models. Bayesian information criteria were used to evaluate which random effects should be included in the final model and to compare models. To further enhance the interpretation of rate, the age at which a child reaches 90% of their limit (age-90)12 was calculated for each child. The mean and 95% confidence limits of age-90 for each group were calculated by performing an ordinary nonparametric bootstrap with 3000 resamples; p<0.05 was considered statistically significant. The analyses and graphs were performed with the statistical software R, version 2.8.0 (R Development Core Team;


  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References

Assisting Hand Assessment scores increased over time in 37 children, more than the smallest detectable difference measurable with the AHA, although there were considerable differences between individuals. In three children the increase was smaller than the smallest detectable difference, and for three children there was both an increase and a subsequent decrease that exceeded the smallest detectable difference during the study period. To evaluate whether development curves were related to children’s ability level at age 18 months, the groups showing high versus low 18-month AHA scores were compared. A cut-off level at –3 logits between the groups was used. This cut-off level was chosen for two reasons. First, by studying the plotted results for age 18 months a distinct gap was observed at –3 logits. Second, on looking more closely at characteristics of the AHA scores and item response pattern for the children around this measure level it was seen that this ability level was the divide between children who mostly use a grip for holding objects (above –3 logits) and children who mostly do not hold objects with their affected hand (below –3 logits). In Figure 1, the observed development for each child, together with the estimated average development for each group and the high and low 18-month AHA score, is plotted. Children with a low AHA score at age 18 months have a lower AHA score limit of development; the parameter estimates and 95% confidence intervals (CI) are presented in Table I. Limits of development and rates of reaching the respective limits were significantly different between the groups (Table II). Expressed as age-90, the higher rate in the group with high 18-month AHA score means that these children reach 90% of their limit at an average age of 3 years, whereas children with a low 18-month AHA score who have a lower rate reach 90% of their limit at about age 7 years (Table I).


Figure 1.  Observed and predicted Assisting Hand Assessment (AHA) development. Separate curves for predicted development are plotted for the groups. High (n=27) or low (n=16) 18-month AHA score (above or below –3 logits on the AHA at age 18mo).

Download figure to PowerPoint

Table I.   Parameter, rate, and age-90 estimates for the two models 18-month Assisting Hand Assessment (AHA) and the Manual Ability Classification System level (MACS)
ModelnLimitRateAge-90 (mo)
Parameter estimate 95% CIParameter estimate 95% CIParameter estimate 95% CI
  1. CI, confidence interval; age-90, the age at which 90% of the limit level AHA is reached.

18-month AHA
 High272.141.36; 2.920.0640.057; 0.07136.535.3; 38.4
 Low16−0.24−1.41; 0.930.0290.023; 0.03687.178.8; 103.5
MACS level
 I74.423.17; 5.660.0560.039; 0.07245.141.4; 53.2
 II251.320.66; 1.990.0530.045; 0.06148.143.7; 58.4
 III11−0.16−1.48; 1.160.0270.016; 0.03811389; 182
Table II.   Comparisons between different limits and rates
ComparisonDifferenceSEM95% CIp
  1. SEM, standard error of the mean; CI, confidence interval; AHA, Assisting Hand Assessment; MACS, Manual Ability Classification System.

Limit (logits)
 High 18mo AHA – low 18mo AHA2.390.720.98; 3.81 0.001
 MACS I – MACS II3.090.72  1.68; 4.50<0.001
 MACS I – MACS III4.610.93  2.77; 6.39<0.001
 MACS II – MACS III1.490.76  0.01; 2.96 0.051
 High low 18mo AHA0.0340.005  0.025;  0.044<0.001
 MACS I – MACS II0.0030.009−0.015;  0.021 0.775
 MACS I – MACS III0.0290.010  0.009;  0.048 0.005
 MACS II – MACS III0.0260.007  0.012;  0.040<0.001

In the second model, MACS levels determined at the final AHA assessment were used to illustrate development (Fig. 2). Limits of AHA score were 4.42 logits for MACS level I and 1.32 logits for level II, which were significantly different, and the limit for MACS level III was –0.16 (see Tables I and II). The rates of reaching the limit had a similar pattern for MACS levels; children with lower limits took longer to reach their limit. However, MACS groups I and II had very similar age-90 values, whereas MACS group III had on average not yet reached 90% of their limit at age 8 years when this study ended.


Figure 2.  Observed and predicted Assisting Hand Assessment (AHA) development. Separate curves for predicted development are plotted for each Manual Ability Classification System (MACS) level. Level I, n=7; level II, n=25; level III, n=11.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References

The development curves, based on the AHA, illustrate how the use of the affected hand changes over time. A significant difference both in the limit and rate of development was found between the two groups in the 18-month AHA model. Differences in limits could be traced between MACS levels. The two models have different applications. The 18-month AHA model is useful for the approximate prediction of assisting hand development. The score from the first AHA assessment with young children can be used to discuss the prognosis of future development of affected hand use for children with unilateral CP. The MACS is currently only valid for children from 4 years of age and we do not yet know the stability of the MACS levels over time.10 Therefore, MACS is not useful at present as a predictor for development in young children. However, we found the graph showing development for children at different MACS levels valuable for illustration.

There are interesting clinical interpretations based on high and low 18-month AHA scores, although there were considerable differences within the groups. Typical hand-use behaviours can be described for the two groups on the basis of their AHA score profiles. Children with a low 18-month AHA score (raw score range 22–40) rarely perform bimanual play tasks independently at the age of 18 months. They do not spontaneously grasp objects with the hemiplegic hand but may hold on to objects placed in the hand by an adult and may use the arm to stabilize objects by weight or support. This group reaches 90% of their average limit level at about age 7 years. At this age they usually engage in self-initiated holding of objects, although the grip is not always stable. Sometimes they prefer to stabilize objects by weight or support without using a grip. They will now be able to perform many bimanual tasks independently. Children with a high 18-month AHA score (raw scores>40) are already performing bimanual tasks at 18 months, although with some difficulty. They typically use their well-functioning hand to pick up objects, which they thereafter grasp with the hemiplegic hand. The grip with the affected hand is not always reliable, so that objects sometimes slip. Children with a high 18-month AHA score reach 90% of their average limit at about age 3 years. At this age they usually manage bimanual tasks more effectively and use the hemiplegic hand as a fairly useful assisting hand. This means that they still prefer to use the dominant hand for picking up and putting down objects, but thereafter quickly transfer them to and from the assisting hand. In the assisting hand, objects are now most often reliably stabilized and well oriented for successful manipulation with the dominant hand.

An important finding in this study was that children who have a higher AHA limit reach their limit at a younger age than children with a lower limit (e.g. in the 18mo AHA groups). This is the opposite of the pattern described for gross motor function, in which children at GMFCS level I reach their limit later than children at levels III to V.12 This indicates that gross and fine motor development do not follow the same time course. The underlying brain lesion and other characteristics of the two groups with high and low 18-month AHA scores have not yet been investigated, but they may provide a further explanation of the different rates of development in the two groups.

A point worth noting is that all children in this study had access to local habilitation services with the opportunity for periods of intensive treatment, which has a documented effect on hand function.21–24 In Sweden all families with children with special needs are entitled to advice, support, and interventions from multidisciplinary teams at local paediatric rehabilitation centres, free of charge. Common interventions for children with unilateral CP are periods of functional group training, intensive training such as modified constraint-induced movement therapy, splinting and therapy after upper limb surgery, or spasticity-reducing treatments such as injections of botulinum toxin type A. This means that development of hand function for children with unilateral CP living in Sweden has to be understood as being influenced by interventions. These results, therefore, cannot be generalized to children with other service conditions.

In this study we used a nonlinear model for calculating average development curves, similar to that used by Rosenbaum et al.12 and named the ‘Stable Limit’ model in their follow-up study.25 The main reason for using this model was that it has a shape that reflects both our assumptions about development and empirical data. Furthermore, we found the calculation of age-90 used by Rosenbaum et al.12 very helpful for interpreting the rate parameter. The age-90 provides ages with confidence intervals until which it is likely that the children will continue to develop the use of their affected hand, which can be useful when discussing prognosis with professionals and families.

A limitation of this study was the small sample, which limits generalization to the population as a whole. This also resulted in large confidence intervals in the MACS model, in which the sample was split into three groups. The demographic data (sex, gestational age, and affected side) are representative of the population of children with unilateral CP.7 Despite this there is low representation of children in MACS level I in comparison with a population-based study from southern Sweden, in which most of the children with hemiplegia were found in MACS level I.20 There is a need for a replication of this study with a larger sample with established representativeness of the population.


  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References

This study has provided some new insights into how hand function develops in children with unilateral CP. The longitudinal design was a prerequisite to exploring how individuals develop, to describing the large differences in development between individuals, and to illustrating how an early AHA score can predict outcome at age 8 years, as in this sample. This material could be used in the further exploration of common factors and even predictors for different patterns of development.


  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. Conclusion
  7. References
  • 1
    Pehoski C. Object manipulation in infants and children. In: HendersonA, PehoskiC, editors. Hand Function in the Child: Foundations for Remediation. St. Louis: Mosby Elsevier, 2006: 14360.
  • 2
    Forssberg H, Eliasson A-C, Kinoshita H, Johansson R, Westling G. Development of human precision grip I: basic coordination of force. Exp Brain Res 1991; 85: 45157.
  • 3
    Forssberg H, Kinoshita H, Eliasson A-C, Johansson R, Westling G, Gordon A. Development of human precision grip II: anticipatory control of isometric forces targeted for object’s weight. Exp Brain Res 1992; 90: 39398.
  • 4
    Folio M, Fewell R. Peabody Developmental Motor Scales. Chicago: Riverside Publishers, 1983.
  • 5
    Haley S, Coster W, Ludlow L, Haltiwanger M, Andrellos P. Pediatric Evaluation of Disability Inventory. Boston: New England Medical Center Hospitals and PEDI Research Group, 1992.
  • 6
    Eliasson A-C, Gordon A, Forssberg H. Basic co-ordination of manipulative forces of children with cerebral palsy. Dev Med Child Neurol 1991; 33: 66170.
  • 7
    Uvebrant P. Hemiplegic cerebral palsy. Aetiology and outcome. Acta Paediatr Scand Suppl 1988; 345: 1100.
  • 8
    Krumlinde-Sundholm L, Eliasson A-C. Comparing tests of tactile sensibility: aspects relevant to testing children with spastic hemiplegia. Dev Med Child Neurol 2002; 44: 60412.
  • 9
    Kuhtz-Buschbeck J, Krumlinde-Sundholm L, Eliasson A-C, Forssberg H. Quantitative assessment of mirror movements in children and adolescents with hemiplegic cerebral palsy. Dev Med Child Neurol 2000; 42: 72836.
  • 10
    Eliasson A-C, Krumlinde-Sundholm L, Rösblad B, et al. The Manual Ability Classification System (MACS) for children with cerebral palsy: scale development and evidence of validity and reliability. Dev Med Child Neurol 2006; 48: 54954.
  • 11
    Singer J, Willet J. Applied Longitudinal Data Analysis. New York: Oxford University Press, 2003.
  • 12
    Rosenbaum P, Walter S, Hanna S, et al. Prognosis for gross motor function in cerebral palsy. JAMA 2002; 288: 135763.
  • 13
    Hanna S, Law M, Rosenbaum P, et al. Development of hand function among children with cerebral palsy: growth curve analysis for ages 16 to 70 months. Dev Med Child Neurol 2003; 45: 44855.
  • 14
    Fedrizzi E, Pagliano E, Andreucci E, Oleari G. Hand function in children with hemiplegic cerebral palsy: prospective follow-up and functional outcome in adolescence. Dev Med Child Neurol 2003; 45: 8591.
  • 15
    Eliasson A-C, Forssberg H, Hung Y-C, Gordon A. Development of hand function and precision grip control in individuals with cerebral palsy: a 13-year follow-up study. Pediatrics 2006; 118: e122636.
  • 16
    Krumlinde-Sundholm L, Eliasson A-C. Development of the Assisting Hand Assessment: a Rasch-built measure intended for children with unilateral upper limb impairments. Scand J Occup Ther 2003; 10: 1626.
  • 17
    Holmefur M, Aarts P, Hoare B, Krumlinde-Sundholm L. Retest and alternate forms reliability of the assisting hand assessment. J Rehab Med (Forthcoming).
  • 18
    Krumlinde-Sundholm L, Holmefur M, Kottorp A, Eliasson A-C. The Assisting Hand Assessment: current evidence of validity, reliability and responsiveness to change. Dev Med Child Neurol 2007; 49: 25964.
  • 19
    Holmefur M, Krumlinde-Sundholm L, Eliasson A-C. Interrater and intrarater reliability of the Assisting Hand Assessment. Am J Occup Ther 2007; 61: 7984.
  • 20
    Delhusen Carnahan K, Arner M, Hägglund G. Association between gross motor function (GMFCS) and manual ability (MACS) in children with cerebral palsy. A population-based study of 359 children. BMC Musculoskelet Disord 2007; 8: 50.
  • 21
    Eliasson A-C, Krumlinde-Sundholm L, Shaw K, Wang C. Effects of constraint-induced movement therapy in young children with hemiplegic cerebral palsy: an adapted model. Dev Med Child Neurol 2005; 47: 26675.
  • 22
    Hoare B, Wasiak J, Imms C, Carey L. Constraint-induced movement therapy in the treatment of the upper limb in children with hemiplegic cerebral palsy. Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.: CD004149. DOI: 10.1002/14651858.CD004149.pub2.
  • 23
    Fehlings D, Rang M, Glazier J, Steele C. An evaluation of botulinum-A toxin injections to improve upper extremity function in children with hemiplegic cerebral palsy. J Pediatr 2000; 137: 33137.
  • 24
    Hoare B, Imms C. Upper-limb injections of botulinum toxin-A in children with cerebral palsy: a critical review of the literature and clinical implications for occupational therapists. Am J Occup Ther 2004; 58: 38997.
  • 25
    Hanna S, Rosenbaum P, Bartlett D, et al. Stability and decline in gross motor function among children and youth with cerebral palsy aged 2 to 21 years. Dev Med Child Neurol 2009; 51: 295302.