Abstract
Background Limited information is available regarding physical therapy use for individuals with cerebral palsy (CP).
Objectives The purpose of this study was to evaluate the association of Gross Motor Function Classification System (GMFCS) level, age, race, sex, and type of insurance with the total physical therapy units received over a 1-year period for individuals with CP in this outpatient pediatric medical setting.
Design This was a cross-sectional study.
Methods Four hundred twenty-five individuals with CP (GMFCS level I, 36%; level II, 15%; level III, 13%; level IV, 19%; and level V, 17%) were identified retrospectively through their electronic medical records. A one-way analyses of variance (ANOVA) was performed for each explanatory variable followed by a multiway ANOVA that adjusted for other variables to find the best model to explain total physical therapy units received.
Results A significant difference in total therapy units received was found among GMFCS levels (F=6.91; df=4,420; P<.001), age groups (F=4.76; df=3,421; P=.028), and type of insurance (F=8.09; df=2,422; P=.004). No significant difference in physical therapy received was found for the factors of sex and race. The final multifactorial model indicates a significant main effect of insurance and a GMFCS by age interaction accounting for 19% of the variability (F=4.45; df=21,403; P<.001).
Limitations This study is cross-sectional and examines physical therapy services received in a pediatric medical setting in 1 geographic region of the United States.
Conclusions The results of this study provide insight into how therapy received varies for individuals with CP. Future studies should evaluate additional variables that may affect physical therapy services received.
Cerebral palsy (CP) describes a group of permanent disorders of movement and posture that are attributed to nonprogressive disturbances in the developing brain1 affecting 3.3 per 1,0002 children. It is a heterogeneous condition in terms of etiology and severity, and often is accompanied by disturbances of sensation, cognition, communication, perception, behavior, or seizures.1 It is estimated that between 765,000 and 1,000,000 children and adults with CP are living in the United States.2
Cerebral palsy is the most common neurological diagnosis of children seen by pediatric physical therapists.3 Physical therapists play a major role in providing services that help individuals with CP reach their full potential in their homes and communities. Decisions about many medical interventions, such as surgery and spasticity management, rely on input from a physical therapist.3 Despite these facts, a limited amount of literature exists that quantifies physical therapy service use for those with CP. Furthermore, the factors that influence the amount of physical therapy service used by individuals with CP are not known.
The American Physical Therapy Association's revised research agenda identified an important aim: “Describe patterns of physical therapy use and identify factors that contribute to variation in utilization.”4(p173) This need is supported by evidence from families: 68% of parents of children with CP expressed a need for information about services their children might receive in the future, and 57.7% of parents expressed the need for information about planning for their children's future well-being.5
Studies with small sample sizes,6 limited age ranges,7 and qualitative classification of function8 have diminished the ability to understand the relationship between physical therapy service use and gross motor function in those with CP. Studies of children with a variety of special health care needs examined all therapies together, making it difficult to draw specific conclusions about physical therapy.9–12 However, severity of the condition was found to be significantly associated with an increased risk of unmet needs10 and greater need for therapy services.11 Furthermore, it was recommended that functional limitation be considered for program planning.11
The Gross Motor Function Classification System (GMFCS) is an internationally used method for clearly describing functional ability of children and young adults up to age 18 years with CP, with an emphasis on sitting, transfers, and mobility. It identifies 5 levels of gross motor function within several age bands from level I, walks without limitations, to level V, transported in a manual wheelchair. The GMFCS has had a major effect on the health care of children with CP.13 It is reliable,14,15 valid,16 and stable over time17 and has been used in research to successfully prognosticate motor function.18 Although the GMFCS was intended for use in service planning, there are limited reports of its use for that purpose.13 Previous studies of service use lack large samples of children with CP across all age groups and information regarding children's function using the GMFCS. Knowledge of how physical therapy service use varies among individuals with CP at different levels of functioning can have implications for policy, decision making, and service planning.
The purpose of this study was to evaluate the association of GMFCS level, age, race, sex, and type of insurance with the total physical therapy units received over a 1-year period for individuals with CP in this outpatient pediatric medical setting. We hypothesized that GMFCS level and age, not race, sex, or type of insurance, would be associated with total physical therapy units received.
Method
Design
This was a cross-sectional study of the 2008 electronic records (Chart Links Rehab Systems, version 4.0.8, Chart Links LLC, New Haven, Connecticut) of 1 large pediatric tertiary care hospital's outpatient division of occupational therapy and physical therapy, which serves individuals at 10 locations in and around a southwestern Ohio metropolitan area. Data were collected retrospectively for individuals 1 year or older who received physical therapy services at this institution's outpatient therapy clinics in the year 2008 with the International Classification of Diseases, ninth revision, (ICD-9)19 diagnosis of CP, 343.0–343.9. Children younger than 1 year with CP were not included because it is difficult to assign a GMFCS level to children this young. We have experienced an increase in the number of individuals with chronic childhood conditions such as CP who continue to receive care at our pediatric facility as they transition to adulthood. Therefore, we chose not to have an upper age limit for participants.
The primary independent variable was the child's GMFCS level. The dependent variable, total therapy units received in the year, was treated as a continuous variable. One therapy unit is equal to 15 minutes of care. Secondary independent variables included age, sex, race, and type of insurance. Institutional approval was granted before the primary investigator (PI) and trained research assistant began systematically collecting data from the electronic record for each of the variables. In cases where the medical record did not include the GMFCS level, the PI reviewed the therapy and medical notes and identified the level that best described the child's ability. If documentation in the medical record was not sufficient to assign a GMFCS level, the PI contacted the rehabilitation physician or treating physical therapist familiar with the child to assign a level or confirm the level chosen by the PI. We abstracted the age of the child from the first physical therapy visit in the record. For consistency, we chose the insurance type listed that coincided with the first therapy visit documented in the medical record. The data were de-identified and maintained in a secure database that was managed by the PI to minimize access and the potential for errors during data entry. The PI and a trained research assistant performed data cleaning. Finally, to minimize random errors in data extraction, the PI reviewed every chart for accuracy.
Data Analysis
We used SAS version 9.2 (SAS Institute Inc, Cary, North Carolina) for data analysis, and descriptive statistics were generated for all independent variables: GMFCS level, age, race, sex, and type of insurance. In addition, frequencies of the total number of therapy units received in the year were calculated. In the analyses, we treated the independent variables as categorical variables. We grouped age into 4 categories: 1 year to ≤5 years, >5 to 12 years, >12 to 18 years, and >18 years. We chose these age groupings because they represent meaningful transition points for individuals when a change in service use might occur, such as at the start of kindergarten through elementary school years, advancing to middle and high school, and graduating high school. Because of the small numbers in some race categories, we categorized race into Caucasian or non-Caucasian. Ethnicity was not available. Type of insurance was categorized into private, public, or “none.” Because of the retrospective nature of this study, limited information was available as to why an individual had no insurance coverage. The category of “none” could include several reasons, such as private pay, insurance carrier denial of coverage, international residence, or lapses in coverage. Tests for normality were applied to the dependent variable to assess the normality of total therapy units received. The dependent variable was not normally distributed and was positively skewed. Therefore, the data were transformed logarithmically so that parametric tests could be used for analyses.20 We report geometric means, which typically approximate the median and are the best measure of central tendency when data are positively skewed such as in this study. Arithmetic means are less appropriate to report when the data are skewed and would overestimate the median. A one-way analysis of variance (ANOVA) was performed for each explanatory variable individually to determine differences in the total therapy units received in 2008 among the GMFCS levels, age groups, races, sexes, and types of insurance. The nominal significance level was controlled for multiple pair-wise comparisons by the Tukey procedure, with a family-wise α level set at .05.
A multiway ANOVA was used, adjusting for age, sex, race, and type of insurance, to find the best model to explain the outcome variable, total number of therapy units received in the year. Least-square means were compared using the Tukey procedure for multiple comparisons. Two-way interactions between the covariates and the key independent variable were assessed. The differences between the observed values and those estimated by the model, the residuals, were examined to assess the goodness of fit of the ANOVA model obtained.
Role of the Funding Source
This work was partially supported by the Foundation for Physical Therapy Promotion of Doctoral Studies Scholarships awarded to Ms Bailes, 2009–2012.
Results
A review of the 2008 electronic medical records revealed that 425 individuals with CP were seen in the outpatient division of occupational therapy and physical therapy in 2008. Table 1 shows the frequency and distribution of the sample characteristics and total therapy units received. The total number of physical therapy units received in the year was 15,139, equating to 3,784.75 hours of therapy. The percentage of children who were classified into each GMFCS level were as follows: 36% into level I, 15% into level II, 13% into level III, 19% into level IV, and 17% into level V. Limb distribution was as follows: hemiplegia, 30% (n=128); diplegia, 25% (n=105); quadriplegia, 32% (n=137); triplegia, 1% (n=6); and unspecified CP, 12% (n=49). The oldest age group consisted of individuals with a mean age of 24 years (median=23, SD=4, range=18–33). Medians and interquartile ranges are reported for each variable of interest.
Descriptive Characteristics and Total Therapy Units of Study Sample (N=425)
Individual analyses for each explanatory variable demonstrated a significant difference in total therapy units received among GMFCS levels (F=6.91; df=4,420; P<.001), age groups (F=4.76; df=3,421; P=.028), and types of insurance (F=8.09; df=2,422; P=.004). Results indicate no significant differences in total therapy units received for the factors of sex and race.
The significant variables, GMFCS level, age, and type of insurance, were entered into the final model, and results of the multiway ANOVA are presented in Table 2. Findings indicate a significant main effect of insurance and a significant GMFCS × age interaction. Nineteen percent (R2=.19) of the variability in total amount of therapy units received in the year was explained by the final model (F=4.45; df=21,403; P<.001). Multiple comparisons were made for each variable in the final model, adjusting for all other terms in the model, and are shown in Table 3. The average number of physical therapy units received per individual was greatest for those classified as GMFCS level III. A significantly greater average number of therapy units per individual was received by those in levels II, III, and IV than those in level V. The average number of therapy units received per individual in level I was not significantly different from those in level V. Individuals in the ≤5 years age group received an average number of therapy units that was significantly greater than that for individuals in the >12 to 18 years group but not different from that of the other age groups. Those with private insurance received an average number of therapy units per individual that was significantly greater than those with public insurance. On visual inspection, the residuals were normally distributed, indicating that the normality assumption was reasonable for the final model.
Results of 3-Way Analysis of Variance on 3 Factors (Gross Motor Function Classification System [GMFCS], Age, and Insurance Type)
Full Model With Multiple Comparisons Ordered From Greatest to Least Use Using Tukey Adjustmenta
The Figure graphically displays the geometric mean total therapy units received per individual for the year for each GMFCS level by age group. In general, individuals received greater amounts of therapy at younger ages. However, we observed a spike in services received for individuals >18 years of age for GMFCS levels II and III. We observed a decrease in services received for the >12 to 18 years age group for all GMFCS levels except for GMFCS level V, which was slightly higher. The amount of therapy units received appears to not differ by functional level during the 12 to 18 years age range. Individuals in GMFCS level III demonstrated a large amount of physical therapy units received, except in the >12 to 18 years age group. For individuals classified as level IV, services received decreased across age groups. Individuals at the greatest mobility (level I) and lowest mobility level (level V) received the fewest physical therapy units.
Graph of estimated geometric mean number of therapy units by age group and Gross Motor Function Classification System level.
Discussion
This study describes the amount of therapy units received in 1 year for individuals with CP in this outpatient pediatric academic medical facility and evaluates the association between total therapy units received and functional level, sex, age, race, and type of insurance. The results indicate a significant difference in total amount of therapy units received in 1 year among GMFCS levels, age groups, and type of insurance, but not with regard to sex or race. In addition, findings provide evidence of a significant interaction between GMFCS level and age with regard to total amount of physical therapy units received in 1 year. The lack of interaction of insurance with other variables indicates that the effect of insurance is consistent across GMFCS levels and age groups.
Comparing our study results with previous work is challenging for several reasons. First, some studied unmet service needs rather than amount of actual therapy received and found a significant association between severity of functional limitation and increased risk of having an unmet need5,10,11 for children with special health care needs. Only Palisano and colleagues'5 results were specific to CP with utilization of the GMFCS to classify function. Dusing et al10 and Nageswaran et al11 combined all types of therapy, whereas Palisano et al5 explored needs not specific to any service type in the areas of information, support, community services, and finances.
Palisano et al5 reported the highest needs for children and youth in GMFCS levels IV and V, whereas we report the lowest number of therapy units received for individuals in level V. It is not clear what the relationship might be between reported needs and actual therapy received. It could be that individuals in level V had physical therapy needs not met by this institution. However, some needs might be beyond the scope of direct physical therapy. Palisano et al5 did not find differences in reported needs by age groupings and did not include individuals over 21 years of age.
Interestingly, 1 study compared different time periods and did not find any trends in amount of therapy service use for children with CP.12 Although our study did not evaluate trends over time, this topic justifies further study, especially because children with CP are living longer.21
Two studies,11,22 not specific to children with CP or to physical therapy, assessed the relationship between insurance and functional limitation in children with special health care needs and showed that those with severe limitations reported a higher proportion of public insurance. One study11 did not examine therapy use. The other study,22 which compared service use between school and beyond school settings, showed that those children with public insurance were more likely to use services beyond the school setting compared with those with private or no insurance. When controlling for other variables in the model, our study showed that those with private insurance received significantly greater service compared with those with public insurance. Type of insurance can change, or there can be gaps without insurance throughout the year, and this study did not capture whether this was the case for any of the individuals. In addition, families may have secondary insurance for their children that was not accounted for in this study. For these reasons, future studies should continue to explore the relationship between type of insurance, functional limitation, and physical therapy services received for individuals with CP.
Three studies examined actual therapy use.6,8,12 Our results are in disagreement with the findings of 2 studies,8,12 in which severity of functional limitation was associated with increased use of health services. Parkes et al8 grouped severity of functional limitation into moderate (having abnormal gait) and severe (no walking ability), whereas Boulet et al12 defined severe as answering “yes” to at least 2 or more functional limitation questions in their survey. Although these studies were specific to CP, none of them used the GMFCS to describe the children. All therapies were grouped together in one study,12 whereas the other study8 was specific to physical therapy. Parkes et al8 did not find a difference in service use across ages in their study of 4- to 15-year-old children. Our study included wider age ranges and showed that younger children received more therapy than those in older age groups. Our results are in agreement with findings of a study by Majnemar et al,6 in which severity of delay did not appear to influence likelihood for receiving service. Majnemar et al,6 who studied a small sample of children <5 years of age, assigned severity of delay by comparing functional age equivalents from standardized assessments with chronological age into severe, moderate, or mild. Their sample size was small for the severe group, and they were able to detect variation in service delivery only between children with mild delay and those with moderate delay. Our study included a large sample of individuals across all age ranges and used the GMFCS to classify functional level, which has been found to be more useful than past methods of describing function.13 In addition, past studies varied by health care delivery system. Two studies were in countries with universal health coverage,8,6 whereas the other study12 was completed on a US sample. Further work is warranted to study similarities and differences among countries with varying health care systems.
Our results suggest that the amount of physical therapy received appears not to differ by GMFCS level for the 12 to 18 years age group. There could be several possible explanations for this finding. Individuals in this age group may be better served in their communities or schools. Medical professionals may believe that less physical therapy is necessary during these years. Mobility issues during these years may not be as high a priority as other things and, therefore, less physical therapy service is pursued by the family. Moreover, families may be more comfortable with caring for their child during these years and do not require visits to medically based professionals as often.
Although the total number of physical therapy units received was greater in GMFCS level I, the mean number of units received per individual was lower for this classification. Population-based studies23,24 indicate a larger percentage of those with CP are functioning at higher levels, which was similar for our study. However, individuals at higher levels typically do not require physical therapy services to initiate walking, probably explaining the lower mean number of units received per individual for those classified at level I. Our results show a spike in the average number of physical therapy units received per individual in GMFCS levels II and III in the oldest age group. It has recently been reported that individuals in levels III, IV, and V may lose skills as they become young adults.25 In our study, the spike in services for those in level III may be related to this phenomenon of losing skills. However, we did not observe an increase in physical therapy services received for children in levels IV and V. This finding could be due to the small numbers of individuals in these age and functional groups and warrants further study. In addition, we did not examine why individuals were continuing to seek care at this pediatric institution. Recent literature26 reports individuals with chronic childhood conditions such as CP continue to use pediatric centers as they transition into adulthood because of disease complexity and poor transition planning services.27,28 Future studies with a larger sample of individuals in the >18 years to adulthood age group are needed to examine physical therapy service utilization for this population.
The findings of this study reflect care offered at one large pediatric academic institution, where we provide a broad array of services. Individuals with CP may see providers in many divisions such as orthopedics, rehabilitation, neurology, ophthalmology, nursing, neuropsychology, and health care professions services. During 2008, this facility offered specific interventions that may have included Botox (Allergan Inc, Irvine, California), Baclofen (Watson Laboratories Inc, Corona, California) pumps, nerve blocks, individual therapy, adaptive equipment/technology assistance, postorthopedic surgical care, casting, splinting and orthotics, wheelchair clinics, and aquatic programs. Whereas all these services were available to all individuals with CP, our institution does not offer any specific services for those who are >18 years of age. The findings from this study cannot be generalized to other types of settings where individuals might receive services such as in a rural setting, community program, and different cities across the United States or in different countries.
Our findings account for only 19% of the variance in physical therapy use. These results suggest there are other factors not included in this study that probably contribute to the amount of physical therapy services received. Furthermore, the results point to the need to involve families in developing relevant and useful research questions to gain their perspective on service use.29 Finally, future work should utilize mixed-method study designs to capture both quantitative and qualitative data to broaden our understanding of how physical therapy service use can be influenced by personal and social factors.30
Strengths and Limitations
Several limitations must be taken into consideration when evaluating the results of this study. The cross-sectional nature of this study does not allow causal inferences to be made. This study does not include other variables that probably contribute to the variation in physical therapy services received, such as motivation, readiness to learn, and personal or family characteristics, which might be better addressed with qualitative research methods. This study examined the amount of physical therapy units received in a pediatric academic medical setting in 1 geographical region of the United States and cannot be generalized to services, physical therapy or otherwise, which individuals with CP receive in other settings or other geographical regions. In this geographical region, children are eligible to receive therapy services through early intervention programs when they are between the ages of 0 and 3 years, and school-based services are available from age 3 through graduation from high school. Therefore, this study underestimated the total physical therapy service use of individuals with CP. We recorded type of insurance at 1 time point during the year, and this approach may not show the complete picture regarding insurance status, as it can change throughout the year. Future studies may include additional data and information regarding insurance status at more than 1 time point throughout the year, as well as document secondary insurance coverage. There is a possibility that selection bias could have occurred if some individuals were incorrectly identified with or without CP. However, this sample's distribution of children in each functional level is similar to other population-based studies of children with CP,23,24,31,32 suggesting that only a small amount of selection bias, if any, could have occurred.
Regardless of its limitations, this is the first large study of individuals with CP that examined physical therapy services received across functional levels as described by the GMFCS, age, race, sex, and type of insurance. Past studies were not specific to physical therapy or did not use the GMFCS to classify function. Future studies of trends in physical therapy services received are warranted, especially as individuals with CP transition to adulthood. Their findings will inform families and provide information for the development of policies for care providers, educators, and payers to improve services for individuals with CP.
Conclusion
This is the first study to use a large sample of individuals with CP to examine the relationship between physical therapy services received and GMFCS level. It is the first to assess the relationship between demographic variables, such as age, sex, and race, and physical therapy services received. Overall, results support a significant difference in physical therapy services received for individuals with private versus public insurance and a significant interaction between an individual's GMFCS level and age and the total physical therapy services received. We observed a spike in services received for individuals >18 years of age in this sample who are in levels II and III, but this finding must be interpreted with caution because of the small number of individuals at this level and age within this sample. Finally, the final model accounts for 19% of therapy use variation, suggesting that other variables contribute to the variation in physical therapy services received and should be explored in future studies utilizing mixed-method designs.
Footnotes
-
Ms Bailes provided concept/idea/research design, writing, data collection, project management, fund procurement, study participants, facilities/equipment, and institutional liaisons. Both authors provided data analysis. Dr Succop provided consultation (including review of manuscript before submission).
-
The authors thank Michelle Wallace, BS, for her assistance with database management and Kari Dunning, PT, PhD, for her advice regarding study analysis.
-
Part of the data were presented in poster format at the Combined Sections Meeting of the American Physical Therapy Association; February 17–20, 2010; San Diego, California.
-
This work was partially supported by the Foundation for Physical Therapy Promotion of Doctoral Studies Scholarships awarded to Ms Bailes, 2009–2012.
- Received November 3, 2011.
- Accepted July 11, 2012.
- © 2012 American Physical Therapy Association
References
Issue highlights
