OBJECTIVE: To summarize evidence on the rates of co-occurring impairments, diseases, and functional limitations with cerebral palsy into succinct clinical messages.
METHODS: A search was conducted of the databases PubMed, Medline, CINAHL, and PsycINFO, and the results were supplemented with hand searches. Two independent reviewers determined whether retrieved abstracts met the following inclusion criteria: human subjects; >90% were children or adults with cerebral palsy; published after 1999; and population-based data. Articles were appraised, analyzing design, participants, level of evidence, rates of impairments, and functional implications. Methodologic quality was rated by using a standardized checklist.
RESULTS: A total of 1366 papers were identified in the search; 82 were appraised and 30 were included in the meta-analyses. High-level evidence existed, as rated on the Oxford 2011 Levels of Evidence: 97% of prevalence studies were level 1. The data were of a moderate to high quality grade (with the exception of sleep disorders), allowing plain English clinical messages to be developed.
CONCLUSIONS: Among children with cerebral palsy, 3 in 4 were in pain; 1 in 2 had an intellectual disability; 1 in 3 could not walk; 1 in 3 had a hip displacement; 1 in 4 could not talk; 1 in 4 had epilepsy; 1 in 4 had a behavior disorder; 1 in 4 had bladder control problems; 1 in 5 had a sleep disorder; 1 in 5 dribbled; 1 in 10 were blind; 1 in 15 were tube-fed; and 1 in 25 were deaf.
- CI —
- confidence interval
- GMFCS —
- Gross Motor Function Classification System
- ID —
- intellectual disability
- SCPE —
- Surveillance of Cerebral Palsy in Europe
- UKCP —
- United Kingdom Cerebral Palsy register
Parents of children with cerebral palsy often ask pediatricians and allied health professionals, “will my child walk?,” “will my child talk?” and “will my child work and live independently?”1 In essence, parents want to know “how bad is it?” and what will their child’s future look like.1,2 The answer to the question depends on the severity of physical disability, the type of motor impairment, and the presence of comorbid conditions. Research exists to help clinicians identify impairments and predict future function, such as walking, in children with cerebral palsy. Yet parents report that they are rarely given prognostic information.3 Parents believe professionals withhold prognostic information in an attempt to protect them from bad news. Research, however, suggests that the absence of prognostic information makes it more difficult, not easier, for parents to cope.4 Dissatisfaction with delayed receipt of diagnostic information has been linked to higher rates of parental depression. In qualitative studies, parents advise professionals that they want and need prognostic information to assist them with planning services.3 In addition, parents recommend that medical information be presented in “parent-friendly” language to facilitate their understanding and acceptance of information.5
Cerebral palsy is the most common physical disability in childhood, occurring in 2 to 2.5/1000 births.6 Pediatricians and allied health professionals therefore need to have up-to-date prognostic information readily available to communicate with families at diagnosis and throughout the child’s life span to develop interventions. The definition of cerebral palsy has recently been expanded to include the impairments commonly associated with the condition because of their substantial impact on the child. “Cerebral palsy describes a group of permanent disorders of the development of movement and posture, causing activity limitation, that are attributed to non-progressive disturbances that occurred in the developing fetal or infant brain. The motor disorders of cerebral palsy are often accompanied by disturbances of sensation, perception, cognition, communication, and behaviour, by epilepsy, and by secondary musculoskeletal problems.”7 The inclusion of secondary impairments and functional limitations within the definition heightens the importance of understanding and communicating the impact of these co-occurring impairments, diseases, and functional limitations to parents to help predict outcomes. Furthermore, although the brain injury in cerebral palsy is nonprogressive, the co-occurring impairments, diseases, and functional limitations change over time, reducing function and quality of life.8 Definitive information about the real extent of co-occurring impairments, diseases, and functional limitations with cerebral palsy is essential for parents in choosing services and for providers in terms of pediatric resource allocation.4 Moreover, the psychological and physical health of parents of children with cerebral palsy is strongly influenced by the child’s behavior and the complexity of caregiving demands.9 Thus, the opportunity to introduce preventative mental and physical health measures for parents and children exists when professionals proactively identify and forecast the extent of a child’s disability.
The diagnosis of cerebral palsy is typically made in the toddler years, after metabolic and degenerative conditions have been ruled out.10 In the period of “wait and see” leading up to diagnosis, parents experience great distress.4 The description and ultimately the diagnosis of cerebral palsy can in some cases be made earlier; for example, if a preterm child has abnormal brain imaging coupled with abnormal motor signs on tools with good predictive psychometrics (such as General Movements or the Test of Infant Motor Performance).11 MRI is useful for identifying the presence and location of an injury in ∼89% of children with cerebral palsy.10 The location and type of brain injury generally correlate with subtype of cerebral palsy, which can provide physicians with some guidance in prognosticating future function.12 Despite this, significant limitations exist in the accuracy of MRI predicting the severity of cerebral palsy and long-term functional outcomes.13 In addition, cerebral palsy is an umbrella term for many different brain lesions (with the type and size of the lesion responsible for the different motor impairments and accompanying impairments), which explains why professionals encounter so many difficulties in giving an accurate prognosis to parents. Despite the lack of an evidence-based tool to prognosticate the severity of cerebral palsy early, many parents anecdotally report being told that their child will be profoundly disabled and will not walk, talk, or work. For this reason, prognostic evidence drawn from more sources than an abnormal MRI is needed to help accurately answer parents’ questions.
Because no guideline or systematic review about prognosis of cerebral palsy existed, we conjectured that this might explain why clinicians have not yet developed clear clinical messages for communicating to parents. The lack of synthesized information may compound professionals’ preference to avoid giving bad news. Recommendations for intervention might be based on specialists’ opinions about their own area of expertise based on experience, rather than on an overview of evidence from cerebral palsy population data. Given that interdisciplinary care is believed to be best practice, we considered the lack of synthesized clinical messages to be a major gap in cerebral palsy knowledge. Our answerable clinical questions therefore were: What are the rates of co-occurring impairments, diseases, and functional limitations in cerebral palsy? In addition, what is the prognosis of individuals with cerebral palsy? The objective of the current study was to systematically review the highest levels of evidence available relating to rates of co-occurring impairments, diseases, and functional limitations in cerebral palsy, presented in parent-friendly clinical prognostic messages. We hypothesized that it was possible to develop from the literature a series of succinct clinical messages in plain English for clinicians to communicate to parents of children with cerebral palsy, to promote understanding and to inform intervention planning.
We conducted this systematic review using a protocol based on recommendations for conducting systematic reviews from the Cochrane Collaboration and PRISMA statements and in accordance with the quality of reporting of meta-analysis of observational studies statements.14,15,16 We identified relevant articles by searching PubMed (1999–2011), Medline (1999–2011), CINAHL (1999–2011), and PsycINFO (1999–2011). Searches were supplemented by hand searching bibliographies of included articles and review articles. Relevant studies known to the investigators through previous research work were also included. The search of published studies was originally performed in January 2009, and an updated search was conducted in January 2011.
We searched electronic databases by using EBSCO host software including the following search terms: (i) cerebral palsy OR hemipleg* OR dipleg* OR quadripleg*; AND (ii) incidence OR prevalence OR prognosis OR rate OR proportion; AND (iii) behavior*/behavior*/emotion*/psycholog*/psychiatr*/autis* OR continen*/incontinen* OR drool*/saliva/sialorreah OR eat*/dysphagia/mealtime*/feed*/aspirat*/nutrition*/videofluoroscopy/video fluoroscopy/swallow*/deglutition OR epilepsy OR hearing OR hip/disloc*/displac*/scoliosis/lordosis/kyphosis/windsw* OR intellectual/intelligen*/mental* retard*/learning disab*/cognit* OR pain OR sleep OR communication/nonverbal communication/speech/intelligibility/articulation/dysarthria/anarthria/apraxia/dyspraxia/language/augmentative communication/alternative communication/AAC/speech generating device/SGD/communication methods/communication aid* OR walk*/motor/ambulat* OR vision/visual/nystagmus/strabismus/hemianopia/blind*; AND (iii) population OR registry OR register. The list of associated impairments searched were generated from the definition of cerebral palsy and seminal texts,6 and assigned plain English terms for parents.
Studies published in English regarding the prevalence of co-occurring impairments, diseases, and functional limitations fulfilling the following criteria were included:
Type of study: Observational studies were specifically sought, because it is neither logical nor ethical to prospectively randomize individuals to risk association studies. To improve the quality of the review and minimize bias, studies using a population sample (ie, a cerebral palsy register) were preferentially sought. There was a scarcity of research about some functional limitations (eg, behavior, bladder control, dribbling, pain, sleep), and more generous inclusion criteria were required to achieve the study aim for these topics (eg, nonpopulation studies). The search strategy only included studies with greater levels of bias when no higher-quality observational studies were available.
Type of outcome: studies that involved ascertaining the rates of co-occurring impairments, diseases, and functional limitations for children or adults with cerebral palsy.
Types of subjects: studies that explicitly involved human subjects in which >90% of the participants were children or adults with cerebral palsy.
Studies were excluded from the systematic review if: (1) they were nonpopulation\-based surveys (unless population evidence was unavailable); (2) a second publication of the same study was published, which found the same results; (3) they were not available in English; and (4) they were published before 1999, given the marked advances in obstetric and neonatal care in the last 10 years.
A data abstraction sheet was developed based on the Cochrane recommendations. Two independent assessors determined eligibility of the abstracts identified from searches for further review. Abstracts were retained for full review if they met the inclusion criteria or if more information was required from the full text to ascertain eligibility. The same 2 reviewers then reviewed full-text versions of all retained articles and all additional articles identified by hand searching. Full-text articles were retained if they met inclusion criteria; 100% agreement was reached on inclusion and exclusion assignment for the full-text articles. Data extracted from included studies included: authors and date of study; number of participants; participants’ diagnosis; study design; and description of findings. The data extracted from each included study are summarized in Table 1. All the data required to answer the study questions were published within the papers, so no contact with authors was necessary.
Evidence experts acknowledge that conforming to gold standard systematic review conventions is difficult when critiquing etiologic, risk factor, and prognostic studies.17,18 This difference is because evidence hierarchies heavily favor treatment studies.18 Our review of the cerebral palsy literature confirmed this problem: a lack of agreement existed in literature regarding a preferred quality rating scale for observational studies, and studies had for the most part examined only single groups, which therefore precluded standard meta-analyses.17 In the absence of an endorsed systematic review guideline for epidemiologic studies, the following course of action was chosen: First, we rated the level of evidence on the Oxford 2011 Levels of Evidence as per convention.19 Second, we assessed quality by using the standardized checklist developed by Boyle.20 Third, the overarching body of prevalence evidence was rated by using the GRADE system.21
The Human Research Ethics Committee waived the need for ethical approval because the study did not involve any contact with humans.
Our intention was to conduct a meta-analysis if the retrieved studies were of sufficient clinical and statistical homogeneity. We encountered 2 problems. First, the prevalence studies predominantly only studied 1 group (ie, the cerebral palsy population). Thus, conducting 2-group prevalence meta-analyses by using standard Mantel-Haenszel techniques was not possible.14 For the prevalence meta-analysis, we therefore needed to use simple descriptive statistics, calculating population prevalence mean impairment rates with 95% confidence intervals (CIs). To increase the rigor of these descriptive analyses, studies had to meet these additional stringent criteria for inclusion within the meta-analyses: (1) study sample was >80% of population (as per public health conventions) or the total sample size needed to be >500 cases of cerebral palsy; (2) study sample included all subtypes of cerebral palsy so as not to skew the estimates; and (3) case data were not duplicated in other publications. Studies meeting these extra criteria are denoted in Table 1. When studies included near-duplicate data, older and smaller duplicates were preferentially excluded to improve rigor, unless the newest study had collapsed data, in which case the next best available study was included. Review articles were not included in the meta-analysis, as per convention. The second problem was that although we originally intended to conduct the meta-analysis by using Gross Motor Function Classification System (GMFCS) levels, there were insufficient numbers of studies reporting GMFCS data for this to be possible.
A total of 1366 citations were identified in our search; 82 articles met the inclusion criteria for full appraisal, but only 30 studies met the additional stringent inclusion criteria for meta-analyses (Fig 1).
Levels of Evidence
Evidence from these studies was generally very high when rated by using the Oxford 2011 Levels of Evidence; 97% of the prevalence articles included in the meta-analysis were rated as level 1 evidence (“most relevant local and current random sample survey or census”) because they were cerebral palsy register studies capturing data from >80% of the population. Only 1 lower-level evidence (level 5) prevalence article, on the topic of sleep difficulties, was included in the meta-analyses because no higher-level evidence could be found in the literature.
Rating of Study Quality
Prevalence of Co-occurring Impairments, Diseases, and Functional Limitations
Average rates of impairments, diseases, and functional limitations associated with cerebral palsy were calculated by aggregating the findings of the included prevalence studies to identify a mean rate and a 95% CI. The findings were: (1) behavior (818 cases for analysis): 26% (95% CI: 24–28) had abnormal behavior (the rate and CI were from Parkes et al22 because new CIs were unable to be calculated because only 1 study met inclusion criteria); (2) bladder and bowel control (601 cases for analysis): 24% had bladder control problems (95% CI unable to be calculated because only 1 study met inclusion criteria); (3) dribbling (1119 cases for analysis): 22% had excessive drooling (95% CI unable to be calculated because only 1 study met inclusion criteria); (4) eating (1299 cases for analysis): 6% (95% CI: 3–9) were tube-fed; (5) epilepsy (12 140 cases for analysis): 35% (95% CI: 26–42) had epilepsy at some point, and 24% (95% CI: 17–29) had active epilepsy; (6) hearing (9492 cases for analysis): 4% (95% CI: 2–6) had a severe hearing impairment or were deaf; (7) hips and spine (632 cases for analysis): 28% (95% CI: 21–34) had displaced hips (ie, migration >30%) and 7.5% (95% CI: 7–8) had dislocated hips if they had not received hip surveillance, but this was reduced to 0% dislocation with surveillance; (8) intellectual function (12 053 cases for analysis): 49% had an intellectual disability (ID) (IQ <70) (95% CI: 34–64) and 28% had a severe ID (IQ <50) (95% CI: 21–35); (9) pain (1224 cases for analysis): 75% were in pain (95% CI: 72–78); (10) sleep (173 cases for analysis): 23% had a pathologic sleep disorder (95% CI unable to be calculated because only 1 study was included); (11) talking (4872 cases for analysis): 23% (95% CI: 19–27) were nonverbal; (12) walking (21 350 cases for analysis): 28% (95% CI: 16–40) could not walk, 16% walked with assistance (95% CI: 8–24), and 56% walked independently (95% CI: 32–80), with GMFCS (2924 cases for analysis) level I at 36% (95% CI: 26–46), level II at 22% (95% CI: 9–35), level III at 11% (95% CI: 8–14), level IV at 14% (95% CI: 8–20), and level V at 17% (95% CI: 13–21); and (13) vision (19 076 cases for analysis): 11% (95% CI: 5–17) were functionally blind.
The data regarding average rates of impairments were combined with prognostic data from the included studies regarding the likelihood of unwanted events occurring in the future. From these data, we developed a series of simple, plain English clinical messages that could be communicated to people with cerebral palsy and their families about the prognosis of the condition (Table 3).
In this systematic literature review regarding the rates of co-occurring impairments, diseases, and functional limitations with cerebral palsy, we found that large volumes of either high-quality or moderate-quality evidence existed to provide guidance to parents about “how bad is it?” Included prevalence studies were nearly all rated as level 1 evidence on the Oxford Scale because they were cerebral palsy population registry studies. Furthermore, the methodologic quality of almost all studies was very high with low levels of bias, providing high degree of certainty about the plain English messages that were developed from the meta-analysis. The exception were data regarding behavior, bladder control, dribbling, pain, and sleep, in which more research, including that of a higher quality, was needed.
From the evidence appraised, we found that among children who have cerebral palsy: 3 in 4 were in pain; 1 in 2 had an ID; 1 in 3 could not walk; 1 in 3 had a hip displacement; 1 in 4 could not talk; 1 in 4 had epilepsy; 1 in 4 had a behavior disorder; 1 in 4 had bladder control problems; 1 in 5 had a sleep disorder; 1 in 5 dribbled; 1 in 10 were blind; 1 in 15 were tube-fed; and 1 in 25 were deaf. Rates of co-occurring impairments, diseases, and functional limitations were strongly linked to the severity of the motor impairment, with the exception of pain and behavior disorders. Pain was very likely to be present at all levels of physical disability. Interestingly, behavior disorders were more common in children with milder levels of physical disability. In general, findings could potentially be explained to parents by saying, “All children with cerebral palsy will have physical challenges. The bigger the child’s brain injury, the more likely the child is to have other co-occurring impairments, diseases, and functional limitations accompanying the physical disability, except for pain and behavior, which are common regardless of the level of physical disability.” If these objective and simple messages are given to parents, this information may help alleviate the stress parents experience while trying to envisage the future at the time of diagnosis4 and in planning realistic and achievable services.3
The clinical messages we developed are consistent with cerebral palsy messages published elsewhere in the literature based on smaller, single-country, nonaggregated samples; for example, the Swedish register,23 the Danish register,2 and the Canadian register.12 However, to the best of our knowledge, this is the first review to aggregate data from all the published international registries on a multitude of clinical problems associated with cerebral palsy to produce a suite of parent-friendly clinical messages.
Parents generally maintain a remarkably optimistic view of the future despite receiving bad news about their child’s prognosis.24 Research suggests that parental recall about diagnostic information shared does not match what professionals claim to have communicated. Parents are thought to actively block the recall of bad news as a coping strategy. It is therefore paramount that professionals understand that when communicating bad news with families, “truth disclosure is a process, not an event.”25 The literature provides guidance about how professionals might communicate these cerebral palsy prognostic clinical messages. Parents recommend and prefer the following communication strategies when receiving bad news: (1) use of an honest, upfront, specific, and transparent communication style to enable parents to seek their right supports after receiving the news; (2) discussion about the child’s strengths to convey hope and reframe the future; (3) person-centered and respectful treatment of the child, so that the news feels personalized; (4) provision of a list of frequently asked questions to help parents prepare their own questions; (5) delivery of news to both parents simultaneously, or to one parent accompanied by a support person, to assist with information recall; (6) provision of a follow-up interview to assist with information recall, understanding, and acceptance; (7) provision of key information in writing; and (8) involvement of an advocate/keyworker/case manager to help with planning after the receipt of news.5,24,26,27 In light of this parental counsel about preference for a strengths-based approach, professionals could positively reinforce that the level of physical disability associated with cerebral palsy does not predetermine a child’s “happiness” and quality of life (high-quality GRADE).28,29
Our review was limited to evaluation of rates of co-occurring impairments, diseases, and functional limitations in studies published in English. Nevertheless, in most topic areas, publications were sourced from a wide range of developed countries across Europe, the United Kingdom, North America, and Australia using cerebral palsy register population data. However, accurate cerebral palsy population data from developing countries are difficult to ascertain, and the issues may not be comparable. In the context of developing countries, public health strategies such as rubella vaccines and rhesis management are often unavailable, thus changing the profile of cerebral palsy. Publication bias is another potential study limitation. Studies may have evaluated the interaction of impairments and prognostic outcomes but published only those that were statistically significant. For vision and hearing impairments, there was a lack of consistency in coding of impairments between authors and registries, as some authors used the term “some impairment.” To improve the rigor of the estimates calculated in the meta-analysis, only severe hearing and severe vision impairments were included because these impairments are formally tested. Rates were calculated with a CI, and the body of supporting evidence was graded as moderate to high, indicating that more research is likely to change our understanding and the precision of these estimates.
Implications for Research
There is a vital need for population studies that more precisely investigate the rates of behavioral problems, bladder control problems, dribbling, feeding problems, pain, and sleep disorders in cerebral palsy. These 6 problem areas have not received sufficient attention in the literature, and more data of a higher quality are likely to change and enhance our understanding of these problems. Moreover, more data would inform the design of future intervention studies, which may create the possibility for improved outcomes. The findings of this study suggest an underresearched link between chronic pain, sleep disorders, and behavioral problems. The relationship between pain, sleep, and behavior warrants urgent attention, as effective evidence-based intervention strategies exist in other diagnostic groups that may offer great promise to individuals with cerebral palsy.
Future research and cerebral palsy register data sets collecting information on vision and hearing impairments associated with cerebral palsy should use agreed data dictionary terminology to code impairments and therefore improve data quality and ease of interpretation.
There is high-quality grade evidence that among children with cerebral palsy: 1 in 3 cannot walk; 1 in 4 cannot talk; 1 in 4 had epilepsy; and 1 in 25 were deaf. There is moderate-quality evidence that 3 in 4 were in pain; 1 in 2 had an ID; 1 in 3 had a hip displacement; 1 in 4 had a behavior disorder; 1 in 4 had bladder control problems; 1 in 5 dribbled; 1 in 10 were blind; 1 in 15 were tube-fed. There is low-quality evidence that 1 in 5 had a sleep disorder. Children and adults unable to walk are more likely to experience these accompanying impairments. The risk for pain and behavioral problems occurs equally at all levels of physical disability. There is insufficient evidence to be certain about the rates of sleep disorders, and more research is warranted.
- Accepted June 19, 2012.
- Address correspondence to Iona Novak, PhD, MSc (Hons), BAppSc (OT), Head of Research, Cerebral Palsy Alliance Research Institute, PO Box 184, Brookvale NSW 2100, Australia. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- Baird G,
- McConachie H,
- Scrutton D
- ↵Stanley F, Blair E, Alberman E. Cerebral Palsies: Epidemiology and Causal Pathways. London: Mac Keith. Clin Dev Med. 2000
- McAdams RM,
- Juul SE
- Higgins JPT,
- Green S
- Dickersin K
- Glasziou P,
- Vandenbroucke JP,
- Chalmers I
- ↵OCEBM Table of Evidence Working Group. The Oxford 2011 Table of Evidence: Oxford Centre for Evidence-Based Medicine
- Boyle M
- Guyatt GH,
- Oxman AD,
- Vist GE,
- et al.,
- GRADE Working Group
- Arnaud C,
- White-Koning M,
- Michelsen SI,
- et al
- Surman G,
- Bonellie S,
- Chalmers J,
- et al
- Australian Cerebral Palsy Register Group (ACPR). Report of the Australian Cerebral Palsy Register, Birth Years 1993–2003. Sydney Australia: Cerebral Palsy Alliance; 2009
- Shevell MI,
- Dagenais L,
- Hall N,
- REPACQ Consortium
- Himmelmann K,
- McManus V,
- Hagberg G,
- Uvebrant P,
- Krägeloh-Mann I,
- Cans C,
- SCPE collaboration
- Jarvis S,
- Glinianaia SV,
- Arnaud C,
- et al.,
- SCPE collaboration of European Cerebral Palsy Registers
- Saunders KJ,
- Little JA,
- McClelland JF,
- Jackson AJ
- Andersen G,
- Mjoen T,
- Vik T
- Hanna SE,
- Bartlett DJ,
- Rivard LM,
- Russell DJ
- Wu YW,
- Day SM,
- Strauss DJ,
- Shavelle RM
- Zafeiriou DI,
- Kontopoulos EE,
- Tsikoulas I
- Hägglund G,
- Andersson S,
- Düppe H,
- Lauge-Pedersen H,
- Nordmark E,
- Westbom L
- Humphreys P,
- Deonandan R,
- Whiting S,
- et al
- McClelland JF,
- Parkes J,
- Hill N,
- Jackson AJ,
- Saunders KJ
- Copyright © 2012 by the American Academy of Pediatrics