Variation by Diagnostic and Practice Pattern Groups in the Mobility Outcomes of Inpatient Rehabilitation Programs for Children and Youth.

Physical therapists providing services in inpatient rehabilitation settings conduct examinations and evaluations, determine diagnoses and prognoses, and intervene to alleviate symptoms, remediate movement restrictions (functional limitations), and prevent disability.[1] Individualized intervention programs are designed to promote childrens' and youths' return to their age-expected social roles after hospitalization.

The Guide to Physical Therapist Practice[1] (the Guide) describes physical therapist practice and can aid patient/client management. The Guide is supposed to enhance positive outcomes of physical therapy practice and serve as a framework for organizing questions for clinical research.[1] As clinicians and administrators involved in pediatric inpatient programs explore the degree to which intervention programs assist in reaching anticipated outcomes, grouping children into meaningful categories for aggregate analyses will be useful. The practice patterns within the Guide may be helpful in categorizing children and judging whether outcomes have been met within or across programs. The Guide describes patient/client management processes through a series of preferred practice patterns. Practice patterns are defined for 4 major patient/client groups: (1) musculoskeletal, (2) neuromuscular, (3) cardiopulmonary, and (4) integumentary. In an inpatient pediatric rehabilitation hospital, the most common preferred practice pattern groups are musculoskeletal and neuromuscular. Limited information is available to date as to the benefit of the practice patterns in providing a structure for outcomes description and analyses.

We believe that functional outcome measurements, including an assessment of mobility, are a fundamental element of a comprehensive pediatric rehabilitation program. Assessment of function as part of a clinical examination can identify a child's capabilities in the hospital setting and becomes an indicator of the skills that will be carried over into the discharge setting.[2] Documenting gains in mobility within groups using a uniform functional measure allows program clinicians to evaluate mobility outcomes, develop high-quality improvement initiatives, compare findings from other programs, and meet accreditation requirements for performance measures and benchmarking.[3-5] Some authors recommend using functional outcome measures to evaluate pediatric programs and suggest that such measures should be broad in scope, should reflect the goals of the program, and should be sensitive to the characteristics of the children being served.[6-8]

In the Guide, the use of the Nagi model of disablement[9,10] is recommended for the description of outcomes within physical therapy. Using this model, functional mobility limitations are defined as restrictions in the ability to perform actions that promote movement and mobility in an efficient and age-appropriate manner. Disability is the lack of ability to take part in the age-expected movement functions that are expected within the child's environment of home, school, and community. Definitions of functional limitations and disability similar to those of the Nagi model have been retained in recent publications by the National Center for Medical Rehabilitation Research[11] and the Institute of Medicine.[12] At least one group of authors[13] contend that the National Center for Medical Rehabilitation Research model should be used as the framework for the description of outcomes across a broad variety of pediatric clinical programs.

In most outcome studies examining pediatric brain injury, including studies of children with traumatic brain injury, the time since injury, rather than hospital discharge, is used as a follow-up point. Thus, very little outcome information is available to interpret changes that take place during the inpatient hospital episode. Jaffe and colleagues,[14,15] for example, followed children at 6 months, 1 year, and 3 years postinjury, but did not report mobility outcomes at the time of hospital discharge. Surprisingly, we were able to find only 2 published studies in which variables relating to mobility were used as outcome measures for children with acquired brain injury at the time of discharge from inpatient rehabilitation. Vander Schaaf et al[2] studied the outcomes of children with traumatic brain injury and anoxia, and Philip et al[16] studied the outcomes of children who had brain tumors. Other researchers examining the recovery of children with traumatic brain injury during inpatient rehabilitation did not report on functional mobility outcomes.[17,18]

The variations of functional mobility achieved during inpatient rehabilitation are an element of program outcome evaluation.[19,20] Children receiving inpatient rehabilitation services vary as to types and severity of problems. Groups of children, therefore, will differ in the level and amount of change in mobility associated with inpatient rehabilitation. The level is the average of admission and discharge mobility scores, which we believe provides an estimate of the overall mobility status of the child. The amount of change expresses the difference in mobility status from admission to discharge. For outcomes analyses, we contend that it is useful to describe the variation in groups of children by use of both of these 2 variables.

Most pediatric researchers have chosen to illustrate variation in groups by use of traditional diagnostic categories. Children are often classified as having disorders based on neurological, orthopedic, or other conditions such as cardiopulmonary illnesses, burns, and infections. The Guide[1] provides a potentially new way to categorize variation in outcomes for pediatric patients. By understanding variation in outcomes, program administrators can more readily make changes designed to improve care. We compare the use of diagnostic and practice pattern groupings to describe the variation (level and amount of change) in functional mobility for children in an inpatient rehabilitation program. We predict that children within subgroups based on diagnosis and practice patterns will demonstrate different levels and amounts of change in mobility during inpatient rehabilitation and that we will be able to document this variation.

Method

Subjects

Through a search of the Franciscan Children's Hospital and Rehabilitation Center (FCH) database and by confirmatory chart review, we identified all children and adolescents who were admitted to the Physical Rehabilitation Program beginning in August 1994 and discharged by April 1997. This period of time was selected to coincide with an assessment of records for a 3-year period completed as an internal program quality review. Children were included in the study if they were admitted to the rehabilitation program for at least a 3-day period. Three children were excluded because they had incomplete discharge data. Another 5 children were excluded because they had conditions that placed them in practice pattern groups with very small samples (cardiopulmonary, n = 3; integumentary, n = 2). Our final study sample consisted of 138 children and youths between 1 and 22 years of age. Table 1 shows the practice pattern groups that were included in this study. The overall demographic characteristics are shown in Table 2, as well as the characteristics categorized by diagnosis. The characteristics of the sample organized by practice pattern group are given in Table 3.

Table 1.
Preferred Practice Pattern Groups Included in Study(1)

Preferred Practice Pattern Groups   Pediatric Conditions

Musculoskeletal

  Pattern G: Impaired joint         Traumatic injuries (lower
    mobility, muscle performance,     extremity)
    and range of motion
    associated with fracture

  Pattern I: Impaired joint         Open-reduction internal fixation
    mobility, motor function,         (ORIF), fusions, osteotomies,
    muscle performance, and range     external fixators, rod
    of motion associated with         procedures, multiple fractures,
    bony or soft tissue surgery       fascial release procedures,
                                      soft tissue realignment, muscle/
                                      tendon/ligament repair or
                                      reconstruction

  Pattern J: Impaired motor         Amputation
    function, muscle performance,
    range of motion, gait,
    locomation, and balance
    associated with amputation

Neuromuscular

  Pattern C: Impaired motor         Traumatic brain injury, anoxic
    function and sensory              brain injury, brain tumors,
    integrity associated with         cerebral hemorrhages, seizures
    nonprogressive disorders of       (surgical and nonsurgical),
    the central nervous               infectious disease that affects
    system--congenital origin or      the central nervous system
    acquired in infancy or            (meningitis, encephalitis, human
    childhood                         immunodeficiency virus)

  Pattern G: Impaired motor         Guillain-Barre syndrome,
    function and sensory              mitochondrial myopathy
    integrity associated with
    acute or chronic
    polyneuropathies

  Pattern H: Impaired motor         Traumatic spinal cord injury,
    function, peripheral nerve        nontraumatic spinal cord injury
    integrity, and sensory            (including benign spinal tumors)
    integrity associated with
    nonprogressive disorders of
    the spinal cord

  Pattern I: Impaired arousal,      Anoxic brain injury, traumatic
    range of motion, and motor        brain injury, cerebral infarct,
    control associated with coma,     brain tumor, seizures (surgical
    near coma, or vegetative          and nonsurgical)
    state
Table 2.
Demographic Variables by Diagnostic Group(a)

                                          Traumatic   Nontraumatic
                                          Brain       Brain
                           Overall        Injury      Injury
                           (N=138)        (n=49)      (n=50)

Males (N, %)               84; 60.9%      30; 61.2%   33; 66.0%

Ethnic group (N, %)
  White                    93, 67.4%      36, 73.5%   33, 66.0%
  Black                    17, 12.3%       7, 14.3%    6, 12.0%
  Hispanic                 17, 12.3%       6, 12.2%    3, 6.0%
  Other                    11, 8.0%        0, 0.0%     8, 16.0%

Age at admission (y)        9.4, 5.2       9.8, 0.7    8.1, 0.8

Length of inpatient        55.1, 4.4      49.6, 6.6   66.9, 8.9
  stay (d)

Admission PEDI             33.8, 2.4      35.5, 4.4   32.0, 3.9
  "Functional Skills"
  scale mobility score

Admission PEDI             32.8, 2.4(*)   33.4, 4.4   30.1, 4.1(**)
  "Caregiver Assistance"
  scale mobility score

                                        Neuro-
                           Orthopedic   logical
                           (n=16)       (n=23)      P

Males (N, %)               10; 62.5%    11; 47.8%   NS

Ethnic group (N, %)
  White                     8, 50.0%    16, 69.6%   NS
  Black                     4, 25.0%     0, 0.0%
  Hispanic                  3, 18.8%     5, 21.7%
  Other                     1, 6.3%      2, 8.7%

Age at admission (y)       13.5, 1.0    8.3, 1.1    .002

Length of inpatient        41.4, 10.3   50.5, 8.8   NS
  stay (d)

Admission PEDI             33.7, 5.0    33.9, 5.6   NS
  "Functional Skills"
  scale mobility score

Admission PEDI             39.5, 5.6    32.4, 5.7   NS
  "Caregiver Assistance"
  scale mobility score

(a) All values are mean and standard error of mean unless otherwise
noted. NS=not significant, PEDI=Pediatric Evaluation of Disability
Inventory. Asterisk indicates score was based on 137 responses; double
asterisk indicates score was based on 49 responses.
Table 3.
Demographic Variables by Practice Pattern Group(a)

                           NM-C(b)        NM-G(c)      NM-H(d)
                           (n=89)         (n=8)        (n=11)

Males (N, %)               54, 60.7%       2, 25.0%     7, 63.6%

Ethnic group (N, %)
  White                    59, 66.3%       5, 62.5%    10, 90.9%
  Black                    12, 13.5%       0, 0.0%      0, 0.0%
  Hispanic                  8, 9.0%        3, 37.5%     1, 9.1%
  Other                    10, 11.2%       0, 0.0%      0, 0.0%

Age at admission (y)        8.8, 0.6       8.6, 1.3     9.3, 1.5

Length of inpatient        55.3, 5.8      44.1, 16.1   64.6, 12.5
  stay (d)

Admission PEDI             38.3, 3.1      40.4, 7.2    27.8, 7.6
  "Functional Skills"
  scale mobility score

Admission PEDI             36.5, 3.1(g)   37.1, 9.0    29.1, 8.0
  "Caregiver Assistance"
  scale mobility score

                           NM-I(e)      MS-all(f)
                           (n=14)       (n=16)       P

Males (N, %)               11, 78.6%    10, 62.5%    NS

Ethnic group (N, %)
  White                    11, 78.6%     8, 50.0%    NS
  Black                     1, 7.1%      4, 25.0%
  Hispanic                  2, 14.3%     3, 18.8%
  Other                     1, 6.3%      1, 6.3%

Age at admission (y)        8.3, 1.4    13.5, 1.0    .01

Length of inpatient        68.2, 15.1   41.4, 10.3   NS
  stay (d)

Admission PEDI              6.2, 1.2    33.7, 5.0    .01
  "Functional Skills"
  scale mobility score

Admission PEDI              1.7, 1.1    39.5, 5.6    <.001
  "Caregiver Assistance"
  scale mobility score

(a) All values are mean and standard error of the mean unless
otherwise noted. NS=not significant, PEDI=Pediatric Evaluation of
Disability Inventory.

(b) Neuromuscular practice pattern C (congenital and acquired central
nervous system disorders).

(c) Neuromuscular practice pattern G (polyneuropathies).

(d) Neuromuscular practice pattern H (spinal cord injuries).

(e) Neuromuscular practice pattern I (coma/near coma).

(f) Musculoskeletal practice patterns (all).
(g) n = 88.

The program administrator (HMD) classified each child into a diagnostic group. The program administrator had been program director for 5 years and had been involved with organizing the program evaluation system. The diagnostic groupings were internal program classifications based on International Classification of Diseases, 9th Revision,[21] (ICD-9) codes for 4 major clinical groups: traumatic brain injury, nontraumatic brain injury, orthopedic conditions (mainly post-operative lower-extremity surgery), and a series of conditions without brain injury (neurological) affecting the neurological system, such as spinal cord injury and neuropathies. Children with traumatic brain injury included those with neurological or motor sequelae resulting from physical trauma. Children with nontraumatic brain injury included those with conditions such as stroke (infarct or hemorrhage), tumors, seizures (surgical and nonsurgical), anoxia, or infections. The numbers of children in the nontraumatic category were: stroke (n = 11), tumor (n = 11), seizures (n = 11), anoxia (n = 5), and infection (n = 12). Classification as to the physical therapy preferred pattern[1] was done retrospectively by consensus of the authors (SMH, HMD) and was based on the initial conditions present at admission. The first author (SMH) has over 20 years of experience as a pediatric physical therapist, and the second author (HMD) is a board-certified specialist in pediatric physical therapy with 16 years of experience.

Assessment

The Pediatric Evaluation of Disability Inventory (PEDI)[22] is a standardized pediatric functional assessment commonly used in inpatient hospital settings to evaluate functional change. The PEDI has been standardized for children between the ages of 6 months and 7 1/2 years, but the instrument is often used to assess the performance of older children and adolescents who demonstrate basic motor functioning during recovery. The PEDI was initially validated using a normative sample (N = 412) and on children receiving rehabilitation services due to brain injury and developmental delay.[22] Content validity was established using a panel of 31 experts in the field of pediatric rehabilitation.[23] The range of values for inter-interviewer reliability for the clinical standardization sample has been reported to be between .84 and .99.[22] Reliability studies have demonstrated what we consider good interrater reliability[24] and intrarater reliability.[22,24,25] Concurrent validity has been established with other pediatric measures such as the Peabody Developmental Motor Scales[25] and the Functional Independence Measure for Children.[22] Construct validity studies have shown that the PEDI can be used to discriminate between children with and without disabilities[25,26] and among levels of severity in children with osteogenesis imperfecta.[27,28] Researchers have reported good responsiveness of the PEDI in children with spastic cerebral palsy who were treated with selective posterior rhizotomy[29] and in children with traumatic brain injury followed in the community 6 months after injury.[30] Haley[31] provides a summary of technical information on the PEDI. We believe that physical therapists in an inpatient rehabilitation hospital often use the "Mobility" domain of the PEDI as a standardized outcome instrument to detect changes in functional limitations and disability.

The PEDI assesses 2 distinct mobility constructs. The "Functional Skills" scale is used to assess a child's capability to perform basic mobility activities that are considered to be part of important daily functional skills used within the hospital setting. The "Functional Skills" scale examines functional limitations as defined by Nagi.[10] "Caregiver Assistance," the second level included in the PEDI, is used to assess the amount of help required by the child to do complex (multi-step) mobility activities. This level addresses the extent to which children are able to conduct mobility tasks that allow them to independently function in the environment and is consistent with the level of disability described in the Guide.[1]

"Functional Skills" scale. The "Functional Skills" scale of the "Mobility" domain of the PEDI consists of 59 dichotomous items. Children are scored as either "capable" (ie, capable of performing the item in most situations) or "unable" (ie, unable to perform the item in most situations). In order to be scored as "capable" on any item, the child must demonstrate a consistent ability in the context of daily routines, in this case, the hospital setting. The "Functional Skills" items are organized into 13 clusters, including toilet transfers, chair or wheelchair transfers, car transfers, bed mobility and bed transfers, tub transfers, method of indoor locomotion, indoor locomotion distance and speed, carrying objects, method of outdoor locomotion, outdoor locomotion and speed, ambulation over outdoor surfaces, and walking up and down stairs. Scaled summary scores for the "Functional Skills" scale are calculated by summing the scores for items the child is capable of performing, then transforming the summated score to a score of 0 to 100, with higher scores indicating greater capability. Transformed scores are based on a model of item difficulty calibrations, as defined by the original standardization sample. In the PEDI, a form of item response theory method is used to place each of the 59 items on a location that provides an estimate of the relative degree of difficulty for each item on the same metric as the 0-100 summary scoring system.[32,33]

"Caregiver Assistance" scale. The "Caregiver Assistance" scale of the "Mobility" domain of the PEDI consists of 7 items. This scale is used to measure the amount of help needed to carry out functional mobility tasks. Items are scored on a 6-point scale with scores ranging from "independence" to "total assistance." The 7 "Garegiver Assistance" items are: chair/toilet transfers, car transfers, bed mobility/transfers, tub transfers, indoor locomotion, outdoor locomotion, and stairs. Scaled summary scores for "Caregiver Assistance" scale are calculated by summing the ratings for the 7 items, then transforming the summated score to a score of 0 to 100, with higher scores indicating greater independence.

Procedure

The Physical Rehabilitation Program at FCH is designed for children and adolescents with neurological or musculoskeletal impairments of recent onset (less than 30 days), regression, or progression and who have a prognosis for improvement in the areas of self-care, mobility, safety, communication, cognition, and behavior. Prognosis is determined by the attending physician based on: time since injury, severity, the child's past medical history, comorbidities, and response to intervention in the acute care setting. In addition to physical therapists, the core inpatient treatment team at FCH includes pediatric physiatrists, rehabilitation nurses, occupational therapists, speech-language pathologists, therapeutic recreation specialists, special education teachers, dietitians, school liaisons, social workers, and case managers. The physiatrist works with the child and his or her family or guardian and with the treatment team to develop and implement a rehabilitation program based on the individual needs of the child and family. Children and adolescents are provided a structured environment with regulated stimulation and a consistent approach to behavior management and enhancement of memory and cognitive skills.

Mobility activities are a focus of each child's or youth's plan of care. Children receive a minimum of 3 hours per day of rehabilitation services, typically including physical therapy twice per day for 45 to 60 minutes per session. Intervention by physical therapists at FCH is based on a physical therapist's examination and may include therapeutic exercise; functional training in self-care and home management; community and work (job/school/ play) integration or reintegration; prescription, application, and fabrication of devices and equipment; electrotherapeutic modalities; physical agents; and mechanical modalities.

Since August 1994, the PEDI has been used at the time of admission and at all planned discharges to assess the functional performance of children and adolescents admitted to the Physical Rehabilitation Program at FCH. The physical therapy, occupational therapy, and speech therapy staff administers all 3 domains of the PEDI ("Self-Care," "Mobility," and "Social Function"). The therapists at FCH were trained in the use of the PEDI through a formal instructional course given by 2 of the PEDI's developers and through periodic staff training. Case studies in the PEDI manual[22] are used for orientation of all new staff. Therapists had attended in-service training sessions and had completed the case studies in the PEDI manual[22] in preparation for administering the PEDI in clinical practice. In this report, results from only the mobility assessments are described.

Mobility data were collected by each child's physical therapist, who administered the "Mobility" domain of the PEDI within 3 days of admission and approximately 1 to 2 days before discharge. The physical therapist observed the child's mobility skills in the hospital setting and obtained additional information from parents and other hospital staff about the child's present mobility performance. Physical therapists contributed to all sections of the PEDI, but were responsible for administering and completing the "Mobility" domain of the PEDI at the time of admission and at the time of discharge. The same physical therapist administered the admission and discharge assessments of the PEDI. After the admission examination and the discharge re-examination, the raw and scaled PEDI scores were entered in a program database for analysis. In these analyses, admission and discharge data were available for 93% (n = 138) of the intended sample.

Data Analysis

Differences in demographic variables across the diagnostic and practice pattern groups were tested by one-way analyses of variance (ANOVAs) or chi-square analyses as appropriate. For each PEDI "Mobility" domain scale, total individual raw scores were calculated and transformed to scaled scores using the tables in the PEDI manual.[22] In order to examine the variation in mobility level and change in the diagnostic and practice pattern groups, individual repeated-measures ANOVAs were performed for each dependent variable ("Functional Skills" and "Caregiver Assistance" scales) with separate between and within factors. One-tailed paired t tests with a Bonferroni correction to adjust the critical values were performed to examine within-group changes between admission and discharge scores for the diagnostic and practice pattern groups. One-way ANOVAs were used for the change scores for the diagnostic and practice pattern groups to determine whether groups differed in the amount of change. Planned contrasts were conducted with the Dunnett multiple-comparison test,[34] using specific contrast groups. This was done to avoid random comparisons and to minimize the possibility of a Type I error. For the diagnostic groups, traumatic brain injury was identified as the contrast group because it is normally the largest diagnostic group in inpatient pediatric rehabilitation. For the practice patterns, the neuromuscular practice pattern C (congenital and acquired central nervous system disorders) was designated as the contrast group. This group was by far the largest (64.5%) of the practice pattern groups within this sample. All tests were conducted with an alpha level of .05 unless otherwise noted.

Results

Demographic Data

Group comparisons with the diagnostic categories indicated that there were differences among the groups in age, as the average age of the children in the orthopedic group ([bar]X = 13.5 years) was greater than the average ages of the children in the other 3 groups. No differences were noted in other demographic variables, length of stay, or admission PEDI scores using the diagnostic categories (Tab. 3). Children in the combined musculoskeletal practice pattern group were older than children in the other practice pattern groups. Admission PEDI scores (both "Functional Skills" and "Caregiver Assistance" scale scores) for the neuromuscular practice pattern I (coma or near coma) were lower than for the other 4 practice pattern groups.

Diagnostic Groups

Between-group (level = average of admission and discharge) comparisons across admission and discharge tests indicated that there no were differences across diagnostic groups for the "Functional Skills" scale or the "Caregiver Assistance" scale. Within-group comparisons identified a large admission-discharge effect across all diagnostic groups for the "Functional Skills" scale (F = 131.57, P = .001) and the "Caregiver Assistance" scale (F = 131.57, P = .001). In addition, changes between admission and discharge tests were identified for each diagnostic group for the "Functional Skills" scale (range = 18.3-34.4) and "Caregiver Assistance" scale (range = 17.3-34.1). Differences in the magnitude of change scores were found across the diagnostic groups for the "Functional Skills" scale (range = 6.2-11.4) (F = 4.81, P = .003) and the "Caregiver Assistance" scale (range = 1.5-11.7) (F = 4.55, P = .005). Planned contrasts on "Functional Skills" scale change scores showed that the average amount of change between admission and discharge in the nontraumatic brain injury diagnostic group ([bar]X = 18.3) and the neurological diagnostic group ([bar]X = 19.6) was less than in the traumatic brain injury diagnostic group ([bar]X = 34.4). Similarly, planned contrasts on "Caregiver Assistance" scale change scores showed that the average amount of change between admission and discharge in the nontraumatic brain injury diagnostic group ([bar]X = 17.3) and the neurological diagnostic group ([bar]X = 20.5) was less than in the traumatic brain injury diagnostic group ([bar]X = 34.1). In Tables 4 and 5, these results are summarized for the "Functional Skills" and "Caregiver Assistance" scales by reporting the means and standard deviations for the admission and discharge scores, the mean mobility level for each diagnostic group, the specified post hoc comparisons in levels between the contrast group and other groups, the mean changes between the admission and discharge data, and the specified post hoc comparisons in change scores between the contrast group and other groups. Figure 1 illustrates the admission and discharge PEDI scores for the diagnostic groups.

[GRAPH OMITTED]

Table 4.
Changes in Pediatric Evaluation of Disability Inventory "Functional
Skills" Scale Mobility Scores Between Time of Admission and Time of
Discharge by Diagnostic Group

                           Admission                Discharge
Diagnostic
Group          N    [bar]X   SD     Range    [bar]X   SD     Range

Traumatic      49    35.5    30.8   0-100     69.9    22.7   11.4-100
  brain
  injury
Nontraumatic   50    32.1    27.5   0-100     50.4    32.6      0-100
  brain
  injury
Orthopedic     16    33.7    19.9   0-68.7    59.2    11.9   37.1-79.8
Neurological   23    33.9    26.7   0-100     53.5    24.5    6.1-100

                              Repeated
                              Measures(a)

                            Group             Paired t Tests
                            Contrast
Diagnostic          Mean    (Level            Mean
Group          N    Level   Effect)(c)   P    Change   P

Traumatic      49   52.7    Contrast Group    34.4     .001
  brain
  injury
Nontraumatic   50   41.3      -11.4      NS   18.3     .001
  brain
  injury
Orthopedic     16   46.5       -6.2      NS   25.5     .001
Neurological   23   43.7       -9.0      NS   19.6     .001

                    One-Way
                    ANOVA(b)

                    Group
                    Contrast
Diagnostic          (Change
Group          N    Scores)(c)   P

Traumatic      49   Contrast Group
  brain
  injury
Nontraumatic   50     -16.1      .001
  brain
  injury
Orthopedic     16      -8.9      NS
Neurological   23     -14.8      .03

(a) Within-subject effect: F=131.74, P<.0005; between-subject effect:
F=1.89, P=.13. NS=not significant.

(b) ANOVA=analysis of variance: F=4.81, P=-.003.

(c) All post hoc comparisons were 2-tailed.
Table 5.
Changes in Pediatric Evaluation of Disability Inventory "Caregiver
Assistance" Scale Mobility Scores Between Time of Admission and Time of
Discharge by Diagnostic Group

                          Admission                Discharge
Diagnostic
Group          N    [bar]X   SD     Range    [bar]X   SD     Range

Traumatic      49    33.4    31.1   0-100     67.5    22.5      0-100
  brain
  injury
Nontraumatic   50    30.1    28.4   0-100     47.4    30.7      0-100
  brain
  injury
Orthopedic     16    39.5    22.5   0-70.5    64.4    14.3   42.7-100
Neurological   23    32.4    27.5   0-82.5    52.9    26.6      0-89.4

                                 Repeated
                                 Measures(a)

                               Group             Paired t Tests
                               Contrast
Diagnostic             Mean    (Level            Mean
Group          N       Level   Effect)(c)   P    Change   P

Traumatic      49      50.5    Contrast Group    34.1     .001
  brain
  injury
Nontraumatic   50      38.8      -11.7      NS   17.3     .002
  brain
  injury
Orthopedic     16      52.0        1.5      NS   24.9     .001
Neurological   23      42.7       -7.8      NS   20.5     .001

                    One-Way
                    ANOVA(b)

                    Group
                    Contrast
Diagnostic          (Change
Group          N    Scores)(c)   P

Traumatic      49   Contrast Group
  brain
  injury
Nontraumatic   50     -16.8      .001
  brain
  injury
Orthopedic     16      -9.2       NS
Neurological   23     -13.6      .048

(a) Within-subject effect: F=124.98, P<.0005; between-subject effect:
F=2.21, P=.089. NS=not significant.

(b) ANOVA=analysis of variance: F=4.55, P=.005.

(c) All post hoc comparisons were 2-tailed.

Practice Pattern Groups

Between-group comparisons across admission and discharge scores indicated that there were differences across practice pattern groups for the levels of the "Functional Skills" scale (F = 7.71, P = .005) and the "Caregiver Assistance" scale (F = 7.73, P = .005). Planned contrasts on the levels showed that there were differences between the neuromuscular practice pattern I group and the neuromuscular practice pattern C group for both the "Functional Skills" scale (35.1) and the "Caregiver Assistance" scale (34.7). Within-group comparisons identified a large admission-discharge effect across all practice pattern groups for the "Functional Skills" scale (range = 20.1-26.2) (F = 70.94, P = .0005) and the "Caregiver Assistance" scale (range = 20.3-29.4) (F = 75.31, P = .0005). In addition, changes between admission and discharge tests were identified for each group for both the "Functional Skills" and "Caregiver Assistance" scale (Tabs. 6 and 7). However, no differences were found in the amount of change between the target group (neuromuscular practice pattern C) and the other practice pattern groups for either the "Functional Skills" scale or the "Caregiver Assistance" scale. Figure 2 illustrates the admission and discharge PEDI scores for the practice pattern groups.

[GRAPH OMITTED]

Table 6.
Changes in Pediatric Evaluation of Disability Inventory "Functional
Skills" Scale Mobility Scores Between Time of Admission and Time of
Discharge by Practice Pattern Group

Practice                  Admission              Discharge
Pattern
Group(a)       N    [bar]X   SD     Range      [bar]X   SD

Neuro-         89   38.3     29.4      0-100   64.5     27.0
  muscular C
Neuro-          8   40.4     20.4   6.1-58.2   65.9      7.8
  muscular G
Neuro-         11   27.8     25.3     0-89.2   49.1     22.6
  muscular H
Neuro-         14    6.2      4.4     0-15.2   26.3     29.8
  muscular I
Musculo-       16   33.7     19.9     0-68.7   59.9     27.3
  skeletal

                                      Repeated
                                     Measures(b)

                                   Group               Paired t Tests
Practice       Discharge           Contrast
Pattern                    Mean    (Level              Mean
Group(a)       Range       Level   Effect)(d)    P     Change    P

Neuro-           6.1-100   51.4     Contrast Group     26.2     .001
  muscular C
Neuro-         56.5-75.2   53.2       -1.8       NS    25.5     .005
  muscular G
Neuro-          11.4-100   38.5      -13.0       NS    21.3     .001
  muscular H
Neuro-             0-100   16.3      -35.1      .001   20.1     .031
  muscular I
Musculo-       37.1-79.8   46.8       -4.6       NS    26.2     .001
  skeletal

                   One-Way
                   ANOVA(c)

               Group
Practice       Contrast
Pattern        (Change
Group(a)       Scores)(d)   P

Neuro-         Contrast Group
  muscular C
Neuro-         -0.7         NS
  muscular G
Neuro-         -4.9         NS
  muscular H
Neuro-         -6.1         NS
  muscular I
Musculo-        0           NS
  skeletal

(a) See Table 3 footnotes for descriptions of practice pattern groups.

(b) Within-subject effect: F = 70.94, P<.0005; between-subject
effect: F = 7.71, P<.0005. NS = not significant.

(c) ANOVA = analysis of variance: F=0.28, P=.89.

(d) All post hoc comparisons were 2-tailed.
Table 7.
Changes in Pediatric Evaluation of Disability Inventory "Caregiver
Assistance" Scale Mobility Scores Between Time of Admission and Time of
Discharge by Practice Pattern Group

Practice                   Admission           Discharge
Pattern
Group(a)       N    [bar]X   SD     Range    [bar]X   SD

Neuro-         89   36.5     29.4   0-100    61.2     26.3
  muscular C
Neuro-          8   37.1     25.5   0-63.3   66.5      8.4
  muscular G
Neuro-         11   29.1     26.4   0-72.7   49.4     26.5
  muscular H
Neuro-         14    1.7      4.2   0-11.7   26.6     31.1
  muscular I
Musculo-       16   39.5     22.5   0-70.5   64.4     14.3
  skeletal

                                      Repeated
                                     Measures(b)

                                   Group               Paired t Tests
Practice       Discharge           Contrast
Pattern                    Mean    (Level              Mean
Group(a)       Range       Level   Effect)(d)    P     Change    P

Neuro-             0-100   48.9     Contrast Group     24.7     .001
  muscular C
Neuro-         56.1-78.3   51.8      2.7         NS    29.4     .007
  muscular G
Neuro-         3.9-89.4    39.3     -9.6         NS    20.3     .001
  muscular H
Neuro-             0-100   14.2    -34.7        .001   24.9     .011
  muscular I
Musculo-        42.7-100   52.0      3.1         NS    24.9     .002
  skeletal

                   One-Way
                  ANOVA(c)

               Group
Practice       Contrast
Pattern        (Change
Group(a)       Scores)(d)   P

Neuro-         Contrast Group
  muscular C
Neuro-          4.7         NS
  muscular G
Neuro-         -4.4         NS
  muscular H
Neuro-          0.2         NS
  muscular I
Musculo-        0.2         NS
  skeletal

(a) See Table 3 footnotes for descriptions of practice pattern groups.

(b) Within-subject effect: F=75.31, P<.0005; between-subject effect:
F=7.73, P<.0005. NS=not significant.

(c) ANOVA=analysis of variance: F=0.18, P=.95.

(d) All post hoc comparisons were 2-tailed.

Discussion

As we predicted, both the diagnostic and practice pattern groupings were effective in allowing for descriptions of within-group change between admission and discharge scores. In general, greater variability in the level was detected using the practice pattern groups, whereas variation in amount of change was more evident using the diagnostic groups.

The arrangement of children's outcomes by practice patterns yielded a difference among levels, whereas there was no difference in levels when children were classified by diagnostic group. This result was most likely due to the relatively low level of motor functioning (for both "Functional Skills" and "Caregiver Assistance" scales) at the time of admission and discharge for the children in the neuromuscular practice pattern I (coma or near coma) group. Even though the level was lower for the children in the neuromuscular practice pattern I group than for the target group (neuromuscular practice pattern C), the amount of change for children who were admitted in coma or near coma ([bar]X = 20.1 for the "Functional Skills" scale and [bar]X = 24.9 for the "Caregiver Assistance" scale) was not different than the amount of change for the target group ([bar]X = 26.2 for the "Functional Skills" scale and [bar]X = 24.7 for the "Caregiver Assistance" scale).

Greater variability in the amount of change across groups was detected using the diagnostic groupings than the practice pattern groupings. When applying the practice pattern groupings, most of the children in the traumatic brain injury and nontraumatic brain injury diagnostic groups were combined in neuromuscular practice pattern C group, thus diluting the effect of any large changes that could have occurred in the traumatic brain injury diagnostic group. Because of the relatively large amount of recovery of motor skills and the gains in independence seem across all the children, there were within-group changes as defined by the diagnostic groups and practice patterns. A 20-point or more change in the PEDI scores reflects a noticeable improvement in mobility capability and level of independence based on item map calibrations.[22] For example, a 20-point change in "Functional Skills" scale scores at the middle of the score range is indicative of a child who, on admission, can walk for very short distances indoors (within a room) to being able to walk outdoors without support at the time of discharge.

By separating children with brain injuries who had traumatic versus nontraumatic brain etiologies, we were able to demonstrate variation in the amount of change across groups. Children in the traumatic brain injury diagnostic group changed the most and more than children in the nontraumatic brain injury and neurological diagnostic groups. These results are consistent with those of Vander Schaaf and colleagues,[2] who found better motor outcomes in children with traumatic brain injury than in children with nontraumatic injuries. These findings may be related to the often nonreversible and progressive nature of nontraumatic illnesses seen in inpatient rehabilitation.

Differences in level of motor function (lower for the neuromuscular practice pattern I grouping than for the neuromuscular practice pattern C group) and in the amount of functional gain (greater in the traumatic brain injury diagnostic group than in the nontraumatic brain injury and neurological diagnostic groups), we believe, should be interpreted with caution because of initial differences in the group demographics. The average age of children in the combined orthopedic and musculoskeletal practice pattern group (same 16 children) was greater than for the other diagnostic or practice pattern groups, respectively. A ceiling effect of the PEDI for older children may be an alternative explanation for the findings. However, no differences existed in the admission functional scores between the combined orthopedic and musculoskeletal practice pattern group and the comparison groups (traumatic brain injury and neuromuscular practice pattern C groups). In these analyses, we used the PEDI scaled scores, which are on a criterion scale of 0 to 100 and are age-independent. With larger samples, adjusting for age statistically could minimize age differences as a possible explanation for differences in outcomes.

All of the practice pattern groups registered about the same amount of change in both the "Functional Skills" and "Caregiver Assistance" scales, even though the 2 PEDI mobility domains represent distinct constructs of functional limitations and disability, respectively. The "Functional Skills" scale provides information regarding the recovery of simple motor activities, whereas the "Caregiver Assistance" scale identifies increasing independence in moving around the environment and in transfers. Children in all diagnostic and practice pattern groups made meaningful gains in motor recovery and independence from admission to discharge. Although we had expected to see certain groups, such as children with spinal cord injury, make more gains in independence (disability) than functional skills capability, we found no evidence to support this view.

A potential limitation in the use of the practice pattern groupings for inpatient pediatrics appears to be the large percentage of children who fall into the neuromuscular practice pattern C grouping. In our sample, nearly 65% of the children and adolescents fit into the neuromuscular practice pattern C grouping. Because this practice pattern grouping includes children with both acquired and congenital central nervous system problems, the current practice pattern structure may not provide sufficient differentiation of groups for interpretation and analyses of future pediatric outcome studies in inpatient rehabilitation programs. One possible option is to define subgroups of children (eg, children with acquired and congenital central nervous system problems) within the neuromuscular practice pattern C grouping to describe outcomes within this large group of children who use inpatient rehabilitation services.

Although the preferred practice patterns in the Guide[1] are intended to be applicable to children and adults, only 2 of the 34 practice patterns are intended specifically for children.[35] We did not include all of the practice patterns that might be applicable for children in an inpatient setting. We excluded children (n = 5) with burns (integumentary practice patterns) and cardiac difficulties (cardiopulmonary practice patterns) because of the small number of children in those categories. We chose to collapse the musculoskeletal preferred practice patterns into one group because of the relatively small number of children (n = 16) in this overall category. Studies using broader, more diverse samples of children across other settings will be needed to examine the benefits of individual musculoskeletal practice patterns and the full spectrum of practice patterns for outcome classification in pediatric physical therapy.

Assignment to practice pattern groups was done by 2 of the authors (SMH, HMD) in a retrospective manner in this study. This was necessary because the children were enrolled in the program before the Guide[1] was published. In addition, we classified the children based on information obtained at the time of admission. In clinical practice, children can move between patterns as they progress through their rehabilitation process and recovery. For example, some children who started in the neuromuscular practice pattern I (coma or near coma) group may have shifted to the neuromuscular practice pattern C group and then even perhaps to a musculoskeletal practice pattern, depending on the extent of orthopedic injuries and the focus of the rehabilitation program. In our study, we did not make changes in practice pattern assignments throughout the inpatient stay, because we relied only on the admission classification for grouping children.

We used the practice patterns to categorize children in order to understand variations in outcomes. In physical therapist practice, the practice patterns are intended to describe the entire spectrum of elements of patient/ client management, including examination, evaluation, diagnosis, prognosis, and intervention. These 5 elements can be specific to each pattern and are designed to help describe the processes within an episode of care. We do not intend to imply that, by categorizing children into one of the practice patterns for outcomes analyses, physical therapists in this study used any or all elements of each practice patterns for patient/client management. In our study, the processes of care used by the physical therapists, or how they differed across patterns, was not controlled or detailed. Ideally, in future outcome studies, researchers should combine the specific practice pattern elements with the corresponding outcomes of care.

Limitations in mobility are a major problem associated with childhood illness or injury or following complicated surgery. Measurement of functional mobility for children in inpatient pediatric rehabilitation programs can help document outcomes in order to better understand variation in results. Feedback to clinical staff based on aggregate functional data can assist physical therapists in determining which children are achieving expected and unexpected outcomes. According to Haley et al,[8] improving physical therapy services requires a consistent and systematic approach to documentation of motor recovery. Outcome analyses can prompt an examination of elements of the practice patterns, intervention approaches, decisions regarding resource allocation, and the appropriateness of admission criteria to the program.

Conclusion

Groups of children and adolescents in an inpatient rehabilitation program can show changes in functional mobility, regardless of whether they are grouped by traditional diagnostic categories or by preferred practice patterns. Variation in the amount of mobility change achieved by the groups of children is best seen when categorized by diagnostic groups. In contrast, categorizing children by preferred practice patterns at the time of admission highlights variation in the level of mobility function. In our sample of children in an inpatient rehabilitation setting, both categorical systems seem to have merit for understanding variation in mobility as an outcome of care.

References

[1] Guide to Physical Therapist Practice. 2nd ed. Phys Ther. 2001;81: 9-744.

[2] Vander Schaaf PJ, Kriel RL, Krach LE, Luxenberg MG. Late improvements in mobility after acquired brain injuries in children. Pediatr Neurol. 1997; 16:306-310.

[3] Hamilton BB, Granger CV. Disability outcomes following impatient rehabilitation for stroke. Phys Ther. 1994;74:494-503.

[4] Wilkerson DL, Johnston MV. Clinical program monitoring systems: current capability and future directions. In: Fuhrer MJ, ed. Assessing Medical Rehabilitation Practices: The Promise of Outcomes Research. Baltimore, Md: Paul H Brookes Publishing Co; 1997:1-41.

[5] Schyve PM. The evolving role of the Joint Commission for the Accreditation of Health Care Organizations. Joint Commission Journal on Quality Improvement. 1996;11(1):S54-S57.

[6] Haley SM, Cioffi MI, Lewin JE, Baryza MJ. Motor dysfunction in children and adolescents after traumatic brain injury. Journal of Head Trauma Rehabilitation. 1990;5(4):77-90.

[7] Haley SM. Motor assessment tools for infants and young children: a focus on disability assessment. In: Forssberg H, Hirschfeld H, eds. Movement Disorders in Children. Basel, Switzerland: Karger; 199:278-283.

[8] Haley SM, Baryza MJ, Webster HC. Pediatric rehabilitation and recovery of children with traumatic injuries. Pediatric Physical Therapy 1992;4:24-30.

[9] Nagi SZ. An epidemiology of disability among adults in the United States. Milbank Memorial Fund Quarterly. 1976;54:439-467.

[10] Nagi SZ. Disability concepts revisited: implications for prevention. In: Disability in America: Toward a National Agenda for Prevention. In: Pope AM, Tarlov AR, eds. Washington, DC: National Academy Press; 1991.

[11] NCMRR. Report and Plan for Medical Rehabilitation Research. Washington, DC: National Institute of Child Health and Human Development, National Institutes of Health; 1992.

[12] Pope AM, Tarlov AR, eds. Disability in America: Toward a National Agenda for Prevention. Washington, DC: National Academy Press; 1991.

[13] Butler C, Chambers H, Goldstein M, et al. Evaluating research in developmental disabilities: a conceptual framework for reviewing treatment outcomes. Dev Med Child Neurol. 1999;41:55-59.

[14] Jaffe KM, Fay GC, Polissar NL, et al. Severity of pediatric traumatic brain injury and neurobehavioral recovery at one year: a cohort study. Arch Phys Med Rehabil. 1993;74:587-595.

[15] Jaffe KM, Polissar NL, Fay GC, Liao S. Recovery trends over three years following pediatric traumatic brain injury. Arch Phys Med Rehabil. 1995;76:17-26.

[18] Philip PA, Ayyangar R, Vanderbilt J, Gaebler-Spira DJ. Rehabilitation outcome in children alter treatment of primary brain tumor. Arch Phys Med Rehabil. 1994;75:36-39.

[17] Boyer MG, Edwards P. Outcome 1 to 3 years after severe traumatic brain injury in children and adolescents, Injury. 1991;22:315-320.

[18] Bowen JM, Clark E, Bigler ED, et al. Childhood traumatic brain injury: neuropsychological status at the time of hospital discharge. Dev Med Child Neurol. 1997;39:17-25.

[19] Kane RL. Looking for physical therapy outcomes. Phys Ther. 1994; 74:425-429.

[20] Dobrzykowski E. The methodology of outcomes measurement. Journal of Rehabilitation Outcomes Measurement. 1997;1:8-17.

[21] Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death, Based on the Recommendations of the Ninth Revision Conference, 1975. Geneva, Switzerland: World Health Organization; 1977.

[22] Haley SM, Coster WJ, Ludlow LH, et al. Pediatric Evaluation of Disability Inventory: Administration Manual. Boston, Mass: New England Medical Center; 1992.

[23] Haley SM, Coster WJ, Faas RM. A content validity study of the Pediatric Evaluation of Disability Inventory. Pediatric Physical Therapy. 1991;3:177-184.

[24] Reid DT, Boschen K, Wright V. Critique of the Pediatric Evaluation of Disability Inventory (PEDI). Physical & Occupational Therapy in Pediatrics. 1993;13(4):57-87.

[25] Nichols DS, Case-Smith J. Reliability and validity of the Pediatric Evaluation of Disability Inventory. Pediatric Physical Therapy. 1996;8: 15-24.

[26] Feldman AB, Haley SM, Coryell J. Concurrent and construct validity of the Pediatric Evaluation of Disability Inventory. Phys Ther. 1990;70: 602-610.

[27] Engelbert RHH, Custers JWH, van der Net J, et al. Functional outcome in osteogenesis imperfecta: disability profiles using the PEDI. Pediatric Physical Therapy. 1997;9:18-22.

[28] Engelbert RHH, van der Graaf Y, van Empelen R, et al. Osteogenesis imperfecta in childhood: impairment and disability. Pediatrics. 1997;99:E3.

[29] Dudgeon BJ, Libby AK, McLaughlin JF, et al. Prospective measurement of functional changes after selective dorsal rhizotomy. Arch Phys Med Rehabil. 1994;75:46-53.

[30] Coster WJ, Haley SM, Baryza MJ. Functional performance of young children after traumatic brain injury: a 6-month follow-up study. Am J Occup Ther. 1994;48:211-218.

[31] Haley SM. The Pediatric Evaluation of Disability Inventory (PEDI). Journal of Rehabilitation Outcomes Measurement. 1997;1:61-69.

[32] Wright BD. Best Test Design. Chicago, Ill: MESA Press; 1978.

[33] Haley SM, Ludlow LH, Coster WJ. Pediatric Evaluation of Disability Inventory: clinical interpretation of summary scores using Rasch rating scale methodology. Physical Medicine Rehabilitation Clinics of North America. 1993;4:529-540.

[34] Glass GV, Hopkins KD. Statistical Methods in Education and Psychology. Baltimore, Md: Prentice-Hall; 1994.

[35] Palisano RJ, Campbell SK, Harris SR. Decision making in pediatric physical therapy. In: Campbell SK, Vander Linden D, Palisano RJ, eds. Physical Therapy for Children. Philadelphia, Pa: WB Saunders Co; 2000: 198-224.

SM Haley, PT, PhD, is Director, Center for Rehabilitation Effectiveness, Sargent College of Health and Rehabilitation Sciences, Boston University, 635 Commonwealth Ave, Boston, MA 02215-1605 (USA) ([e-mail]). Address all correspondence to Dr Haley.

HM Dumas, PT, MS, PCS, is Manager, Center for Excellence for Children With Special Health Care Needs, Franciscan Children's Hospital and Rehabilitation Center, Boston, Mass.

LH Ludlow, PhD, is Associate Professor, Lynch School of Education, Boston College, Chestnut Hill, Mass.

All authors provided concept/research design. Ms Dumas and Dr Ludlow provided writing. Ms Dumas provided data collection, subjects, facilities/equipment, and consultation (including review of manuscript before submission). Dr Ludlow provided data analysis. The authors thank the medical and clinical staff of the Inpatient Pediatric Physical Rehabilitation Program for their contributions in data collection and discussion of results. They also thank Maria Fragala, PT, Clinical Researcher, Franciscan Children's Hospital and Rehabilitation Center, for her helpful comments, and Maggie Foley, for her assistance in manuscript preparation.

The Institutional Review Board of Franciscan Children's Hospital and Rehabilitation Center approved this study.

The results of this study, in part, were presented at the Combined Sections Meeting of the American Physical Therapy Association, February 12-15, 1998, Boston, Mass.

This article was submitted August 21, 2000, and was accepted February 20, 2001.