All the various treatment systems claim good results. It is difficult to decide which system is superior, whether on theoretical grounds or on the basis of a scientific study. Clinical experience of many therapists, as well as my own, has not confirmed the superiority of any one approach.

As therapists ‘want to know what is the best thing to do for children’, there has been a focus on which therapy approach is superior. Perhaps this is not a worthy aim to pursue on theoretical grounds or research studies.

Every therapist wishes to understand ‘why we do what we do’ and, unfortunately, may accept a therapy system because it offers a ready explanation. However, there is no all-encompassing theory which fully explains all the abnormal motor behaviour presented by people with the different cerebral palsies. In addition, theories may not adequately explain the effects of various treatment systems or, in some cases, of specific procedures. Theories also do not fully clarify mechanisms underlying motor behaviour from infancy to adulthood. Each therapy approach is based on scientific evidence available at the time, and the neurofacilita-tion systems are based on a number of controversial neurophysiological hypotheses. There are currently newer theories on motor learning and motor control. However, controversies exist, so current neu-roscience and behavioural psychology do not favour a single model for motor behaviour. Therapists continue to work with an underlying theoretical framework with some scientific evidence but also with assumptions about motor control, motor development or motor learning. Increasing scientific understanding of brain function and motor behaviour will contribute to therapy and its knowledge base. Therapists will need to judge the relevance of such studies and advances and apply them to clinical work. It is still unwise to be dogmatic about a theoretical framework or about procedures which arise from that. Therapy techniques cannot fully rest on different and more current theories underlying motor control or motor learning. We still have to learn to live with these doubts.

Although the therapist should continue to ask herself why she is using a particular method, this enquiry should focus more on the careful observations of motor behaviour and any changes in behaviour after treatment procedures. Increasing clinical studies and research, together with clinical experience, will offer scientific evidence underpinning treatment procedures. Therapists need to draw on those studies which are relevant to their particular patients/clients.

Contemporary theories are discussed in Chapter 3.

These are fraught with many problems and to date no study to compare the value of different treatment systems has convincingly dealt with all the problems. Firstly, the results of treatment are influenced by the methods dictated not only by the concept of an approach but by the severity of motor function and impairments, age and sex of an individual. Secondly, results are influenced by possible associated impairments and disabilities of vision, hearing, communication, perception, cognition as well as by the presence of epilepsy and poor health. A child’s personality and ‘drive’ as well as his home background contribute to the results of therapy. One must also recognise that the therapist’s enthusiasm, personality and her abilities to make positive professional relationships as well as her technical skill may have a strong effect on the results of treatment with any method.

There are other problems. The results of a scientific study would have to be obtained over a long period of time as progress is slow. At least a 6-month follow-up is now generally recommended. Crothers pointed out many years ago that one would really need a follow-up to adulthood to establish the ultimate effects of treatment methods in childhood (Crothers & Paine 1959; Levitt 1962; Paine 1962). However, this view needs to be reassessed with more current information on the deterioration related to specific physiological and psychological factors in older persons with cerebral palsy.

The research studies that have been carried out on therapy interventions have been reviewed by Parette and Hourcade (1984) for the period 1952–1982, by Tirosh and Rabino (1989) for the period 1973–1988, by Siebes et al. (2002) from 1990 to 2001 and by Anttila et al. (2008) from 1990 to 2007. They all found in their reviews that the research designs were not rigorous enough and they discuss the problems facing researchers. Siebes et al. (2002) point out that although the methodology in research studies had clearly improved, this did not lead ‘to a substantial improvement in the scientific foundation’ of the motor interventions for children with or at risk for cerebral palsy.

Tirosh and Rabino (1989) suggest that a much larger number of subjects in a multicentre study should iron out many variables and that more reliable data might be obtained. They pointed out the importance of psychosocial influences. Bower and McLellan (1992) found pitfalls in eight major studies, most of which are included in Tirosh and Ra-bino’s review. A review of studies specifically on the neurodevelopmental treatment (NDT/Bobath) approach by Butler and Darrah (2001), such as Siebes et al. (2002), classifies levels of evidence from I down to V. Both these reviews find that two-thirds of the studies are below Levels I and II. Furthermore, each of these reviews shows that the more scientific the study, the fewer the number of statistically significant results due to therapy.

Palmer et al. (1988) found NDT less effective than a global developmental programme. The NDT review by Butler and Darrah, approved by the American Academy for Cerebral Palsy and Developmental Medicine (AACPDM) Treatment Outcomes Committee Review Panel, has been mentioned above. They state that the ‘preponderance of results did not confer any advantage to NDT over the alternatives to which it was compared’. Brown and Burns (2001) considered 147 articles of which only 17 studies met scientific criteria. These studies found insufficient evidence for NDT and the results were largely inconclusive. There are studies in the past which claim to support NDT, but their design is inadequate. For example, the study by Jonsdottir et al. (1997) did show that ‘NDT seemed more effective than practice for improvement of postural alignment during reaching’, but like other studies it only used a small number of subjects and for a short time of 5 days, without follow-up. In early treatment, Blauw-Hospers and Hadders-Algra (2005) found that specific or general developmental programmes had a positive effect whereas NDT was not beneficial. Of 17 papers, 12 were of high methodology (research levels I and II) in their study of early treatment.

The Doman-Delacato system was found un-proven in results by Sparrow and Zigler (1978) and by Cummins (1988). Vojta’s approach (see Vojta (1989) or von Aufschnaiter (1992) for recent account) received criticism by Jones (1975) and Forss-berg and Hirschfeld (1992). Vojta’s list of postural reflexes is not accepted by Norén and Franzén (1982) as a reliable measure of diagnosis and results. The results of Vojta’s very early therapy for babies at risk as well as any other early intervention such as that of Kong (1987) and Katona (1989) cannot show that it was the intervention that obtained results and not the fact that babies ‘at risk’ might have become normal anyway. Nelson and Ellenberg (1982) and Touwen (1987) point out how unreliable early diagnoses can be. Nelson and Ellenberg found in their large sample of infants suspected of cerebral palsy that there was a high rate of these infants becoming normal. However, neonatal studies by paediatric therapists continue and show positive value for intervention.

Carlsen (1975) showed improvement using methods from Rood (1962), sensory integration, pro-prioceptive neuromuscular facilitation and NDT compared to a functional occupational therapy approach. She had a small sample of 12 children.

Conductive education

Bairstow et al. (1991, 1993) carried out studies on conductive education (CE) compared with a neurophysiological and developmental approach in a selected group necessary for CE. No difference was found between these two groups. However, the assessors were aware which children were in the intervention group and which were in the non-intervention (control) group. Reddihough et al. (1998) carried out a comparative study of traditional therapy and CE which showed little difference in outcomes, provided equal intensity of therapy sessions were used. This is not always so with traditional therapy. Johanna Darrah with three physiotherapists (Darrah et al. 2004) carried out a thorough review of evidence for CE, approved by the AACPDM. Fifteen studies met the inclusion criteria. The studies included many aspects such as motor function, daily living activities, communication, various educational aspects, social skills and parent satisfaction. Only four studies are rated I and II, and there were many difficulties in obtaining information about the subjects’ specific activities and the heterogeneity in various studies. The result is that there is no conclusive evidence in support or against CE as an intervention strategy. The focus of CE on daily life function and education may fit the needs of many families but cost, time, availability and effects on family dynamics need to be considered in view of current lack of strong evidence. Further research is recommended.

Odman and Oberg (2005) have carried out studies on Move&Walk, an intensive Swedish CE programme with conductors educated in Hungary, compared with an intensive rehabilitation programme with an eclectic approach called Lemo with two physiotherapists and a special educational needs teacher. Children were aged 3–16 years. Both programmes emphasised parents’ and assistants’ (carers’) participation. No major difference in function was shown between the two training programmes. Odman et al. (2007) explored parents’ perceptions in Move&Walk and Lemo, finding that high-level service quality and achieved expectations influenced parents more than their perceived functional improvements of their children. No difference was present in their perceived functional outcomes. Most parents perceived high-level service quality. The parents in the Lemo approach perceived better service quality than those in the Move&Walk and were more involved in discussions about expectations and knowledge exchange.

Single-case studies to evaluate specific treatment procedures together with more sensitive and more specific measures for cerebral palsy are suggested in a review of research by Siebes et al. (2002). Single-case studies have increased and are useful for evaluation of procedures such as training methods for strength, for constraint-induced function in hemi-plegia, methods for postural control with arm reach (balance), use of below-knee plaster casts (inhibitory casts), ankle-foot orthoses and similar procedures. For example, the research study of Damiano et al. (1995a) evaluated the effects of specific strengthening methods of quadriceps and showed that these procedures improved crouch gait.

Single-case studies, carefully designed, with an appropriate number of subjects offer useful results for research on therapy. Research designs are discussed below. Single-case study design is discussed by Kazdin (1982), Edwards et al. (1990) and Rid-doch and Lennon (1991).

A series of research studies in cerebral palsy have been carried out by E. Bower and D.L. McLellan, investigating the effect of intensive physiotherapy with specific measurable goals. All the studies use the Gross Motor Function Measure (GMFM; Russell et al. 1989, 2002). The first pilot study (Bower & McLellan 1992) used a controlled series of single-case studies with only seven children. The trend was that increased intensity for 3 weeks with goal setting generally improved the rate of progress as compared to routine physiotherapy. The intensive treatment was carried out by the researcher.

In a subsequent randomized controlled study of 44 children (Bower et al. 1996), the subjects were assigned to four different groups of 11, with careful checks being made that each group was similar. Intensive physiotherapy was given for 2 weeks to two of the groups, one with general aims and the other with specific goals for physiotherapy. The other two groups received routine physiotherapy for 2 weeks, again one with general aims and the other with specific goals. The study showed that intensive therapy together with specific goals can accelerate the acquisition of motor function. There was no follow-up to see whether these gains were subsequently maintained.

A further study of 56 children (Bower et al. 2001) used four groups as before. However, the period of routine or intensive therapy lasted 6 months and, unlike the 1996 study, there was a period of 6-month routine therapy (about half an hour a week) with aims (baseline observation period) before the treatment period and a similar (follow-up) period afterwards. In the baseline period, there was some progress which accelerated with intensive therapy (average 3.5 hours a week). The mean total score on the GMFM improved by 5.9 percentage points on intensive therapy, whereas by only 3.1 percentage points on routine therapy. There was no difference as to whether aims or goals were used in the intensive or routine treatment period. The follow-up 6 months later showed that the average child did not maintain the lead gained during the intensive therapy. However, despite this obvious difference, Bower reports that the lead during the treatment period was not statistically significant.

Studies of rhizotomy by Steinbok et al. (1997a), McLaughlin et al. (1998) and Wright et al. (1998) provide data on rates of progress on the GMFM during intensive physiotherapy alone. These show a range of average gains of 4.2–5.2 percentage points over a period of 9–12 months. This is less than the gain of 5.9 points from 6-month intensive therapy found by Bower et al. (2001).

Today, what evidence of the evaluation of therapy is available to the clinician?