In pediatric physical therapy, multiple intervention programs are used for the treatment of infants with or at risk for developmental disorders, such as cerebral palsy (CP). Programs frequently used during a stay in the neonatal intensive care unit are the Newborn Individualized Developmental Care and Assessment Program, Kangaroo Care, and sensory stimulation programs. Programs used after term age are quite different in approach and vary from treatment according to Vojta, to neurodevelopmental treatment (NDT), to multiple forms of specific and general developmental programs.1 Four recent reviews1–4 did not find conclusive evidence that early physical therapy intervention improves motor outcome in children with or at risk for CP. Literature also does not provide evidence that early physical therapy produces other benefits, such as the enhancement of social-emotional or cognitive function.2,3
A theory-driven evaluation5 might help to understand why evidence of the effectiveness of early intervention programs is lacking. Theory-driven evaluations pay attention to the mechanisms underlying outcome. This means that they evaluate not only the effect of intervention on outcome (eg, “does application of early intervention result in a higher score on the Bayley Scales of infant development?”) but also the pathway along which the outcome is achieved (eg, “is application of a specific physiotherapeutic [PT] action related to outcome?”).
Theory-driven evaluation offers a framework to understand the relationship between the program input, the contents of the program (the black box), and the outcomes.6 Knowledge about the effect of specific elements of an intervention is crucial for improvement of physical therapeutic guidance. Guidance should focus on elements that have a beneficial effect on outcome, and elements that do not affect outcome or have an adverse effect should be discarded. Uncovering the actual contents of physical therapy practice will shed light on potentially effective and ineffective elements of the intervention.
The aim of this study was to develop an instrument to objectify what physical therapists actually do during treatment, that is, to open the black box of physical therapy sessions in daily practice, applied to infants at high risk for a developmental disorder such as CP. It should be noted that CP develops in only a portion of infants at high risk for developmental disorders. Yet, motor problems and difficulties in learning develop in a substantial proportion of these infants at school age.7
To this end, we developed an observation protocol based on the theoretical constructs of the most commonly used physical therapy approaches for high-risk infants in the Netherlands, that is, NDT and functional therapy.8,9 The term functional in the latter approach may give the impression that NDT does not aim at improving the infant's function. This is a false impression; all approaches in infant physical therapy aim at improving the child's function. Nevertheless, some differences exist between NDT and functional therapy. For instance, NDT is characterized by a broad repertoire of therapist-infant contact strategies, whereas the functional approach relies entirely on infant self-produced motor behavior. Primary questions addressed in this study deal with the psychometric properties of the protocol. (1) Is the observation protocol complete, that is, is it possible to classify the majority (>90%) of the contents of physical therapy sessions? (2) Are the observation items mutually exclusive? (3) Can the items be assessed reliably in terms of inter- and intraassessor agreement? Secondary questions address the actual content of the physical therapy sessions during early infancy: (4) Which observable therapeutic strategies can be distinguished? (5) What proportion of time is spent on the different strategies?
Development of the Observation Protocol
Protocol development started with a systematic analysis of the literature on the application of physical therapy in infants with or at risk for developmental motor disorders such as CP. This analysis gave us clues to the most important strategies that are used by pediatric physical therapists. These strategies are based on the theoretical constructs of the most commonly used physical therapy programs for high-risk infants, that is, NDT and functional therapy. The NDT approach was developed in the 1940s by Karel and Berta Bobath. Inhibition of spasticity and facilitation of normal posture and motor behavior were the main aims of treatment, which the Bobaths tried to accomplish by tone-influencing patterns with the use of various therapeutic handling techniques.10,11 Over the years, NDT gradually evolved based on new theoretical constructs and practical knowledge. Although the basic philosophy and the approach to treatment remained the same, goals, treatment techniques, and treatment application obtained a more functional orientation.11,12 Currently, the main goals of NDT are to increase or improve the skill repertoire, to maintain the functional level of performance, and the general management and minimization of contractures and deformities.11
The way in which NDT is applied depends on the child's age and the severity of the disability. As early as the 1960s, Köng and Quinton started to adapt and develop techniques for early treatment of babies based on the Bobath concept.13,14 The principles of NDT baby treatment13 aim at helping infant motor development by giving sensorimotor experiences of typical motor patterns to the infant. This is achieved by using several forms of facilitation techniques, such as handling or pressure techniques, by using support devices, and by giving sensory experience to the infant. Therapy should not be restricted to the therapy sessions; the therapist aims by means of training the caregivers to achieve a carryover from treatment into activities of daily life.10,11,15
The focus of functional therapy lies on improvement of functional tasks.8,9 Application of the concept of functional therapy to young infants implies that the cornerstone of guidance consists of encouragement of self-produced motor behavior during functional tasks, such as feeding, dressing, bathing, and playing.8,9,16,17 From the literature, it is known that infants at high risk for a developmental motor disorder have a reduced repertoire of motor strategies available for exploration. This is already expressed during the first postnatal months in a limited repertoire of general movements (GMs) and continues when goal-directed motility emerges. The infants also have difficulties selecting the most appropriate solution for a certain task out of their motor repertoire.18 Functional therapy therefore aims through play to promote variation in motor behavior and opportunities to explore the motor possibilities to find appropriate solutions for motor tasks.8,18,19 Part and parcel of the more functional approach is the coaching role of the therapist, that is, the caregivers determine how developmental strategies highlighted by the therapist may be integrated into the child and family's daily life.
In the next stage of protocol development, we analyzed 20 pilot video recordings of infant treatment sessions in terms of directly observable PT actions. All actions that physical therapists performed were documented. These PT actions were classified and defined into categories. The categories varied from classic NDT actions such as handling techniques to categories describing functional activities (self-produced motor behavior), family involvement, and education. The observation protocol is the final result of a cyclical process that incorporated knowledge about physical therapy for infants and the observation of overt and directly observable PT actions during treatment from video recordings. During the development of the observation protocol, we paid attention to the following psychometric requirements: completeness, mutual exclusiveness, reliability, and construct validity.20
The contents of physical therapy sessions were analyzed for a study group of 22 infants, who had been admitted to the neonatal intensive care unit of the University Medical Center Groningen. All infants had an indication for early PT intervention on the basis of the presence of definitely abnormal GMs at the age of 10 weeks of corrected age. The GM method is a standardized technique to assess neurological integrity on the basis of the quality of spontaneous motor behavior.21,22 GMs are spontaneously generated complex movements involving the head, trunk, arms, and legs. Four classes of GM quality can be distinguished: normal-optimal, normal-suboptimal, mildly abnormal, and definitely abnormal. Various studies showed that the presence of definitely abnormal GMs at 2 to 4 months post-term indicates a high risk for developmental disabilities, such as CP.22–24 The GM assessment has a good predictive validity for both major and minor developmental disorders. Also construct and concurrent validity and reliability of the GM method are satisfactory.25 Infants with severe congenital anomalies and infants whose caregivers had incomplete understanding of the Dutch language were excluded from the study. All caregivers of the infants signed an informed consent form, and the research project was approved by the Ethics Committee of the University Medical Center Groningen.
Twenty of the infants were born preterm with a gestational age at birth that varied from 28 to 32 weeks (median, 31) and a birth weight between 630 and 2090 g (median, 1205 g). Two infants were born at term with a median birth weight of 3560 g. Neurological condition was assessed at 6 months of corrected age with the Touwen Infant Neurological Examination.26 Three infants showed clear neurological dysfunction. Fourteen infants showed a high number of signs of minor neurological dysfunction. The remaining 5 infants had a normal-suboptimal neurological condition at 6 months (Table 1).
Recording of Intervention Sessions
All 22 infants received physical therapy between the ages of 3 and 6 months of corrected age. Therapy was delivered by 17 physical therapists who were registered in the Netherlands as pediatric physical therapists. Fifteen therapists treated the infants in their home environments; and 2 applied the intervention in a clinical setting. Treatment duration varied from 12 to 50 minutes per session (mean, 30 minutes per session; SD, 9 minutes per session).
At 4 and 6 months of corrected age, a video recording of a treatment session of each infant was made. To minimize intrusion, the camera was positioned as far away from the physical therapist and the infant as possible. In addition, therapist and caregiver were asked to ignore the person making the video so that the recording would resemble a natural situation. Research has shown that video recording hardly affects the behavior of people who are being filmed because they tend to forget the presence of the camera after awhile.27 Care was taken that the physical therapist, caregiver, and infant were continuously in view of the camera.
The video recordings were analyzed with the standardized observation protocol with help of a Noldus software program, The Observer (version 5.0; Noldus, Wageningen, the Netherlands), a program specially designed for behavioral observation. The program allows the quantification of the duration, frequency, and serial order of defined therapeutic actions. For example, imagine the following sequence of actions: (A) a physical therapist offers an infant a toy in the midline; (B) after 6 seconds, the therapist facilitates rolling behavior with the pelvis or legs as a key point of control; and (C) meanwhile the therapist gives the caregiver instructions on handling during rolling movements. In The Observer program, action A is scored as category H (see Appendix, which is available online at http://links.lww.com/PPT/A8; the infant is challenged to produce motor behavior by himself or herself; activity flows over into therapeutic handling). As the toy is offered in the midline only, the action is performed with little variation (H1). Action B is scored as category C1 (facilitation techniques: handling), but at the same time the physical therapist interferes with the infant's behavior and interrupts his or her activity (category A.4.1). Action C is scored as category A.4.3 (caregiver training). Start and stop of an action are indicated by pressing a key on the computer keyboard. It is important to note that The Observer program allows scoring of multiple actions occurring simultaneously (eg, action B). To compare the results of different treatment sessions, the duration of the PT actions was converted into a relative duration, which represented a percentage of the total treatment time. Relative duration (%) = (time spent on PT action/total time of treatment session) × 100.
Psychometric Quality of the Observation Protocol
Completeness of the protocol was tested by checking whether all possible PT actions observed by the researchers were covered by the categories of the observation protocol. Completeness was considered satisfactory if major part of the PT actions (>90% of the observation time) could be classified into protocol categories. Mutual exclusiveness was determined by means of inter- and intrarater agreement. Substantial agreement indicates that a single PT action is clearly related to a specific category of the observation protocol. Reliability of the observation protocol was measured through inter- and intrarater agreement. Three time intervals, each lasting 5 minutes, were selected from 5 randomly selected treatment sessions (n = 15). The intervals 100 to 400 seconds, 500 to 800 seconds, and 1000 to 1300 seconds were analyzed. The start of the treatment session, that is, the first 100 seconds, was excluded from the analyses because during this period mainly preparatory actions were performed. Interrater agreement was assessed by comparing observational scores of the third author with those of the first author. Intrarater agreement was assessed by comparing the observational scores of the first author who reanalyzed the video sequences after an interval of at least 3 weeks.
The data were analyzed using the computer package SPSS (version 14.0; SPSS Inc, Chicago, Illinois). Inter- and intrarater agreements were calculated by intraclass correlations (ICCs) of frequency and relative duration of PT actions. ICC values between 0.50 and 0.75 were considered to indicate satisfactory reliability; ICC values exceeding 0.75 indicate good reliability.28
For the comparison of the relative duration of the main categories and subcategories of PT actions and the amount of postural support at 4 months with the data at 6 months, Wilcoxon signed rank tests were used.
To get an impression of the current state of daily practice in pediatric physical therapy in the Netherlands, we analyzed whether sessions were dominated by PT actions that are in line with the concept of NDT in baby treatment13 or by PT actions belonging to the functional approach. A treatment session was classified as NDT-like when the time spent on the category facilitation techniques exceeded the 75th percentile of current group data. Similarly, sessions in which time spent on the category “challenging the infant to produce motor behavior by himself or herself–action continued by the infant” exceeded the 75th percentile were classified as functional sessions.
Correlations and differences with a P value <.05 were considered statistically significant.
The observation protocol contains 3 levels of observation. The first level consists of 8 mutually exclusive main categories of PT actions (see the Appendix, available online at http://links.lww.com/PPT/A8). The second level consists of the subcategories of the main techniques. For example, the main category facilitation techniques has the following subcategories: handling, pressure techniques, transitions, and support devices. The third level of the protocol entails concrete PT actions: for example, in the supine or sitting position, the shoulders of the infant function as a key point, hands guide the shoulders of the infant in protraction to control the infant's movements and to facilitate hand-hand contact and symmetry. The degree of postural support provided by the physical therapist or caregiver when the infant was in the prone, side, or sitting position was scored as an additional variable. Postural support was classified into 4 categories: no postural support, minimal postural support, clear postural support, and full postural support.
The contents of 42 physical therapy sessions were analyzed. For 2 infants, only 1 video recording was available for observation. In 1 case, the therapist had already discontinued treatment before 6 months. In the other case, the family was on holiday when the video recording was scheduled.
The majority of the observed PT actions could be classified into the categories of the observation protocol. Only 3% of the duration of physical therapy sessions could not be classified into the categories of the observation protocol, that is, they were classified in the category “not specified.” This indicates that the observation protocol meets the criterion for completeness.
Inter- and intrarater agreement on the frequency of PT actions was moderate to good with ICCs varying from 0.68 to 1.00 (interrater) and 0.70 to 0.98 (intrarater). The same was true for the inter- and intrarater agreement on the relative duration of actions; ICCs ranged from 0.76 to 1.00 (interrater) and 0.69 to 0.99 (intrarater; Table 2). The confidence intervals of the PT actions are presented in Table 2. The data indicate that the PT actions of the observation protocol can be assessed in a reliable way. Moreover, the moderate to good inter- and intrarater agreement points to a sufficient degree of mutually exclusiveness.
Relative Duration of Therapeutic Actions
The data on frequency and relative duration of PT actions were highly correlated (Spearman rank correlation at 4 months: 0.89, at 6 months: 0.88). This suggests that both parameters carry virtually identical information. We therefore decided to restrict our report on the contents of the sessions to 1 parameter. We chose relative duration as it had slightly better inter- and intrarater agreement values than frequency.
The relative duration of the majority of the PT actions applied at 4 months did not differ significantly from those at 6 months (Table 3). The only exceptions were challenging the infant to produce motor behavior by himself/herself, which flows over into handling, which occurred slightly more at 6 months than at 4 months (P = .03) and in the provision of postural support. Less postural support was provided during PT actions in infants aged 6 months than in those aged 4 months (Table 3).
In the following report on the contents of the NDT sessions, we pooled the 4- and 6-month data for those actions in which the relative duration of PT actions at 4 and 6 months were identical. Table 3 shows that the application of physical therapy in infancy is characterized by great heterogeneity. Most of treatment time was spent with PT actions in the categories of motor behavior produced by the infant, that is, motor behavior that was initiated by the infant and behavior that was challenged by caregiver or physical therapist who (1) let the infant try on her or his own or (2) took over control of the infant's movements by means of handling (together ∼40%), and facilitation techniques (29%; Figure 1). We noted that during both challenging the infant to produce motor behavior by himself/herself, which flows over into handling (5.6%), and challenging the infant to produce motor behavior by himself/herself, which is continued by the infant (13.2%), infants were challenged to explore only 1 movement strategy. The most frequently applied subcategory of facilitation techniques was handling. Approximately 9% of treatment time was spent on sensory and passive motor experience. At 6 months, less postural support was provided during PT actions (Table 3).
The data indicated that the 75th percentile of the facilitation techniques category matched 44% of treatment time, whereas the 75th percentile of the category challenging the infant to produce motor behavior by himself/herself, which is continued by the infant was seen during 20% of treatment time. In 10 sessions, more than 44% of the treatment time was spent of facilitation techniques indicating that these sessions could be classified as NDT-like. Another 10 sessions could be classified as functional as more than 20% of treatment time was spent with the category challenging the infant to produce motor behavior by himself or herself, which is continued by the infant, leaving 22 sessions with mixed contents.
During all treatment sessions, 1 of the caregivers was present and observed what the physical therapist was doing. In 12 sessions, caregivers acted together with the physical therapist to guide the attention of the infant. Approximately 4% of the time was devoted to family involvement and educational actions. This usually consisted of educational actions for example interfering with infant's activities. Very little treatment time was devoted to the training of caregivers on how to continue treatment strategies during daily life activities and/or in the home environment. Treatment related to communication between physical therapist and caregiver was more or less equally distributed between information exchange, imparting knowledge, giving feedback, and sharing information regarding handling and the application of intervention strategies into daily routines.
Discussion and Conclusion
This study demonstrated that it is possible to assess PT actions during physical therapy of young infants at high risk for developmental disorders in a systematic, standardized, and reliable way. Before we address physiotherapeutic considerations, we discuss the strengths and weaknesses of our study.
One of the limitations of the study is the sample size, which results, for example, in relatively large confidence intervals for the ICC values of intra- and interrater agreement. Confidence intervals were especially large in infrequently occurring PT actions. The large confidence intervals indicate that the results cannot be generalized to any population. Therefore, it is important to realize that the results of this study should be interpreted with caution.
Another problem is that at young ages CP cannot be diagnosed. This means that physical therapy at very early ages is applied to infants at risk for developmental disorders, including CP, not to infants with CP. This study group was selected on the basis of definitely abnormal GM around the age of 3 months of corrected age. The presence of definitely abnormal GMs at this age is a powerful indicator of developmental disorders, including CP.22–24 Most of the infants participating in the study continued to show neurological dysfunction during the intervention period, indicating that they had need of ongoing physical therapy.
The major strength of this study is that we succeeded in developing a standardized observation protocol that permitted analysis of the black box of physical therapy interventions for infants. The protocol is an instrument that may be used to assess heterogeneity in physical therapy because it offers a tool to describe operationally what therapists do in a treatment session, both across varying frames of theoretical references and across countries. The video analysis of the contents of physical therapy sessions is relatively time-consuming because it is performed on a real-time basis and usually multiple runs are needed to track down simultaneously occurring actions. Experience at our institute indicates, however, that the technique can be learned relatively quickly and reliably. The analysis of PT actions by means of the standardized observation protocol paves the way for an objective interpretation of the contents of various pediatric physical therapy sessions.
Physiotherapeutic Considerations and Implications for Future Research
It was possible to develop an observation protocol to assess the contents of physical therapy for young infants. Eight main categories were developed that allowed the classification of single PT actions. During the developmental process, the observation protocol was checked for psychometric properties, such as completeness, mutual exclusiveness, and reliability. The results showed that the observation protocol caught virtually all PT actions during the treatment of young infants and that it had a good inter- and intrarater reliability. The good reliability also indicated that the categories in the observation protocol were mutually exclusive. The small proportion of treatment time (3%) that could not be classified into the categories of the observation protocol mainly was spent on comforting the infant and physical therapist or caregiver and infant being out of camera view. We may thus conclude that our observation protocol is an appropriate instrument for opening the black box of physical therapy for infants with a high risk for developmental disorders.
The data represent the contents of various pediatric physical therapy sessions. They give an indication of how physical therapy is applied to young infants at high risk for developmental disorders in the Netherlands. The data showed that the relative duration of the applied treatment techniques covers a wide range, implying a large heterogeneity in the practical implementation of physical therapy for infants. Presumably, this does reflect reality because part of the PT actions that are performed belong to the concept of NDT and baby treatment, which by itself are well-known for its heterogeneity in application.29 The combination with PT actions from a more functional point of view, such as the encouragement of self-produced motor behavior, makes treatment application even more diverse. Presumably, the heterogeneity is brought about by the evolution of treatment techniques and theoretical assumptions over the years. First, physical therapists became aware that motor achievements occurring after the application of former treatment techniques, such as those of NDT, did not automatically carry over into activities of daily life. Nowadays, goals are defined more in relation to function, the needs of the caregiver, and new treatment approaches that have been incorporated into older ones.12 With respect to our opening the black box tool, it is important to note that—despite the heterogeneity in the implementation of physical therapy for young infants—virtually all PT actions could be classified with the help of the protocol.
Across the world, treatment application in infancy varies from NDT to Vojta and from developmental interventions to stimulation programs. Thus, it would be interesting to know whether the protocol can be used also to evaluate physical therapy for infants in other countries. This study indicates that also within the application of a specific approach in 1 country heterogeneity is present.
Physical therapists frequently expressed surprise when they got feedback on the contents of their treatment sessions. They became aware of the discrepancy between what they believed that they had been doing and what they actually had been doing. This finding illustrates the difficulty of verbal communication. Success of verbal communication depends not only on the accuracy of the speaker or author and the intention of the listener or reader to interpret as precisely as possible but also on the body of knowledge and the verbal frame of reference present in sender and receiver. These frames of reference may differ more often than we realize, resulting in miscommunication. Our findings suggest that the application of video feedback may serve as an instrument to prevent this type of misunderstanding. Video recording thus may be a useful tool in the training of novice pediatric physical therapists.
This study indicates that the categories challenging the infant to produce motor behavior by himself or herself and facilitation techniques were applied most. Three types of motor behavior produced by the infant were distinguished: behavior that was initiated by the infant and behavior that was challenged by caregiver or therapist, who (1) let the infant try on her/his own or (2) took over control of the infant's movements by means of handling. We observed that self-initiated activity of the infant usually occurred when the therapist started to communicate with the parents, leaving the infant in a situation in which his or her capacities were not challenged. Another observation was that during motor skill practice the physical therapist usually challenged the infant to explore only 1 movement strategy. This means that during the physical therapy sessions of this study, little variation in motor behavior was practiced. This is surprising because current concepts of infant motor dysfunction suggest that varied practice is beneficial for motor development.18
The protocol pays specific attention to family involvement and educational actions. In the latter, we distinguished between educational actions regarding the child and educational actions regarding the caregiver, that is, caregiver training and caregiver coaching. It was remarkable that during treatment, little time was spent training the caregiver on how to apply treatment strategies during daily life activities and/or in the home environment. According to the traditional treatment approach, merely based on NDT, guiding and training of caregivers are the primary technique to achieve carryover from treatment to everyday life.10,12,15 In the functional approaches, caregivers are the key persons in the child's development and education.8,9,30,31 Thus, it seems that actual practice of the therapists involved in this study differed from current notions because caregivers mainly acted as a passive observer of treatment instead of being actively involved and instructed.
Future research should aim firstly at a replication of this study in infants at various ages, with various degrees of neurological dysfunction, and in various countries. A next step will be to relate the contents of infant physical therapy sessions to developmental outcome. This knowledge might shed light on the potentially effective and ineffective elements in infant physical therapy.
In conclusion, this study demonstrated that it is possible to assess PT actions during physical therapy for young infants at high risk for CP in a systematic, standardized, and reliable way despite the heterogeneity in the implementation of infant physical therapy. The study indicated that opening the black box of infant physical therapy has shown that in the Netherlands the application is very heterogeneous and varies between more traditional (NDT-like) to more functionally oriented treatment. Our study may be seen as a first step in the development of a tool to describe therapists' actions during physical therapy interventions. Objective knowledge on the contents of physical therapy sessions is an essential step toward evidence-based practice. The next steps to be taken are replication studies on reliability assessment, studies in different populations (eg, older children with CP or developmental coordination disorder) and in different countries, and studies connecting the contents of physical therapy sessions with developmental outcome.
1. Blauw-Hospers CH, Hadders-Algra M. Fact or fiction: a systematic review on the effects of early intervention on motor development. Dev Med Child Neurol
2. Butler C, Darrah J. Effects of neurodevelopmental treatment (NDT) for cerebral palsy: an AACPDM evidence report. Dev Med Child Neurol
3. Blauw-Hospers CH, de Graaf-Peters VB, Dirks T, et al. Indications that early intervention in infants at high risk for a developmental motor disorder may improve motor and cognitive development. Neurosci Biobehav Rev
4. Spittle AJ, Orton J, Doyle LW, et al. Early developmental intervention programs post hospital discharge to prevent motor and cognitive impairments in preterm infants. Cochrane Database Syst Rev
5. Chen H-T. Theory-Driven Evaluations
. Newbury Park, CA: Sage Publications; 1990.
6. Brazil K, Ozer E, Cloutier MM, et al. From theory to practice: improving the impact of health services research. BMC Health Serv Res
7. Aylward GP. Neurodevelopmental outcomes of infants born prematurely. J Dev Behav Pediatr
8. Ketelaar M, Vermeer A, Hart H, et al. Effects of a functional therapy program on motor abilities of children with cerebral palsy. Phys Ther
9. Ekström Ahl L, Johansson E, Granat T, et al. Functional therapy for children with cerebral palsy: an ecological approach. Dev Med Child Neurol
10. Bobath K, Bobath B. The neuro-developmental treatment. In: Scrutton D, ed. Management of the Motor Disorders of Children with Cerebral Palsy
. Oxford, UK: Clinics in Developmental Medicine 90, Spastics International Medical Publications; 1984:6–18.
11. Mayston MJ. People with cerebral palsy: effects of and perspectives for therapy. Neural Plast
12. Howle JM. Neuro-Developmental Treatment Approach: Theoretical Foundations and Principles of Clinical Practice
. Laguna Beach, CA: Neuro-Developmental Treatment Association; 2002.
13. Bly L. Baby Treatment Based on NDT Principles
. Austin, TX: PRO-Ed;1999.
14. Quinton MB, Nelson CA. Concepts & Guidelines for Baby Treatment
. Albuquerque, NM: Clinician's View; 2002.
15. Finnie NR. Handling the Young Child with Cerebral Palsy at Home
. Oxford, UK: Butterworth-Heinemann; 1996.
16. Resnick MB, Eyler FD, Nelson RM, et al. Developmental intervention for low birth weight infants: improved early developmental outcome. Pediatrics
17. Palmer FB, Shapiro BK, Wachtel RC, et al. The effects of physical therapy on cerebral palsy. A controlled trial in infants with spastic diplegia. N Engl J Med
18. Hadders-Algra M. The neuronal group selection theory: promising principles for understanding and treating developmental motor disorders. Dev Med Child Neurol
19. Damiano DL. Activity, activity, activity: rethinking our physical therapy approach to cerebral palsy. Phys Ther
20. Reynders K. [Kinderrevalidatie in pedagogisch perspectief] Educational Perspective on Child Rehabilitation
. Groningen, the Netherlands: Stichting Kinderstudies; 1992 (PhD thesis).
21. Einspieler C, Prechtl HFR, Bos AF, et al. Prechtl's Method on the Qualitative Assessment of General Movements in Preterm, Term and Young Infants. Clinics in Developmental Medicine No. 167
. London, UK: Mac Keith Press; 2004.
22. Hadders-Algra M. General movements: a window for early identification of children at high risk of developmental disorders. J Pediatr
23. Prechtl HFR. General movement assessment as a method of developmental neurology: new paradigms and their consequences. Dev Med Child Neurol
24. Prechtl HF, Einspieler C, Cioni G, et al. An early marker for neurological deficits after perinatal brain lesions. Lancet
25. Heineman K, Hadders-Algra M. Evaluation of neuromotor function in infancy—a systematic review of methods available. J Dev Behav Pediatr
26. Hadders-Algra M, Heineman KR, Bos AF, et al. Minor neurological dysfunction in infancy: strengths and limitations. Dev Med Child Neurol
. 2009 [Epub ahead of print].
27. Albrecht TL, Ruckdeschel JC, Ray FL III, et al. Portable, unobtrusive device for videorecording clinical interactions. Behav Res Methods
28. Portney LG, Watkins MP. Part IV: Data Analysis: Correlation. 2000. Foundations of Clinical Research. Applications to Practice
. 2nd ed. Upper Saddle River, NJ: Prentice Hall Health; 2000.
29. Bly L. A historical and current view of the basis of NDT. Pediatr Phys Ther
30. Law M, Darrah J, Pollock N, et al. Family-centred functional therapy for children with cerebral palsy: an emerging practice model. Phys Occup Ther Pediatr
31. Rosenbaum P, King S, Law M, et al. Family-centred service: a conceptual framework and research review. Phys Occup Ther Pediatr