Strategic Planning for Children With Chronic Illnesses

Children's Hospitals Today, Winter 2008
by John Neff, M.D. and
Yu-Fang Li, RN, Ph.D.
Center for Children with Special Needs
and
Patrick Hagan, MHSA, BA
President and Chief Operating Officer
Children’s Hospital & Regional Medical Center, Seattle

With the ongoing development of new technology and treatments, children’s hospitals continue to improve the life expectancy of children with chronic conditions. While many children may never be hospitalized or are hospitalized only once during childhood, the children with chronic illnesses are often dependent on a children’s hospital for specialized treatment over months and years.

From a children’s hospital planning perspective that means that an increasing percentage of the population of children served will have more chronic illnesses and will live longer. Over time, as these children grow older, they can expect the need for more frequent admissions and more expensive care. Families with children with special health care needs should take these factors into consideration. And children’s hospitals in collaboration with the health plans and physicians who serve these children need to refine their quality programs and improve planning to meet the immediate and future needs of these children as they enter adulthood.

In their own strategic planning, children’s hospital CEOs can incorporate these expected disease trajectories of chronic condition groups to project long term building needs, to improve budget projections, facility design and enhance process improvement programs so these children can receive the best possible care.

A New Way to Classify Chronic Conditions

Categorizing children with chronic conditions in clinically meaningful ways is particularly difficult because of the general infrequency and variable nature of these conditions. Chronic conditions in childhood do not fit into the adult chronic disease models driven by common degenerative disease groups such as stroke, hypertension and cardiovascular disease. In addition, the chronic conditions that occur most frequently in children, such as asthma and attention deficit and hyperactivity disorders, have variable trajectories that do not match the progressive patterns of other chronic conditions. Recently some of our work tracking children with chronic conditions in health plan data using Clinical Risk Groups has suggested a logical grouping of children with chronic conditions that can provide a framework for identification, strategic planning and outcome measures. The categories may be more meaningful for clinicians and planners when chronic conditions are separated into those that are generally episodic in nature, and those that are lifelong, with a separate group for those with non-benign malignancies.

The Clinical Risk Groups (CRGs) classification software developed by 3M Health Information Systems and NACHRI classifies individuals into non-chronic and chronic condition categories. CRGs use encounter data and ICD-9-CM codes to classify individuals into groups of increasing complexity and severity. Each person is classified into one of nine core health status groups of increasing complexity. There is a further severity grading of up to six severity levels for those classified in the chronic condition groups. Classification is hierarchical, nonduplicative and assigns each person into the highest applicable complexity and severity group. NACHRI has contributed considerable pediatric-specific information in development of CRGs. The CRGs were developed to be used in health plan data for risk adjustment and recently have been used to identify and track children for care management purposes. Until now, CRGs have not been used in hospital data sets. The application of this software to hospital data should be ideal for identifying patient populations and their patterns of admission to a children’s hospital.

Hospitals are accustomed to analyzing activities according to individual hospital stays, admissions or discharges. This type of analysis identifies patterns of admission by expenditures or charges, length of stay and outcomes such as readmissions, certain complications and mortality. These analyses are generally based on principal diagnoses and surgical procedures, either directly or through the groupings of a DRG classification system. The All Patient Refined Diagnosis Related Groups (APR-DRGs) developed with pediatric-specific guidance from NACHRI, is one such system that is widely used by hospitals. The APR-DRGs category and severity subclass is intended to describe the main reason and severity for a hospitalization. While it may be possible to identify many chronic illnesses through a DRG system, the principal axis of the classification is the acute illness that caused the hospitalization, and thus an additional classification system may be necessary to examine admission patterns for a chronically ill population.

Significance of Patient Population Analysis

Analyzing hospital experience according to the patient population provides a different perspective than just an analysis of discharges or admissions. An analysis of a patient population provides an opportunity to track a given child’s hospitalization experience over an extended time frame. It also provides the opportunity to concentrate on the groups of patients that would be expected to have long lengths of stay or frequent readmission rates.

A chronic illness classification system is not an all purpose, stand alone methodology. It has its greatest potential when used in conjunction with an admission-based classification system such as the APR-DRGs. As an example, by using CRGs it is possible to select some specific chronic conditions such as cerebral palsy and use the APR-DRGs to identify the reasons for various hospitalizations, which can be expected to have different lengths of stay, outcomes and costs.

Many hospitals know very little about the patient population they serve except the information that can be gathered through individual specialty groups such as oncology, cystic fibrosis, craniofacial and sickle cell disease. Much of the information available within a hospital’s own general data network relates to individual admissions and severity grading of individual admissions, not the cumulative data that can be gained from analyzing a specific patient’s long term interaction with the hospital. This includes the DRG systems, which are designed to classify individual admissions, not patient populations. Hospital information systems may sometimes use ICD-9-CM lists of certain diagnoses, but these are seldom inclusive of the many rare conditions that occur and do not take into account combinations of conditions.

Seattle Study Identifies Trends

For this project, CRGs were adapted to be used in five years of Seattle Children’s Hospital & Regional Medical Center discharge data to analyze all of the patients and to identify five year trends. The CRG system contains nine core health status groups – 1) healthy, 2) significant acute, 3) single minor chronic disease, 4) minor chronic disease in multiple body systems, 5) single moderate or dominant chronic disease, 6) significant chronic disease in multiple organ systems, 7) dominant chronic disease in three or more organ systems, 8) dominant and metastatic malignancies, 9) catastrophic conditions. These nine groupings were consolidated into five broader categories for inpatient hospital use: non-chronic; episodic chronic; lifelong chronic; lifelong progressive chronic or dependent on technology; and non-benign malignancies.

Frequently occurring diagnostic groups in each category are as follows:

1. Acute illnesses without underlying chronic conditions: newborn respiratory distress and sepsis, bronchiolitis, pneumonia, uncomplicated urinary tract infection, tonsil and adenoid removals, gastroenteritis, acute appendicitis, and various traumatic injuries
2. Episodic chronic conditions: asthma, repeat urinary tract infection with reflux, attention deficit and hyperactivity disorders, depression and most mental health conditions, simple cleft lip and palate, seizure disorders without other dominant conditions, and minor foot deformities and bone cysts
3. Lifelong chronic conditions: Type 1 diabetes, chromosomal anomalies, major respiratory anomalies, major craniofacial disorders, complex cyanotic and major cardiac anomalies, and major birth defects
4. Lifelong and progressive chronic conditions or dependent on technology: congenital quadriplegia, diplegia and hemiplegia, spina bifida, cystic fibrosis, major organ transplant, muscular dystrophy, total parental nutrition and ventilator support
5. Non-benign malignancies: brain and central nervous system, acute leukemias, kidney malignancies, digestive malignancies, and other malignancies

Patient encounter data from Seattle included all patients 0-19 years admitted from the medical and surgical services from 2001 through 2005, including the intensive care units but excluding the rehabilitation and psychiatry inpatient units.

The five-year trends showed that the children admitted became slightly older, had conditions that were more chronic and complex, and were admitted more frequently.

Results and Conclusions

Clinical Risk Groups show trends of increasing lifespan, hospital utilization and treatment costs among children with chronic illnesses. Understanding these trends is vital to planning for the future of children’s hospitals.

At Seattle Children’s Hospital in 2005, there were 9,707 discharges of 6,889 patients who were 19 years old or younger, admitted to at least one non-psychiatric, non-rehabilitation service during the study period. Table 1 demonstrates the distribution of all of these patients according to the five broad hospital use categories, analyzed by total charges per group and median and mean charges per group. Figures 1–4 demonstrate the average age of patients by group, percent of patients admitted in the previous five years, and number of admissions per patient by group.

Of the 6,889 patients discharged in 2005, 67.3 percent had at least one chronic condition, and these patients accounted for 91.7 percent of all hospital charges for that year. The patients with lifelong chronic conditions, including those with progressive chronic illnesses, dependent on technology and with non-benign malignancies, represented 38.8 percent of the patients and 79 percent of the costs. The median and mean charges increased for each more complex patient group as did the average age and number of readmissions.

The average age of the children increased according to complexity group from 4.3 years for those with non-chronic conditions to 9.2 years for those with progressive chronic illnesses or dependent on technology (see Figure 1).

Admission rates increased according to complexity (see Figure 3). Patients with no chronic conditions were often admitted with complete resolution with an admission rate of 1.04 per child per year in contrast to those with lifelong progressive conditions with an admission rate of 2.4 per child per year. Figure 4 shows the number of admissions per patient according to category. Of children with non-chronic conditions, 95.9 percent were admitted only once, and only 0.4 percent were admitted three or more times. In contrast, 50.8 percent of the children with progressive chronic conditions or dependent on technology were admitted only once; 28.0 percent were admitted three or more times. Children with non-benign malignancies had the highest number of admissions per patient of all the groups.

Of special interest: The percentage of children in each group admitted in the previous five years varied significantly according to the complexity of the group (see Figure 2). Only 4.1 percent of children with nonchronic conditions were previously admitted in comparison to 37.3 percent of children with lifelong chronic conditions and 66.3 percent of children with progressive chronic illnesses or dependent on technology.

Just as new technology is helping improve the treatments and life expectancy for children with chronic conditions, so too are new developments in technology and classification systems tools helping children’s hospitals refine their abilities to best respond to the special health care needs of this growing population of children.

Tables and Figures

Selected Bibliography

Hughes, J.S., Averill, R.F., Eisenhandler, J., Goldfield, N.I., Muldoon, J., Neff, J.M., Gay, J.C. (2004). Clinical risk groups (CRGs): A classification system for risk-adjusted capitation-based payment and health care management. Med Care, 42, 81-90.

Neff, J.M. (Fall 2002). Chronic conditions in children, a decade of change, implications for families, providers and children’s hospitals. Children’s Hospitals Today, 14-16. Neff, J.M., Sharp, V.L., Muldoon, J., Graham, J., Myers, K. (2004). Profile of medical charges for children by health status group and severity level. Health Services Review, 39, 73-89.

Neff, J. M., Sharp, V.L., Muldoon, J., Graham, J., Popalisky, J., Gay, J.C. (2002). Identifying and classifying children with chronic conditions using administrative data with the clinical risk group classification system. Ambulatory Pediatrics, 2, 71-79.

Neff, J.M., Sharp, V.L., Popalisky, J., Fitzgibbon, T. (2006). Using medical billing data to evaluate chronically ill children over time. J. Ambulatory Care Management, 29, 283-290.