SCPE-NET Project EU EAHC / 35

Transcription

1 Work package Task WP5 Variations in access to health care and in outcome of health care T5.3 Variations in outcome of health care Authors Addresses External evaluator Peter B Sullivan Guro L. Andersen, Sandra J. Hollung, Torstein Vik EAHC and SCPE-NET associated and collaborating partners Document Date November st draft March 2012 Evaluation date September 2012 Document Title Document Type Communication D20 - Variations in general outcomes of health care for children with cerebral palsy across Europe: growth and nutrition. Deliverable and postal mail SCPE-NET Project EU EAHC scpe-net@h12o.es 0 / 35

2 Report Task 5.3 Task coordinator: (NO) CPRN Variations in general outcomes of health care for children with cerebral palsy across Europe: growth and nutrition. [variations in outcome of health care] Table of Contents Background and description:... 2 Objectives:... 3 Methods:... 3 Workshops... 3 Surveys Survey to SCPE members: Survey to health professionals/clinical networks... 6 Results... 8 Discussion... 9 Appendix I: Accepted article: Gastrostomy tube feeding of children with cerebral palsy: variation across six European countries Appendix II: Accepted abstract: How is growth and nutritional status assessed in children with cerebral palsy (CP)? A cross-sectional survey in 17 European countries Appendix III: Data collection document... 26

3 Background and description: The aim of this task is to document variations in the general outcome of health care. Our emphasis is on growth and nutritional outcomes of children with cerebral palsy (CP), particularly on the quality of nutritional rehabilitation across Europe. Children and adults with CP are at increased risk of suffering from both under/malnutrition and poor growth. Studies have shown that better nutritional care leads to improved growth and health status in children with severe CP. Therefore, our hypothesis is that the prevalence and recording of nutritional problems among children with CP vary significantly within the different countries in Europe. This task followed the same methodological plan as Task 5.2 and, therefore, joint workshops were held to optimize resources. At the start of the project a detailed study protocol was drafted which describes the specific criteria for participation and methods and agenda for workshops. This protocol is included in the Deliverable D10 (M12). The results and analysis of the surveys to CP registers and clinical networks is included in Deliverable 20 (M32). Two types of surveys were completed: one to all SCPE partners in order to gather which data variables can be collected for the study, and one to gather information on how growth and nutrition of children with CP in Europe is assessed. The study population is children with all subtypes and severities of CP across Europe Survey was performed at the CP register level (this first survey was in cooperation with Task 5.2): Initially SCPE partners were asked to answer a questionnaire about if their register can provide the following data on children with CP at GMFCS levels IV-V, born in : Hip luxation Intrathecal baclofen (ITB) Growth (height and/or weight) Feeding difficulties (chewing, swallowing problems/gastrostomy) Once we ascertained which centers could provide data, we sent out a more detailed questionnaire in order to be able to choose which variables would be possible to obtain and use in the study Survey was performed at the Clinical network level: Key informants among health professionals, particularly neuropaediatricians, were identified by a research assistant. The questionnaire included more detailed information such as standardized procedures for the assessment of growth and nutritional status, presence of nutritional teams and use of special equipment for assessing nutritional status To document variations in outcome of health care across Europe, analysis was performed on both surveys and collected data. (Appendix I and II) These documents include recommendations for good clinical practice with focus on nutritional rehabilitation. 2

4 Objectives: The overall aim of the work package and this task is to promote the best practice in health care for children with CP and to contribute to the reduction of inequalities in care between countries. The specific aim of this task is to document variations in health care of growth and nutrition outcomes. Methods: The task followed in accordance with the milestones as presented in the Grant Agreement. Date: Task: Milestone: M2 T5.2/3 Workshop at Plenary meeting P1 in Ljubljana, Slovenia October 26-28, 2009 Task 5.3 presentation made by (NO) CPRN Leader Guro L. Andersen M9 T5.1/2/3 Workshop 5_1: Garder Conference Center, Oslo, Norway, May 31 June 01, 2010 Study protocols were developed and discussions. M13 T5.3/ Survey: first survey sent to all SCPE members M14 T5.2/3 Workshop at Plenary meeting P2 in Riga, Latvia October 7-9, 2010 Task 5.3 presentation made by (NO) CPRN Leader Guro L. Andersen M15 T Survey: second sent to those SCPE members who responded positively to the first survey M18 T5.3 Data submission M21 T5.1/2/3 Workshop 5_2: NTNU, Trondheim, Norway, June 15, 2011 Data presentations and discussions. Development of Survey. M21 T5.3 Growth and Nutrition analysis and article under revision for re-submission to DMCN, Appendix I M25 T5.2/3 Plenary meeting P3 in Pecs, Hungary September 16-18, 2011 Task 5.3 presentation made by (NO) CPRN Leader Guro L. Andersen M27 T Survey: sent to all SCPE centres and other contacts outside of the SCPE M29 T Survey: data analysis and report. Submitted and accepted as Scientific Poster to AACPDM (waiting for results of ICPC acceptance) Appendix II M34 Plenary Meeting P4: Madrid, Spain June 25-27, 2012 Results will be reported. Workshops Workshop 5_1: Garder Conference Center, Oslo Norway, May 31 June 01, 2010 Participants (SCPE-NET Budget): Guro L. Andersen CPRN (NO) 3

5 Torstein Vik, CPRN (NO) Kate Himmelmann, DSBUS (SE) Magnus Påhlman, DSBUS (SE) Jennifer Kurinczuk, 4Child, Oxford (UK) Participants (not paid by SCPE-NET): Prof. Richard Stevenson, Kluges Institute, Virginia, USA Solveig Sigurdardottir, Iceland CP Register Ane Kristine Finbråthen, NTNU Magnus Dahlseng, NTNU Areej Elkamil, NTNU The aim of the workshop was to decide if: a) It is possible to perform the study on access in health care and outcomes in the SCPE registries as planned related to 1. Hip luxation and/or 2. Growth and nutrition. b) It is possible to make a draft of a study protocol: What are the outcome measures? and Where in Europe would it be possible to perform the study/ies? c) It is not possible to perform it as planned, how could it be performed? Conclusion: We concluded that it is possible to perform the two studies A draft for a study protocol for both task 5.2 and 5.3 were made for distribution and comments from the non-participating task partners Some questions regarding the budget was forwarded to Javier de La Cruz. The study population is children with all subtypes and severities of CP across Europe. At present 11,000 children are recorded in the SCPE database. However, these children will not be directly involved in the study. Instead, we will use this established European network to complete the study questions in two parts. Workshop 5_2: NTNU, Trondheim, Norway, June 15, 2011 Participants (SCPE-NET Budget): Guro L. Andersen, CPRN (NO) Torstein Vik, CPRN (NO) Sandra Hollung, CPRN (NO) Catherine Arnaud INSERM U558 UPS (FR) Magnus Påhlman, DSBUS (SE) Karen Horridge, NHCT (UK) Magnus Odin Dahlseng (NO) The aim of the workshop: a) Presentation of data: growth and nutrition results and discussion 4

6 b) Work still to be done, dissemination c) Reports to the EU Conclusion: Agreement on continuing working with the data and prepare a paper for publication. Some clarification on data validity needed from each center A second survey to health professionals/clinical networks is needed to gain information on health care provision Surveys Survey to SCPE members: A survey was first sent to all SCPE partners to identify if and what data were available on growth and nutrition. The survey was internet based, and sent in May A second questionnaire was sent in November 2010 to only those SCPE centers who answered yes, that they have available data on the first survey. The following questions were included on the second survey in order to ascertain which specific variables to be gathered: 1. Are these variables on growth and nutrition recorded in your register? a. Weight b. Height c. Head circumference d. Arm circumference e. Bio-impedance f. Skin-fold thickness g. Segmental length 2. Can your register provide anonymized individual data on children with CP born ? a. Gestational age b. Birth weight c. Sex d. Year of birth e. CP subtype f. Severity GMFCS or walkers/non-walkers g. ICD10codes: i. associated syndromes ii. congenital anomalies iii. or any other severe chronic disease(s) h. Presence of gastrostomy i. Age of placement of gastrostomy j. Feeding and/or swallowing difficulties k. Mother/father level of education l. Mother/father profession m. Family income n. Associated impairments: i. Vision ii. Hearing 5

7 iii. Epilepsy iv. Cognition o. Mortality p. Age at death After reviewing the answers, the registers with available data were asked to submit their data for children born with a deadline of February 28, In Appendix III you will find the SCPE WP5.3 Survey to Growth and Nutrition Data Collection document Survey to health professionals/clinical networks A survey to clinicians at departments of pediatrics, neuropediatrics and/or habilitation centers in Europe was performed to gather information on health care provision. The clinicians were contacted through the SCPE-NET and various European pediatric organizations. The survey was internet based and sent in October-December The following questions were included: 1. How do children and youths with CP - who can stand - have their height/length measured in your clinic? Standardized stadiometer, which is regularly calibrated Standardized stadiometer, not regularly calibrated Non-standardized equipment Tape measure Not at all Other (please specify): 2. How do children and youths with CP - who cannot stand - have their height/length measured in your clinic? Standardized lying measuring equipment Tape measure Segmental measure with tape measure (e.g. knee height, leg length, upper arm length) Segmental measure with calipers (e.g. knee height, leg length, upper arm length) Not at all Other (please specify, including which segmental measurements): 3. How are children and youths with CP weighed in your clinic? Naked or with light underwear only Fully clothed, but removing footwear, splints, etc. Fully clothed including footwear, splints, etc. Not in a consistent way Not at all 6

8 4. How do children and youths with CP - who can stand - have their weight measured in your clinic? Standardized weighing scale, which is regularly calibrated Standardized weighing scale, not regularly calibrated Non-standardized weighing scale Not at all Other (please specify): 5. How do children and youths with CP - who cannot stand or sit - have their weight measured in your clinic? Standardized equipment, regularly calibrated (e.g. chair or bed weight scale) Standardized equipment, not regularly calibrated (e.g. chair or bed weight scale) Non-standardized equipment Held by carer, weight then calculated Not at all Other (please specify, including which standardized equipment): 6. Are there any other anthropometric measurements recorded regularly in your clinic for children and youths with CP? Occipitofrontal circumference Skin fold thickness Bioimpedance 7. Does your clinic have a feeding team for children and youths with CP? If 'Yes', does your feeding team include a nutritionist? Yes No 8. Is gastrostomy available for children and youths with CP? Yes, in the hospital were my clinic is located Yes, by referral to another general or district hospital Yes, by referral to a specialist/university hospital Not available 7

9 Results The results of the first survey indicated that only six registers were able to provide information on feeding problems, use of gastrostomy and had anthropometric data available, although not regularly recorded in all registers data bank. Five of the registers were located in Northern Europe (DK, ISL, NO, S, UK), while the sixth was located in Southern Europe (Portugal). These registers also had some information on feeding difficulties. Despite the limited number of registers, we decided to continue with a study on growth and nutritional status as well as on the prevalence of gastrostomy tube feeding (GTF). The results showed that children with CP were shorter, lighter and thinner than children of the same age in their respective background population. Retarded growth was most significant among children with the most severe disabilities. The use of GTF varied significantly between centers, and interestingly the prevalence correlated with the GINI-coefficient, suggesting some inequalities in access to this treatment in Europe. Most importantly however, growth retardation in the most disabled group was greatest in the center with the lowest prevalence of GTF, while it was least in the center with the highest prevalence. This suggests that the potential inequality to health care may have consequences for the children s health. Finally, the study showed great variation in how feeding problems are recorded, which is an important limitation when decisions on appropriate interventions are needed as well as for comparison of treatment outcomes between countries. Although one of the participating centers developed and used a validated classification scale, this scale has so far not been used and validated in other populations. Thus, we have therefore also identified a need to agree on a common classification of feeding difficulties that may be useful for clinicians as well as researchers. These findings have been described in a manuscript accepted for publication in Developmental Medicine & Child Neurology (DMCN). A copy of the manuscript is attached as Appendix I. In the survey we asked clinicians throughout Europe to describe how they assessed growth and nutritional status in children with CP. This is an important issue since many children with CP are unable to stand and since they may have contractures and scoliosis. In all, 37 (42%) clinics, representing 17 countries responded. The respondents represented Northern Europe (N= 18; DK, FIN, ISL, N, S, UK), central-southern Europe (N = 12; A, D, F, Swiss, Portugal) and former members of eastern Europe (N = 7; BOSN-HER, BUL, LT, LIT, ROM, SLOV). Approximately 90% of the clinics reported that the height and weight of children who were able to stand were measured using standard equipment, such as a stadiometer and a standard weighing scale. On the contrary, in children who were not able to stand, 27 (73%) clinics reported that they used a standard scale, such as a chair or bed weighing scale. Only 7 (19%) used a caliper to obtain segmental measurements as proxies for height. Furthermore, only 8 (22%) clinics used standard equipment to assess nutritional status: Seven clinics used skinfold measurements, and only one clinic used bioimpedance. Head circumference was regularly measured at 26 clinics (70%). There were no clear differences between the regions in the use of adapted equipment. 8

10 In 27/37 (73%) clinics they reported that they had feeding teams and in a majority (85%) the feeding team included a nutritionist. 95% of the clinics reported availability of gastrostomy for children and youths with CP, either in their own clinics or by referral. Discussion The overall aim of Task 5.3 was to promote best practice in health care for children with CP and to contribute to the reduction of inequalities in care between countries. The specific aim of this task was to document variations in health care of growth and nutrition outcomes. We have shown that there are variations in prevalence of gastric tube feeding (GFT) in children with CP between countries in Europe. What we do not know is if this variation is due to differences in access to care (GFT) or to differences in the decision processes. However, we did find that growth retardation in the most disabled group was greatest in the center with the lowest prevalence of GTF, while it was least in the center with the highest prevalence. This strongly suggests that presence of GFT have consequences for the children s health and that the availability of GFT as a treatment option is a mean to measure potential inequality to health care. Secondly, we have also shown that there is a great variation between countries in how feeding problems are recorded. This is an important limitation when decisions on appropriate interventions are needed and for comparison of treatment outcomes between countries. Thirdly, only 1/5 of clinics used standard equipment to assess nutritional status. Further, while most clinics used standard equipment to measure height and weight of children who were able to stand, only 19% used a caliper to obtain segmental measurements as proxies for height who were not able to stand. There were no clear differences between the regions in the use of adapted equipment. Our recommendations: 1. For good clinical practice: Use skin fold measurements to assess nutritional status. In addition to use appropriate assessment of weight, and preferably standardized segmental measurements of height/length. Use better characterization of feeding difficulties to be able to identify feeding difficulties earlier and to be able to make standards and procedures for interventions. 2. New variables to be added to the SCPE common database: Height (in cm) at the time of registration o Segmental measurements taken (yes/no) Weight (in kg) at the time of registration Triceps skinfold thickness (probably not possible but for the final report it would be relevant to have asked about this) Gastrostomy (yes/no) o If yes, age at placement Feeding difficulties (Eating and Drinking Ability Classification System (EDACS), when validated) 9

11 10

12 Appendix I: Gastrostomy tube feeding of children with cerebral palsy: variation across six European countries Magnus Odin Dahlseng 1, Guro L. Andersen 2, Maria da Graça Andrada 3, Catherine Arnaud 4, Rajesh Balu 5, Javier de la Cruz 6, Teresa Folha 7, Kate Himmelmann 8, Karen Horridge 9, Pétur Júlíusson 10, Magnus Påhlman 8, Gija Rackauskaite 11, Solveig Sigurdardottir 12, Peter Uldall 13 and Torstein Vik 1 on behalf of the Surveillance of Cerebral Palsy in Europe Network (SCPE-NET). Correspondence: Magnus O. Dahlseng, Kvinne Barn-senteret, 6. etasje, nord St. Olavs Hospital, Olav Kyrres gt. 11, 7006 Trondheim, Norway. Tel: magnusodin86@gmail.com 1 Department of Laboratory Medicine, Children s and Women s Health, Faculty of Medicine, Norwegian University of Science and Technology, Trondheim, Norway. 2 The Cerebral Palsy Registry of Norway, Vestfold Hospital Trust, Tønsberg, Norway. 3 The Cerebral Palsy Registry of Portugal, Federação das Associações Portuguesas de Paralisia Cerebral. 4 Inserm, UMR 1027, F Toulouse, France Toulouse III University, UMR 1027, F Toulouse, France. Clinical Epidemiology Unit, University Hospital, F Toulouse, France. 5 Paediatric Department, Sunderland Royal Hospital, Sunderland, UK. 6 Clinical Research Unit, Imas12-Ciberesp, Hospital 12 Octubre, Madrid, Spain. 7 Calouste Gulbenkian Cerebral Palsy Rehabilitation Centre, Lisbon, Portugal. 8 Department of Pediatrics, Institute of Clinical Sciences, Queen Silvia Children s Hospital, Sahlgrenska Academy at the University of Gothenburg, Göteborg, Western Sweden. 9 City Hospitals Sunderland NHS Foundation Trust, UK and North of England Collaborative Cerebral Palsy Survey, UK. 10 Department of Paediatrics, Haukeland University Hospital, 5021 Bergen, Norway and Department of Clinical Medicine, Section of Paediatrics, University of Bergen, 5021 Bergen, Norway. 11 Department of Pediatrics, Aarhus University Hospital, Denmark. 12 State Diagnostic and Counseling Centre, Kopavogur, Iceland. 13 Cerebral Palsy Registry, National Institute of Public Health, University of Southern Denmark and Child Department, Righospitalet, Copenhagen University, Denmark. Abstract Aim To compare prevalence of gastrostomy tube feeding (GTF) of children with cerebral palsy (CP) in six European countries. Methods Data on 1295 children with CP born were collected from geographically defined areas in six European countries; four of the areas covered the whole country. Outcome measures were GTF, age at 11

13 placement, feeding difficulties and the children s height and weight for age standard deviation scores (zscores). Results Use of GTF among all children with CP was highest in Western Sweden (22%; CI: 16-29), and lowest in Portugal (6%; CI: 3-10), Northern England (6%; CI: 3-9) and in Iceland (3%; CI: 0-13) (p < 0.001). The difference between areas was greater among children with Gross Motor Function Classification System levels IV-V (non-walkers); in this group, lower height z-scores were more present in the areas with lower prevalence of GTF. The children s age at placement of gastrostomy also varied between areas (p < 0.002). Interpretation The observed differences in use of GTF may reflect differences in access to treatment or clinical practice, or both. Our results suggest that use of GTF may improve growth in height and weight among children with more severely affected gross motor function the group most likely to have associated feeding difficulties. Keywords Cerebral palsy, feeding, gastrostomy tube feeding, growth, nutrition. What this paper adds The use of gastrostomy and age at placement among children with cerebral palsy differ considerably within Europe. Gastrostomy may improve growth in weight/height among children with GMFCS level IV-V. Validated scales for the assessment of feeding abilities among children with CP are needed. Introduction Feeding difficulties and poor growth are common among children with cerebral palsy (CP), in particular among children with bilateral dyskinetic or bilateral spastic CP with four limb involvement 1-3 Multiple factors contribute to feeding difficulties such as disordered oral motor, pharyngeal and/or oesophageal function and gastro-oesophageal reflux, leading to problems with chewing and swallowing. Children with such problems have high morbidity, often associated with aspiration of food, and they are at increased risk of being undernourished. Poor nutrition may in turn lead to impaired growth. Finally, disordered feeding and restricted growth are prognostic factors for poor survival of children and adults with CP

14 Since feeding difficulties often start early, and since growth restriction increases with increasing age, 7 it is important to ensure optimal nutrition as early as possible. In children with disordered oral motor, pharyngeal and/or oesophageal function, adequate nutrition may be achieved through gastrostomy tube feeding (GTF). However, the use of gastrostomy and the age at placement vary between studies. 3, 8-13 We have previously reported that children who had their gastrostomy placed at an early age were less growth retarded than those who were older. 3 Although placement of gastrostomy is a simple procedure the prevalence may differ. This could be as a result of differences in access to this treatment and also variations in clinicians views regarding which children need gastrostomy. The timing of insertion depends on the child s experience, enjoyment of oral feeding and oro-motor function development for speech, risk of infection, potential worsening of gastro-oesophageal reflux etc. Parents views about the psycho-social benefits and pleasures of oral feeding weighed against the requirement for equipment for feeding may also vary in different settings and countries. The aim of the present study was therefore to explore the prevalence of feeding difficulties and differences in the approach to supporting nutrition in children with CP in geographically defined areas in six European countries by assessing the prevalence of GTF. Due to lack of clear indications for insertion of gastrostomy and possible differences in access to care, we hypothesized that the prevalence of GTF would vary considerably across Europe even after controlling for level of motor functioning as described by the Gross Motor Function Classification System (GMFCS). We also hypothesized that the children s age when the tube was inserted would vary significantly, and finally, that growth restriction would be more marked in areas where GTF was less prevalent and where tube feeding was introduced later, compared with areas with high prevalence and where the tube was inserted at an earlier age. Material and methods Study design This study is part of a 3 year program of work, SCPE-NET ( which utilizes the existing collaboration of 21 registers in 13 European countries - Surveillance of Cerebral Palsy in Europe 14 to promote best practice in describing children with cerebral palsy, and to document variations in access to health care and in health outcomes. Data was collected through CP-registries in six European areas - two regional registries in Western Sweden and Northern England, UK, as well as the four national registries in Denmark, Norway, Portugal and Iceland. In Norway, Sweden and Portugal the necessary data were already collected and recorded by the registries, while in Iceland, Denmark and Northern England such data were retrospectively collected from medical records. 13

15 Study population Children born were included in the study; for Portugal only children born in Children with all CP subtypes and GMFCS levels were eligible. A total of 1295 children were included. In all, 144 (corresponding to 100 % of the total number of children with CP in the local registry) children were living in Western Sweden, 218 (100 %) in Northern England, UK, 414 (98 %) in Denmark, 263 (61 %) in Norway, 218 (97 %) in Portugal and 38 (100 %) in Iceland. Variables The primary outcome was the presence of a GTF. Nasogastric tube feeding was not included. Secondary outcomes were age at placement of gastrostomy, growth and feeding difficulties. Anthropometric measurements were obtained by a number of clinicians according to local practice, and reported to, or collected by their respective registry. No standardized segmental measurements of height/length were recorded. Only one measurement of weight and height for each child was recorded. Weight was recorded to the nearest 0.1 kg and height to the nearest 0.1 cm. Standard deviation score (zscore) for weight and height was calculated using region specific growth references. For children in the Nordic registers we used Norwegian growth curves as reference, 15 for children in Northern England we used British curves 16 and for the Portuguese children we used curves developed by the Centers for Disease Control and Prevention (CDC), 17 consistent with clinical practice and in accordance with the guidelines from the Portuguese National Health Department. In addition to the use of regional growth references, we also 18, 19 analyzed our data using WHO growth references. Exposure variables were CP subtypes and gross motor function level. CP subtypes were recorded as spastic, dyskinetic and ataxic, and the spastic subtype was further divided into the unilateral or bilateral subtypes, according to the criteria of SCPE. 14 Gross motor function was recorded according to GMFCS. 20 Children were categorized as GMFCS level I-II, level III and level IV-V. Other variables recorded were sex and age. All registries provided data on CP subtype, GMFCS and use of gastrostomy. In Northern England, GMFCS level was not routinely recorded in earlier birth cohorts and was thus missing in 85 (39 %) cases. In Portugal, data on age at placement of gastrostomy were not available. Data on weight and height were available in 998 (77 %) and 899 (69 %) of the children respectively. Feeding difficulties were recorded as a dichotomized variable (Yes/No). In Portugal and Norway two different, self-developed feeding scales were used. In the four remaining regions this data was collected from free text information in the children s medical records. Information on feeding abilities was missing in 24% of the cases. 14

16 To study potential inequalities in access to care, we related the prevalence of GTF in each area to the Gini coefficient of income in the corresponding country. 21 The Gini-coefficient is a measure of inequality of income distribution calculated by international agencies like United Nations and Eurostat (EU-SILC 2010 data). The Gini coefficient varies from 0 (complete equality) to 1.0 (complete inequality). The majority of countries in the world have coefficients higher than Ethics National ethical guidelines were followed in each country by the local teams collecting the data. Informed consent and ethical approval to collect data and submit them to the SCPE common database was required in Norway, Portugal and Northern England. In Western Sweden and in Iceland ethical approval was given by regional ethical review boards, while in Denmark the CP-register has been approved by the Danish Data protection Agency. Statistics The Statistical Package for Social Sciences (SPSS) version 18 was used for data analyses. The Chi squared test was used to analyze differences in proportions between areas. Differences in proportions were also visualized by presenting 95% confidence intervals (CI) calculated according to Newcombe and Altman, Group differences of continuous variables with non-normal distribution were analyzed using the Mann Whitney and the Kruskal-Wallis tests. One-way analysis of variance with Scheffe s Post Hoc test was used to analyze whether there were differences between areas in mean z-scores for weight and height. In addition a general linear model was used to adjust for differences in age and CP subtypes between the areas. Spearman s rho was used to study the correlation between the Gini coefficient and growth deviation. The significance level was set to Results Among the 1295 children included in this study 754 (58%; CI: 55-61) were boys. There were significant differences between areas in the proportions of various CP subtypes and GMFCS levels (Table 1). Among 1295 children eligible for the study, 42 (3.2%; CI: 2-5) had died at a mean age of 58.6 months (SD: 24.1). There were no differences between the populations regarding mortality (p = 0.468) or age at death (p = 0.108). The Gini coefficient was low in all countries ranging from 0.23 in Sweden to 0.39 in Portugal (Table 1). Gastrostomy tube feeding Among all 1295 children 133 (11%; CI: 9-12) had a gastrostomy tube (Table 2). The prevalence was highest in Western Sweden (31/144; 22 %; CI: 16-29), and lowest in Portugal (11/194; 6 %; CI: 3-10), Northern 15

17 England (12/218; 6 %; CI: 3-9) and in Iceland (1/38; 3 %; CI: 0-13). When we limited the analyses to children with GMFCS level IV-V, the differences were even more marked, ranging from 67 %(29/43; CI: 53-80) in Western Sweden to 12 %(10/84; CI: 7-21) in Portugal. In Norway the prevalence in this group was 44 % (28/63; CI: 33-57), in Northern England 42 %(11/26; CI: 26-61) and in Denmark the prevalence was 26 % (45/170; CI: 20-34) (Table 3). Age at placement of gastrostomy tube Median age at placement of gastrostomy in the total population was 22 months (range: 1-120). The median age ranged from 16 months in Western Sweden (range: months) to 70 months in Northern England (range: months). In Norway, median age was 19 months (range: 1 to 66 months) and in Denmark the median age at placement was 26 months (range: 4-84 months). The children in Northern England were older when they had their gastrostomy inserted than children in Western Sweden, Norway and Denmark (p = vs. Scandinavian children as a group). Iceland (n = 1) and Portugal (age not available) were not included in these analyses. Growth Weight measurements were reported at a mean age of 5.9 years (SD: 2.3) and height measurements at 5.8 years (SD: 2.2). For weight the mean age at recording was 4.3 (SD: 1.5) years in Denmark, 4.5 (SD: 1.0) in Western Sweden, 5.4 (SD: 1.2) in Iceland, 6.0 (SD: 1.3) in Portugal, 7.3 (SD: 1.5) in Norway and 9.0 (SD: 2.1) years in Northern England. There were similar differences between the areas in mean age at recording of height. In the total population mean z-scores for weight was -0.86(SD: 1.71) and for height (SD: 1.50) (Table 4). Children with the spastic unilateral and the ataxic CP subtypes were considerably less growth retarded than children with the spastic bilateral and dyskinetic subtypes. There was no correlation between duration of GTF and z-score for weight and height (data not shown). When we restricted our analyses to children with GMFCS level IV-V, excluding children from Iceland (n = 6) due to low numbers, we found significant differences in z-scores for weight and height between the areas. The Portuguese children were most and the Swedish children least growth retarded (Table 4). Portuguese children had lower weight z-scores compared to Norwegian, Swedish and Danish children (p < 0.001), while their height z-scores were lower only compared with Swedish and Danish children (p = 0.002). These difference in weight and height z-scores between centers persisted when we adjusted for age at measurement in multivariable analyses. For children with GMFCS levels I-II no significant differences in weight (p = 0.114) or height (p = 0.472) z-scores were observed between the different populations. Among children with GMFCS level IV-V there was a weak, inverse correlation between the Gini coefficient and the weight (r = -0.22; p < 0.001) and height z-scores (r = -0.12; p = 0.042). The differences in z-scores between the countries were more marked when we used WHO growth references. However, using the WHO growth 16

18 references also resulted in statistical significant differences for height z-scores between countries among children with GMFCS-levels I-II (Table 5B; supplementary material). Feeding difficulties Only 262 of 980 (27%; CI: 24-30) children had documentation of feeding difficulties (data not shown). The highest prevalence were Western Sweden (52/122; 43 %; CI: 34-51) and lowest in Northern England (18/131; 14 %, CI: 9-21). Discussion We found considerable variations in the use of GTF in children with cerebral palsy across different areas in six European countries. The differences were even more marked when we restricted the analyses to children at GMFCS level IV-V, which is the group most likely to have associated feeding difficulties. For the latter group z-scores for weight and height differed between countries, showing less growth restriction in the areas with high prevalence of GTF and more growth restriction in areas with low prevalence of GTF. The age at introduction of GTF also differed considerably, being lowest in the area with the highest prevalence of GTF. The low p-values suggest that the results are unlikely to be due to chance. Strength of the study is the population based multicentre design and the large number of children providing relatively narrow confidence intervals in the total study population. A limitation of the study is the lack of standardization of anthropometric measurements. This is likely to affect height measurements more than weight, in particular among the most severely affected children, and in this subgroup the results related to height measurements should be considered with caution. In a recent, unpublished survey among European clinicians we found that only a minority of clinicians used standardized methods (i.e. segmental measures with calipers) to measure height/length in children with CP unable to stand, and the use of such standardized methods was not restricted to a specific country or region. It is therefore unlikely that a systematical bias due to different methods or to lack of standardized measurement explain the differences in growth between centers. A further limitation is that we only have one measurement of weight and height, making it impossible to study changes in growth over time. We preferred to use local growth charts in the calculations of z-scores. Since there were no differences in weight and height z-scores among children at GMFCS levels I-II, who are more independently mobile and who are less likely to have nutritional problems, we consider it unlikely that use the use of different growth charts has confounded our results. Moreover, we also analyzed our data using WHO growth standards, resulting in more marked differences in growth between centers. However, WHO growth standards are 17

19 supposed to indicate optimal growth and our results using these standards may therefore partly be confounded by differences in growth in the background population. This interpretation may be further supported by the finding of significant differences in growth among children with GMFCS I-II between regions. The differences in prevalence of GTF and in weight and height/length z-scores between the populations persisted after multivariable analyses making confounding by differences in CP subtypes, gross motor function and the age at assessment of the children unlikely. Moreover, in analyses restricted to GMFCS level IV-V the differences in GTF were even greater. Moreover, in the latter group, an association between high prevalence of GTF and less growth restriction could be observed, an association that was not apparent in the total population. 3, 8, 9, Both the prevalence of GTF and age at placement are consistent with a number of single centre studies. 23 In Western Sweden the prevalence was 22 % (31/144; CI: 16-29), clearly higher than in a population based study conducted in Southern Sweden where the prevalence was 13 %. 6 However, in the latter study the proportion of children with GMFCS level V was lower. The differences in use of GTF across areas observed in this study may be due to differences in access to gastrostomy, clinical decision-making, lack of clear guidelines, and parents views about the psycho-social benefits and pleasures of oral feeding or a combination of these. We did not include use of nasogastric tube feeding in this study, since nasogastric tube feeding is not recommended for longer periods, i.e. exceeding 6 weeks. 24 However, we cannot exclude that such feeding has been given to some patients for a longer time. We originally wanted to study the prevalence of feeding difficulties. However, this was not possible due to lack of a consistent description of feeding difficulties. Only in two areas were feeding scales used, and both scales were developed locally and only one had been validated. 25 Thus, an important implication of our study is that there is a need for international classification scale of feeding difficulties, not only for epidemiological purposes, but also to provide a better basis for a clinical decision to recommend if and when a gastrostomy tube should be inserted. One such scale has been developed in Portugal, 25 another scale is currently being developed in the UK. 26 It is interesting that there was a correlation between the Gini-coefficient and height and weight z-scores among children with GMFCS levels IV -V, and that the prevalence of GTF was lowest in the country with the highest Gini coefficient and vice versa, despite the fact that all countries had low Gini coefficients. These associations could suggest that the observed differences in gastrostomy between the populations may partly reflect differences in access to care. 18

20 An important finding was that among the children at GMFCS levels IV-V, i.e. those at the highest risk of associated feeding and swallowing difficulties, there was a clear association between low prevalence of gastrostomy and more severe growth restriction. These findings suggest that a gastrostomy tube may improve nutritional status and consequently growth even in those children whose motor function is at GMFCS levels IV-V. This is particularly important considering the association between growth restriction and early death. 27 However, concerns have been raised about potential overfeeding. 28 GTF has been reported to 1, 11 have positive effects regarding the quality of life of the individual child and its family. In conclusion we found that the use of GTF varied considerably between areas in six European countries. We postulate that this is due to lack of clear guidelines and differences in the decision making processes. Among children who had gross motor impairments corresponding to GMFCS levels IV-V, those living in areas where GTF was less common were more growth restricted than children in countries where GTF was more prevalent. Acknowledgements: We thank Professor Allan Colver, Newcastle for input in the design of the study as well as for having contributed to the manuscript. We also want to acknowledge all co-workers from the SCPE network who has contributed with input to the paper, discussions and practical help. List of SCPE collaborators: C Cans, E Sellier (RHEOP, Grenoble, FR), C Arnaud, M Delobel, D Klapouszczak (RHE31, Toulouse, FR), L Gibson, A Lyons (Lavanagh Centre, Cork, IE), J Parkes, N Hill (Belfast, UK), M Pahlman, K Himmelmann (University of Gothenburg, Gothenburg, SW), O Hensey, V Dowding (Central Remedial ClinicA, Dublin, IE), A Colver, Karen Horridge (University of Newcastle, Newcastle, UK), J Kurinczuk, G Surman (NPEU, University of Oxford, UK), I Krägeloh-Mann, V Horber (Tübingen University, Tübingen, DE), MJ Platt (University of East Anglia, Norwich, UK), P Udall, S Michelsen (NIPH, Copenhagen, DK), MG Torrioli, M Marcelli (Lazio Cerebral Palsy Register, Rome, IT), G Andersen, T Vik (CPRN, Tonsberg, NO), J De la Cruz, C Pallas (DIMAS-SAMID, Madrid, SP), D Virella, E Calado (Lisbon, PT), D Neubauer, D Osredkar (Ljubljana, SI), K Hollody, G Csabi (Pecs, HU), S Sigurdardottir, I Einarsson (Iceland, IS), V Mejaski-Bosnjak, K Bosnjak-Nadj (Zagreb, CR). The study was funded by the EU Health Programme; Grant number DG SANCO-EAHC

21 20

22 21

23 1 Sullivan PB, Juszczak E, Bachlet AM, et al. Impact of gastrostomy tube feeding on the quality of life of carers of children with cerebral palsy. Dev Med Child Neurol 2004; 46: Fung EB, Samson-Fang L, Stallings VA, et al. Feeding dysfunction is associated with poor growth and health status in children with cerebral palsy. J Am Diet Assoc 2002; 102: Dahlseng MO, Finbraten AK, Juliusson PB, Skranes J, Andersen G, Vik T. Feeding problems, growth and nutritional status in children with cerebral palsy. Acta Paediatr 2012; 101: Strauss DJ, Shavelle RM, Anderson TW. Life expectancy of children with cerebral palsy. Pediatr Neurol 1998; 18: Strauss D, Shavelle R. Life expectancy of adults with cerebral palsy. Dev Med Child Neurol 1998; 40: Westbom L, Bergstrand L, Wagner P, Nordmark E. Survival at 19 years of age in a total population of children and young people with cerebral palsy. Dev Med Child Neurol 2011; 53: Andrew MJ, Sullivan PB. Growth in cerebral palsy. Nutr Clin Pract 2010; 25: Vik T, Skrove MS, Dollner H, Helland G. [Feeding problems and growth disorders among children with cerebral palsy in south and north Trondelag]. Tidsskr Nor Laegeforen 2001; 121: Sullivan PB, Lambert B, Rose M, Ford-Adams M, Johnson A, Griffiths P. Prevalence and severity of feeding and nutritional problems in children with neurological impairment: Oxford Feeding Study. Dev Med Child Neurol 2000; 42: Henderson RC, Grossberg RI, Matuszewski J, et al. Growth and nutritional status in residential center versus home-living children and adolescents with quadriplegic cerebral palsy. J Pediatr : Mahant S, Friedman JN, Connolly B, Goia C, Macarthur C. Tube feeding and quality of life in children with severe neurological impairment. Arch Dis Child 2009; 94: Sullivan PB, Juszczak E, Bachlet AM, et al. Gastrostomy tube feeding in children with cerebral palsy: a prospective, longitudinal study. Dev Med Child Neurol 2005; 47: Lewis EC, Connolly B, Temple M, et al. Growth outcomes and complications after radiologic gastrostomy in 120 children. Pediatric radiology 2008; 38: Surveillance of cerebral palsy in Europe: a collaboration of cerebral palsy surveys and registers. Surveillance of Cerebral Palsy in Europe (SCPE). Dev Med Child Neurol 2000; 42: Juliusson PB, Roelants M, Eide GE, et al. [Growth references for Norwegian children]. Tidsskr Nor Laegeforen 2009; 129: Cole TJ, Freeman JV, Preece MA. British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by maximum penalized likelihood. Stat Med 1998; 17: Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al. CDC growth charts: United States. Adv Data 2000: Page 22 of 35

24 18 World Health Organization. WHO child growth standards. (accessed 29 March 2012). 19 World Health Organization. Growth reference data for 5-19 years. (accessed 29 March 2012). 20 Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997; 39: Central Intelligence Agency. Distribution of family income - Gini index, The World Factbook. (accessed 6 January 2012). 22 Newcombe RG, Altman DG. Proportions and their differences. In: DG Altman, D Machin, TN Bryant, MJ Gardner editors. Statistics with confidence, 2nd edn. Bristol: BMJ books; Stevenson RD, Conaway M, Chumlea WC, et al. Growth and health in children with moderate-tosevere cerebral palsy. Pediatrics 2006; 118: Sullivan PB. Feeding and Nutrition in Children with Neurodevelopmental Disability. London: Mac Keith Press; Andrada M. Validation of assessment scales for communication and oro-motor function in children with cerebral palsy. Developmental Medicine & Child Neurology 2008; 50: Sellers D. The Eating and Drinking Ability Classification System for Individuals with Cerebral Palsy EDACS. (accessed 10 April 2012). 27 Brooks J, Day S, Shavelle R, Strauss D. Low weight, morbidity, and mortality in children with cerebral palsy: new clinical growth charts. Pediatrics 2011; 128: e Sullivan PB, Alder N, Bachlet AM, et al. Gastrostomy feeding in cerebral palsy: too much of a good thing? Dev Med Child Neurol 2006; 48:

25 Appendix II: Abstract accepted: 1. Scientific Poster at the 66th Annual Meeting of the AACPDM, Toronto, ON, Canada, September 12-15, Oral presentation, 4 th International Cerebral Palsy Conference (ICPC), Pisa, Italy, October 2012: How is growth and nutritional status assessed in children with cerebral palsy (CP)? A cross-sectional survey in 17 European countries. By Sandra J. Hollung, Guro L. Andersen and Torstein Vik, on behalf of the SCPE-NET. Background/objectives: A significant proportion of children with cerebral palsy (CP) have feeding difficulties. The most severely affected children with CP are at risk of malnourishment, impaired growth and a poor prognosis for survival. Linear growth of children is most commonly measured in a standing position with a stadiometer. However, for children with CP, who are not able to stand, segmental measures like knee height, leg length and upper arm length have been proposed as proxies for linear growth. As a result of inaccurate height measurements, body mass index (BMI) is also considered an unreliable assessment of nutritional status. Therefore, skinfold thickness measurements have been proposed as a reliable clinical measurement tool. During the last two decades researchers have proposed both segmental measurements as a proxy for height, and skinfold thickness measurements to assess nutritional status. To see if these measurements are commonly used, we carried out a survey among European clinicians caring for children with CP. Study design: Cross-sectional survey Study participants and setting: Clinicians at departments of pediatrics, neuropediatrics and/or habilitation centers in Europe. The clinicians were contacted through the SCPE-NET and various European pediatric organizations. In all, 88 clinicians in 30 countries were contacted. A total of 37 (42%) clinicians, representing 17 countries responded. Only one response per clinic was allowed. Materials/Methods: A questionnaire was constructed using a web based survey tool (Easyfact ) and sent to participants by . The questionnaire included eight questions regarding growth and nutrition and took an average of six minutes to complete. The questions addressed how height and weight were measured in children with CP, and if other measurements such as knee height or skinfold calipers were used. Results: Approximately 90% of the clinics reported that the height and weight of children who were able to stand were measured with standard equipment, such as a stadiometer and a standard weighing scale. On the contrary, in children who were not able to stand, 27 (73%) clinics reported that they used a standard scale, such as a chair or bed weighing scale. Only 7 (19%) used a caliper to obtain 24

26 segmental measurements as proxies for height. Furthermore, only 8 (22%) clinics used standard equipment to assess nutritional status: 7 clinics used skinfold measurements, and only one clinic used bioimpedance. Head circumference was regularly measured at 26 clinics (70%). Conclusions/Significance: Only a fifth of the clinics who participated in this survey used standard equipment to measure height and assess the nutritional status of children with CP who were not able to stand. The results suggest that there is significant potential for improvement in the quality of linear growth measurements and nutritional status assessments of children with severe CP. Funding: EU Health Programme; Grant number DG SANCO-EAHC

27 Appendix III: GROWTH AND NUTRITION SCPE - WP5.3 Data collection Birth Years: All subtypes and all GMFCS levels DEADLINE: February 28, 2011 Guro L. Andersen guro.andersen@siv.no tlf Sandra J. Hollung sandra.julsen.hollung@siv.no tlf Torstein Vik torstein.vik@ntnu.no tlf