Medical care costs are increasingly the focus of public attention and concern in the United States. Healthcare expenditures accounted for 13% of the United States’ gross domestic product in 1999 as compared with 8.8% in 1980 and 5.1% in 1960. 1 Concern over increasing medical care costs and the increase in percentage of gross domestic product has led to different medical care cost containment strategies. The Centers for Medicare and Medicaid Services (formerly Health Care Financing Administration) instituted a system-wide payment strategy in the form of a prospective payment system in 1983–1984. Managed care organizations and private insurance providers have adopted similar prospective payment systems and have developed clinical practice guidelines for quality evaluation and reimbursement.
Under these influences, hospital length of stay in acute care settings in the United States has decreased during the past 20 yrs. Days of care per 1,000 population decreased from 1297.0 days in 1980 to 566.1 days in 2000. Likewise, average length of stay in days decreased from 7.5 days in 1980 to 4.9 days in 2000. For hospitalized children, days of care per 1,000 population decreased from 341.4 days in 1980 to 178.3 days in 2000, whereas average length of stay remained stable at 4.5 days. 2
Cost-containment pressures in rehabilitation are similar to those noted in acute care settings, but less is known of the secular trends in inpatient rehabilitation. Much of the information about length of stay in rehabilitation comes from studies of adult patients with traumatic brain injury and stroke. A multicenter review of length of stay for adults with traumatic brain injury noted a significant decrease in length of stay during the years 1990 through 1996. 3 We found no similar reports of changes in length of stay for the population of children in rehabilitation. Rehabilitation effectiveness is the change in an outcome measure between admission and discharge and can be used to evaluate the effect of changes in therapeutic interventions. Studies of the effectiveness of rehabilitation are limited by the diversity of patients in rehabilitation and by varying therapeutic program intensity. Even with these limitations, several studies in adult patients have shown that interdisciplinary intensive rehabilitation is more effective than less aggressive and less coordinated rehabilitation efforts. 4 Similarly, a study of children with traumatic brain injury showed a trend toward better outcome with more comprehensive rehabilitation. 5
Finally, all children in the United States with disabilities are eligible for free and appropriate public education in the least restrictive environment under the Individuals with Disabilities Education Act. 6 Children in rehabilitation often have motor and cognitive impairments that limit full reintegration into the educational system without modifications. Educational placement settings range from least restrictive, in a regular classroom with an outside tutor, to most restrictive, with homebound education, as shown in Table 1. The type of educational setting at discharge from rehabilitation provides information about the ability of the child to return to normal childhood roles and activities. The purpose of this study was to assess changes in the length of stay and its effect on effectiveness and return to school in an inpatient pediatric rehabilitation unit during a 5-yr period from fiscal year 1997 through 2001.
A cohort of 321 children discharged from the Kluge Children’s Rehabilitation Center during fiscal years 1997 through 2001 were included for analysis. Kluge Children’s Rehabilitation Center is a pediatric inpatient and outpatient rehabilitation facility and a unit of the Children’s Medical Center at the University of Virginia. Comprehensive interdisciplinary and family-centered care is provided to children, adolescents, and young adults (ages 0–21 yrs) with a range of acquired disabilities, including brain injury, spinal cord injury, traumatic orthopedic injuries, and burns. For this study, we obtained permission from the University of Virginia Human Investigation Committee to review patient records and utilize administrative databases. Data from children who were aged ≥2 yrs and who had complete records were included. Using these criteria, three children were excluded for age of <2 yr, and 18 children were excluded because of incomplete records. Baseline characteristics of children included year of discharge, age, sex, admission diagnosis, admission and discharge functional ratings, and type of health insurance. There was no significant difference between the groups by year for age at admission, sex, diagnosis, discharge WeeFIM® rating, and insurance type. There was a significant difference between the years for admission WeeFIM rating, with lower median WeeFIM rating during 2000 (Kruskal-Wallis, P < 0.05) (Table 2).
Length of Stay.
Average length of stay was calculated for all patients by year of hospitalization. Regression models were used, with the logarithm of length of stay as the dependent variable, to determine differences in length of stay across years, after adjusting for age, sex, diagnosis admission functional rating, and insurance type.
The FIM™ instrument for children (WeeFIM®) was used to evaluate a child’s functional ability. The WeeFIM instrument is an 18-item instrument that measures performance in essential daily functional skills across the domains of self-care, mobility, and cognition. Each item is rated from 1 to 7, and total scores range from 18 to 126, with 126 representing the expected functional skills of an 8-yr-old child. The WeeFIM instrument has been shown to be valid and reliable in older children and adolescents through age 21 with alterations in functional ability. 7–9 Because we administered the WeeFIM instrument to children of >8 yrs of age, we reviewed the data for possible ceiling effects.
The raw WeeFIM rating can be converted to a functional quotient that is a percentage based on chronological norms. 7 Effectiveness was calculated both as the change in functional outcome WeeFIM rating between admission and discharge by year of discharge and as the difference in WeeFIM functional quotient by year of discharge.
Return to School.
Educational placement before injury and at discharge were documented. For infants and young children, educational placement was no formal educational services or early intervention. For children preschool and school-aged, educational placement was classroom based, home schooled, or homebound. For older adolescents, placement was graduated high school or dropped out of high school.
Length of Stay.
Length of stay was significantly different during the 5 yrs of the study, with mean length of stay of 58.9, 43.5, 30.7, 40.9, and 24.0 days in 1997 through 2001, respectively (Kruskal-Wallis, P < 0.001) (Fig. 1). Five children had extremely long lengths of stay in 2000 (196, 171, 123, 116, and 114 days). Excluding records from these children, the mean length of stay for 2000 was 25.1 days. Regression models that included all children indicated that the length of stay differed significantly across years (P < 0.001), even after adjusting for age at admission, admission WeeFIM rating, sex, diagnosis, and insurance type.
Effectiveness when measured by change in unadjusted WeeFIM rating was not significantly different during the 5 yrs of the study. Changes in effectiveness remained nonsignificant when measured using the WeeFIM functional quotient. Ceiling effects did not seem to be an issue in the analyses. Of the 321 subjects, only 16 (5%) had WeeFIM scores equal to 126. There was no significant difference across fiscal years in the percentage of subjects with discharge WeeFIM scores equal to 126.
Return to School.
Discharge educational placement to a classroom-based setting decreased significantly during the 5 yrs of study. Logistic regression analysis of the group of children enrolled in a classroom-based setting before injury showed that there was a significant difference across years in return to classroom setting, controlling for insurance type, diagnosis, and discharge WeeFIM rating (P < 0.05). There was no significant difference across years in the numbers of children who received early interventions services, were home schooled, or who did not qualify for educational services (graduated high school or dropped out of high school) (Table 3).
We present a cohort of children discharged from a pediatric rehabilitation unit between 1997 and 2001 and note a dramatic decrease in length of stay during this period. These results are consistent with national trends in length of stay in adult rehabilitation. We are not aware of any therapeutic changes occurring during this time period to account for the decrease in length of stay.
Prospective payment systems are generally not in place for pediatric rehabilitation that would provide a set hospital fee for specific diagnoses, such as traumatic brain injury. If they were, the hospital would have an incentive for early discharge because it would receive the same payment regardless of length of stay. Thus, we speculate that the most likely explanation for the decreased length of stay is pressure from third-party payers to reduce costs by earlier discharge from inpatient to outpatient rehabilitation services. Typically, insurers, both public and private, require that children in inpatient rehabilitation receive two 30-min therapy sessions by at least three disciplines (e.g., physical and occupational therapy, speech pathology, psychology) each day. Generally, an aggressive outpatient program would consist of 30-min sessions three times a week by one to three therapies, quickly reduced in frequency because of limitations in the total number of allowable therapy sessions per patient per year. On average, inpatient rehabilitation costs $850 per day, or approximately $5,950 per week, in contrast to the $600 per week cost of outpatient therapy (six 30-min outpatient therapy sessions).
We noted no change in the effectiveness of rehabilitation during the 5 yrs of this study—that is, there was no difference over time in the amount of functional change noted for either the raw WeeFIM ratings or the WeeFIM functional quotients. This means that the same functional results were achieved in a shorter time frame.
This may provide good news for insurance companies and, possibly, for patients and families. Certainly, regaining basic functional skills of ambulation, communication, and self-care before discharge is an important achievement. However, the WeeFIM instrument as a functional assessment has its limitations. Patients may have similar total point scores but have very different function. One patient may be walking, whereas another has strengths in communication and is nonambulatory. Subtleties of cognitive function, processing speed, and emotional and behavioral stability are not considered in the cognition portion of the WeeFIM instrument and are not meant to supplant more precise psychological measures. Careful analysis of each domain should be considered when dealing with an individual. Thus, there may be significant residual disability, not accounted for by the basic functional skills measured by the WeeFIM instrument, that would benefit from continued intensive inpatient rehabilitation treatment.
To illustrate this concern, we will present two brief vignettes of children with similar age, length of stay, and discharge WeeFIM ratings. DB is a 14-yr-old young man with traumatic brain injury who was discharged with a total WeeFIM score of 97, with relative strengths in mobility and supervised self-care. His cognitive processing speed and judgment were impaired to the point that he needed constant one-on-one supervision. He was discharged from rehabilitation because he was independently mobile and could receive outpatient therapy for his persisting motor impairment. This placed a significant burden on his family members, who were concerned for his safety, unaccustomed to their roles of being constantly vigilant, and needed to return to their places of employment. DB’s need for supervision prevented his family from returning to work and to their former routines. In contrast, AC is a 17-yr-old young woman with traumatic brain injury and lower limb fractures who was discharged with a total WeeFIM score of 90. She had significant motor impairment and limitations in her self-care (could not transfer without assistance), but her cognitive function was relatively preserved. The burden to her family and caregivers was much less than that presented for DB. The family experienced fewer, albeit different, disruptions to their former routines. Given the similarity in age, length of stay, and WeeFIM scores, these two patients represented very different discharge scenarios to their families and their rehabilitation team.
One clear detriment of shortened length of stay is the observed decrease in return to a classroom-based educational setting. A decrease in length of stay may make re-entry into a classroom setting more difficult, especially for children with brain injury (who represent the largest group of children receiving rehabilitation services), because cognitive changes continue after improvement in motor abilities has reached a plateau. Although children with brain injury are discharged when they are mobile and can perform age-appropriate self-care, their cognition continues to improve for at least 1 yr after discharge. 10 The educators at Kluge Children’s Rehabilitation Center note that, at the time of discharge, many children are making significant cognitive improvement but are not ready for a classroom setting. Longer length of stay in previous years may have allowed children to reach a higher level of cognition and behavioral control before discharge, facilitating return to the classroom.
There will be persistent pressures to maximize cost-effectiveness in all areas of health care, including pediatric rehabilitation. The 5-yr trends reported in this study demonstrate that basic functional skills can be regained in a shorter period than previously thought. However, issues of safe behavior, ability to re-enter the family home and community, local resources for continuing rehabilitation therapies, and the capacity of the family to manage a child still recovering from a severe injury or illness must be considered in discharge planning.
Finally, children in inpatient rehabilitation represent a diverse population, and there may not be a simple response to the question of how best to facilitate their return to home and community. Different measures (e.g., functional ability vs. re-entry into a school setting) show different outcome. Policy and reimbursement decisions based on a limited set of outcome measures may provide misleading information about outcomes important to children and their families, such as return to school. Therefore, it is imperative that families, physicians, and policy makers assess the results of therapeutic interventions, such as inpatient rehabilitation, mindful of the wide variation in outcome based on the choice of different outcome measures. Well planned clinical trials with different outcome measures will provide information about the most effective therapeutic interventions, answering the questions foremost in the minds of all parties involved.
1. Gross Domestic Product, National Health Expenditures, Federal and State and Local Government Expenditures: United States, 1960–1999. Washington, DC, Centers for Disease Control, National Center for Health Statistics, 2001
2. Discharges, Days of Care, and Average Length of Stay
in Non-Federal Short Stay Hospitals, According to Selected Characteristics: United States, Selected Years 1980–1999. Washington, DC, Centers for Disease Control, National Center for Health Statistics. 2001
3. Kreutzer JS, Kolakowsky-Hayner SA, Ripley D, et al: Charges and lengths of stay for acute and inpatient rehabilitation
treatment of traumatic brain injury, 1990–1996. Brain Inj 2001; 15: 763–74
4. Semlyen JK, Summers SJ, Barnes MP: Traumatic brain injury: Efficacy of multidisciplinary rehabilitation
. Arch Phys Med Rehabil 1998; 79: 678–83
5. Swaine BR, Pless IB, Friedman DS, et al: Effectiveness of a head injury program for children: A preliminary investigation. Am J Phys Med Rehabil 2000; 79: 412–20
6. Individuals with Disabilities Education Act Amendments of 1997. Public Law 105-17
7. WeeFIM SystemSM Clinical Guide, Version 5.01. Buffalo, State University of New York at Buffalo, 2000
8. Ottenbacher KJ, Msall ME, Lyon N, et al: Measuring developmental and functional status in children with disabilities. Dev Med Child
Neurol 1999; 41: 186–94
9. Ottenbacher KJ, Msall ME, Lyon NR, et al: Interrater agreement and stability of the Functional Independence Measure for Children (WeeFIM): Use in children with developmental disabilities. Arch Phys Med Rehabil 1997; 78: 1309–15
10. Jaffe KM, Polissar NL, Fay GC, et al: Recovery trends over three years following pediatric traumatic brain injury. Arch Phys Med Rehabil 1995; 76: 17–26
New Rapid Communication
The American Journal of Physical Medicine & Rehabilitation is committed to publishing important research and scientific discussion as quickly as possible. Therefore, a new category of manuscript submission has been developed to expedite publication of timely research articles and other vital scientific communications. This new category of submission will be referred to as “Rapid Communication.” Submissions considered by the editors to be extremely timely or of vital importance to the readers will receive an expedited peer-review process and upon acceptance will be published in the next available issue of the Journal. Authors may request that the editors consider a particular manuscript submission for expedited review and publication as a “Rapid Communication.” Each research paper selected as a “Rapid Communication” will also receive simultaneous early publication on the Journal’s web site at www.amjphysmedrehab.com