Objective. Clinical redesign of processes in hospitals that care for children has been limited by a paucity of severity-adjusted indicators that are sensitive enough to identify areas of concern. This is especially true of hospitals that analyze pediatric patient care using standard Centers for Medicare and Medicaid Services (CMS) diagnosis-related groups (DRGs). The objectives of this study were to determine whether 1) utilization of all-patient refined (APR)-DRG severity-adjusted indicators (length of stay, cost per case, readmission rate) from the National Association of Children's Hospitals and Related Institutions (NACHRI) database could identify areas for improvement at University of Michigan Mott Children's Hospital (UMMCH) and 2) hospital staff could use the information to implement successful clinical redesign.
Methods. The APR-DRG Classification System (version 20) was used with the NACHRI Case Mix Comparative Database by severity level comparison from 1999 to 2002. Indicators include average length of stay (ALOS), case mix index, cost per case, and readmission rate for low acuity asthma (APR-DRG 141.1). UMMCH cases of 141.1 (n = 511) were compared with NACHRI 141.1 (n = 64 312). Although not part of the standard report, mortality rates were calculated by NACHRI for UMMCH and an aggregate of NACHRI member children's hospitals.
Results. Data from 1999 revealed that in noncomplicated asthma cases (level 1 severity), the UMMCH ALOS versus NACHRI ALOS was slightly longer (UMMCH 2.16 days vs NACHRI 2.14 days), and the cost per case was higher (UMMCH $2824 vs NACHRI $2738), whereas levels 2, 3, and 4 cases (moderate, major, and extreme severity) indicated the ALOS and cost per case were lower than the national aggregate. This showed that the APR-DRG system was sensitive enough to distinguish variances of care within a diagnosis according to severity level. After analysis of internal data and meeting with clinicians to review the indicators, 3 separate clinical processes were targeted: 1) correct documentation of comorbidities and complications, 2) standardized preprinted orders were created with the involvement of the pediatric pulmonologists, and 3) standardized automatic education for parents was started on the first day of admission. Yearly data were reviewed and appropriate adjustments made in the education of both residents and staff. In 2002, the UMMCH ALOS dropped to 1.75 ± .08 days from 2.16 ± .09. In 2002, the NACHRI ALOS was 2.00 days ± 0.01 versus the UMMCH ALOS of 1.75 days ± 0.0845, indicating that the UMMCH ALOS dropped significantly lower than the NACHRI aggregate database over the 3-year period. Cost per case of UMMCH compared with NACHRI after the 3 years indicated that UMMCH increased 12%, whereas the NACHRI aggregate increased 18%. These data show that length of stay and cost per case relative to the national database improved after clinical redesign. Improvements have been sustained throughout the 3-year period. Readmission rates ranged from 2.97% to 0.80% and were less than the national cohort by the third year. There were no mortalities in the UMMCH inpatient asthma program. This demonstrates that clinicians believed that the data from the APR-DRG acuity-adjusted system was useful and that they were then able to apply classical clinical redesign strategies to improve cost-effectiveness and quality that was sustained over 3 years.
Conclusions. Severity-adjusted indicators were useful for identifying areas appropriate for clinical redesign and contributed to the improvement in cost-effective patient care without a detriment in quality indicators. This method of using a large comparative database, having measures of severity, and using internal analysis is generalizable for pediatric hospitals and can contribute to ongoing attempts to improve cost-effectiveness and quality in medical care.
Over the last several decades, there has been increasing attention to the quality and cost-effectiveness of medical care.1,2 Tools that are available to measure utilization in the pediatric population have been relatively limited compared with the Medicare population, for which the design and development of diagnosis-related groups (DRGs) began at Yale in the late 1960s.3 The initial motivation for the development of DRGs was to create a framework for monitoring quality of care and the utilization of services in a hospital setting. In 1982, the Tax Equity and Fiscal Responsibility Act modified Medicare hospital reimbursement from cost per diem to a fixed amount per discharge. In 1983, Congress amended the Social Security Act to include a national DRG-based hospital prospective payment system for all Medicare patients. Because DRGs were focused on the Medicare population, ongoing development and refinement was principally done with adult discharge data. However, many commercial insurance companies and state Medicaid plans also based their payments to hospitals on DRGs. This caused significant problems for children's hospitals when they were judged for their cost-effectiveness and utilization by the Center for Medicare and Medicaid Services (CMS) DRG system.3
Clinicians in children's hospitals became very critical of the DRG data, as it had little relevance to the actual acuity of their patient care. Because DRG categories in the 0 to 17 age range are classified without any comorbidity-complication adjustment, tertiary care children's hospitals were at a disadvantage as they cared for large populations of severely chronically ill children who had many complications and comorbidities.3 For example, a child who was admitted with pneumonia and also had cystic fibrosis would simply be listed as pneumonia age 0 to 17 despite the significant comorbidity. Adult-age DRGs allowed the addition of comorbidities and complications to adjust the case mix index or level of severity.
The National Association of Children's Hospitals and Related Institutions (NACHRI) began extensive research in the 1980s to evaluate and reformulate the DRG categories for neonates and other pediatric patients. The system initially developed by NACHRI was called Pediatric Modified DRGs, which included additional DRGs for neonates that took into account birth weight, duration of mechanical ventilation, multiple major problems, and surgical procedures. All-patient refined DRGs (APR-DRGs), were co-developed by NACHRI and 3M HIS and introduced in 1991. The APR-DRGs expand the basic CMS DRG structure by adding subclasses to each DRG. There are 4 subclasses that describe patient differences relating to severity of illness and 4 subclasses for risk of mortality. The 4 severity-of-illness subclasses are minor, moderate, major, and extreme severity of illness and are denoted as the suffix to the DRG (eg, a minor asthma case would be denoted as APR-DRG 141.1, or 141 for asthma and .1 for minor). The process of determining subclasses for APR-DRGs begins by first assigning a severity-of-illness level and risk of mortality to each secondary diagnosis. The assignment takes into consideration not only the level of secondary diagnosis but also the interaction among the secondary diagnoses, age, principal diagnosis, and the presence of certain surgical procedures or nonsurgical procedures.3
There have been several revisions to the APR-DRGs since 1991. The current version is APR-DRG V20. The 25 Major Diagnostic Categories in V20 break down into 314 APR-DRGs, which further subdivide into the 4 severity classes, resulting in 1258 APR-DRGs.
The development of the APR-DRGs made available a patient classification system that could be used for 1) comparison of the pediatric populations across a wide range of resource and outcome measures, 2) the evaluation of differences in inpatient mortality rates, 3) the implementation and support of practice guidelines, 4) the identification of opportunities for continuous quality improvement, 5) the basis of internal management and planning systems, and 6) the evaluation and management of payment arrangements.3–6
New APR-DRGs are developed on the basis of clinical logic and tested extensively with historical data. To illustrate, data used in the development of version 15 included a nationwide set of data that included 5.7 million discharges from 657 general hospitals and all discharges from 40 children's hospitals in the United States. Comparison of payment with actual cost (including indirect and direct costs) for various DRG systems showed improvement for children's hospitals; specifically, the ratio was 0.69 with standard DRGs versus 0.832 with APR-DRGs.6
Development of APR-DRGs provides the first operational means of accurately defining and measuring a pediatric hospital's case mix complexity. Many attributes go into case mix, including severity of illness, which refers to the extent of physiologic decomposition or organ system loss of function. This information is derived by coded International Classification of Diseases, Ninth Revision—Clinical Modification diagnoses and procedure codes obtained from the medical record abstract performed by hospital medical record coders. These codes are processed by software that uses an algorithm that contains 18 steps in looking at each secondary diagnosis. The algorithm sets priorities as to which diagnoses take precedence and determines the overall level of acuity. Management resource intensity refers to the relative volume and type of diagnostic therapeutic services used in the management of a particular illness. In general, a hospital that has a more complex case mix index from an APR-DRG perspective would mean that the hospital treats patients who require more resources, although not always, for instance, a terminally ill child with cancer would have high acuity, but the major care would be aimed toward comfort and therefore diagnostic or operative resources would not be used extensively. The algorithms adjust for these variances.
The University of Michigan Mott Children's Hospital (UMMCH) began using APR-DRGs for clinical redesign in 1998. The NACHRI Case Mix Comparative Database was reviewed by clinicians in the Office of Clinical Affairs, who specifically reviewed APR-DRG categories that indicated that our length of stay or cost per case had significant variation from other NACHRI hospitals.
High-volume cases showing significant variation in average length of stay (ALOS) and cost per case became targets for clinical redesign. Specifically, we targeted APR-DRG 141 asthma, which showed that UMMCH had a higher cost per case and longer length of stay in our level 1 cases as compared with national data. Levels 2 to 4 cases were lower than the mean. This made historical sense in that low-acuity asthma was cared for by our general pediatric services versus higher acuity patients, who were usually admitted to the pulmonary service. Our general services lacked standardized protocols for treatment of asthma patients. Subsequently, we formed a team of pediatric pulmonologists, hospitalists, respiratory therapists, and nurse clinicians to review the data and make plans for changes. Data were presented to the UMMCH Executive Committee and to the pediatric faculty for their comments and suggestions.
The NACHRI Case Mix Comparative Database for calendar years 1999–2002 was used for the initial investigation of UMMCH performance. In 2002, total discharges for UMMCH were 8489 compared with a database of 525 154 from 67 children's hospitals, 283 053 from 105 major teaching general hospitals (those with resident-to-bed ratios of >0.25), and 737 628 patients from 1202 general hospitals (not major teaching). Therefore, the total database contained 1 545 835 patient encounters. The standard report lists ALOS, cost per case, and acuity level for each diagnostic code that is obtained from administrative data reported by each hospital from the standard medical record abstract performed by medical record coders. These individuals are educated and certified as medical record coders and work under strict governmental regulations and auditing because their abstract ultimately leads to the facility bill for each patient. NACHRI tracks the rate of uncodable diagnoses for each hospital to monitor compliance; however, no formal interrater reliability is available. The administrative data are deidentified as to patient name or hospital registration number. Costs reported in the database are estimates of total cost, including direct and indirect, derived from a ratio of costs to charges. Costs therefore are estimates derived from hospital charges, and they are most relevant in looking at relative change over time.
Once the UMMCH Executive Summary report was reviewed, each APR-DRG that fell outside the norm was reviewed. The total number of cases compared with level 1 asthma in 1999 at the start of the study was 116 for UMMCH and 17 505 for NACHRI. Once level 1 was identified as a higher cost and length of stay compared with the mean for the other NACHRI hospitals, 3 separate processes were targeted. 1) Coders pulled a sample of asthma patient medical records and compared documentation with the level of coding. Review by clinicians and coders indicated that in a small percentage of cases, documentation of comorbidities and complications was lacking; therefore, higher acuity cases were mixed into level 1, thus affecting accurate analysis of ALOS. E-mail messages were relayed to faculty regarding the appropriate documentation of all diagnoses, especially when an asthma patient had other problems, such as hypoxia, electrolyte abnormalities, fever, nutritional abnormalities, altered state of consciousness, pneumonia, or other chronic diseases, all of which could potentially increase the acuity level. 2) A meeting was held with pediatric pulmonologists and hospitalists to create a standard set of orders to allow respiratory therapists and nurses to wean nebulization treatments and oxygen according to standard parameters rather than wait for changes to be made on twice-a-day rounds. Allowing weaning according to protocol increased the rapidity of changes in therapy and decreased ALOS. Medication dosing protocols were also standardized. 3) An education coordinator was notified on the day of the admission, who contacted the family to discuss previous care and teach or reteach the family and patient what was considered necessary and the asthma action plan for the patient. This decreased discharge delays, which previously occurred as a result of families' and patients' not being prepared to assume care at home.
There were no other specific changes as far as creation of a short-stay care unit, changes in attending or resident staffing, or changes in nurse-to-patient ratios. Physician education about documentation coding took place through faculty e-mails about specific cases and general education sessions at faculty meetings. Formalized education of families about asthma on the first day of admission started with the introduction of the order sets in 2000; however, we do not have a specific record of the percentage of the time the goal was achieved. The educators reported to us informally that they believed that they were able to meet with families on the first day >80% of the time. There was no Institutional Review Board involvement with the collection of administrative data, as it is not reported in patient identifiable form and the care was not provided outside accepted norms. Data were collected in a confidential manner under the auspices of quality improvement about which patients are notified during their admission process. Review of data and updating order sets occurred within the core group approximately every 3 months, and annual reports were submitted to pediatric residents, faculty, and the UMMCH Executive Committee.
Statistical analysis regarding length of stay and cost per case was performed using the 2-tailed t test. P < .05 was considered significant with data presented as the mean plus the standard error of the mean.
From 1999 to 2002, UMMCH saw ∼116 to 130 patients per year with a diagnosis of asthma at level 1 severity (total n = 511). These data were compared with the NACHRI children's hospitals aggregate database, which had a range of 15 179 to 17 050 level 1 asthma patients per year (total n = 64 312). In 1999, UMMCH had an ALOS for level 1 asthma of 2.16 days versus NACHRI at 2.14 days with the cost per case at UMMCH at $2824 versus NACHRI at $2738. ALOS at UMMCH for 1999 compared with the NACHRI ALOS was not statistically different (P = .8); however, the UMMCH data for levels 2, 3, and 4 was lower than the national average. We believed, therefore, that we should try to improve level 1. Over the 3 years, UMMCH's ALOS dropped to 1.75 days (standard error ± 0.08) as compared with NACHRI's ALOS of 2 days (standard error ± 0.01), showing a significant difference (P = .0039; Figs 1 and 2) . The cost per case from 1999 to 2002 in UMMCH and in NACHRI hospitals increased; however, UMMCH increased 12%, whereas NACHRI increased 18% (Fig 3). With attention toward the level 1 cases and clinical redesign occurring with the 3 steps above, improvements were made in ALOS and slight improvements were made in adjusted cost per case, with UMMCH now below the national average for children's hospitals for ALOS and cost per case. The readmission rates for asthma dropped from 2.97% in 1999 to 0.80% in 2002, whereas the national average readmission rate remained between 2.0% and 2.3% (Tables 1 and 2). There were no deaths in the UMMCH program. The mortality rate of level 1 asthma across the country is <0.005%.
For many years in children's hospitals, the only data available to analyze cost per case and length of stay across multiple national databases used CMS DRGs, which do not take into account the comorbidities and complications in pediatric patients. Because of this, there was an overall cynicism about these data, and few clinicians in UMMCH would base clinical redesign efforts on national DRG utilization data. Using the NACHRI APR-DRG Case Mix Comparative Database has changed that perception. With the ability to separate out acuity levels within a diagnosis, we now have a sensitive tool to determine variances in care. Clinicians have accepted APR-DRG reports and continue to review them to identify variances in their practice. Clinical redesign may be initiated for many purposes, such as to adopt new practices that are based on classic medical research. Redesign to improve efficiency and cost-effectiveness has been limited by the lack of severity-adjusted data that are widely available or accepted in pediatrics. The NACHRI APR-DRG Case Mix Index Database now offers a sensitive tool for identifying utilization issues and allows significant statistical power as a result of the ∼1.5 million patient discharges per year from both children's and general hospitals that can be analyzed. This analysis can be performed from discharge data already collected in the administrative databases used by all hospitals.
Our study exemplifies the situation in which patterns for care at a tertiary hospital can be different for low-acuity versus high-acuity cases. At the study children's hospital, our overall care of asthma patients showed utilization patterns that are very similar if not slightly better than the aggregate database. However, our care of level 1 asthma was done by general pediatric services without the standardized protocols, and ALOS lagged behind other hospitals, that were often more efficient at the level 1 care. This trend has recently been reported in national literature, ie, large academic institutions were less efficient than community hospitals in low-acuity cases.2 This stimulated us to pursue aggressively improving the care to make us more efficient and enable us to improve our performance. The method of using a large pediatric comparative database that includes levels of severity along with internal analysis is generalizable to other hospitals that care for children in an attempt to improve cost-effectiveness and quality in pediatric care.
We thank Patricia Kneeland for preparation of the manuscript and John Muldoon for comments.
- ↵Srivastava R, Homer CJ. Length of stay for common pediatric conditions: teaching versus nonteaching hospitals. Pediatrics.2003;112 :278– 281
- Muldoon J. Structure and performance of different DRG classification systems for neonatal medicine. Pediatrics.1999;103 :302– 318
- Averill R, Muldoon J, Vertrees J, et al. The evolution of DRG case mix measurement using diagnosis related groups. In: Goldfield N, ed. Physician Profiling and Risk Adjustment. 2nd ed. Gaithersburg, MD: Aspen Publishers, Inc; 1999;391–454
- ↵Muldoon J. Pediatrics and DRG case mix classification. In: Goldfield N, Boland P, eds. Physician Profiling and Risk Adjustment. 1st ed. Gaithersburg, MD: Aspen Publishers, Inc; 1996;252–270
- Copyright © 2004 by the American Academy of Pediatrics