Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The Influence of Chronic Disease on Resource Utilization in Common Acute Pediatric Conditions

The Influence of Chronic Disease on Resource Utilization in Common Acute Pediatric Conditions ObjectivesTo estimate the resource utilization in hospitalizations for common pediatric conditions or procedures involving patients with chronic disease vs those with no chronic disease and to develop an economic model of hospital per-patient profit (or loss) when insurance contracts fail to account for the presence of chronic disease.Setting and DesignA retrospective analysis of selected acute pediatric conditions found in the 1991 and 1992 MedisGroups National Comparative Data Base.PatientsWe studied 30,379 pediatric admissions for common acute conditions, including concussion, croup, pneumonia, appendicitis, gastroenteritis, fractures, cellulitis, urinary tract infection, and viral illness.Main Outcome MeasuresHospital length of stay and total hospital charges.ResultsFor patients without chronic disease, mean (geometric) length of stay was 2.53 vs 3.05 days (P<.001) for patients with at least 1 chronic disease. For patients without chronic disease, mean (arithmetic) total hospital charge was $2614 vs $3663 (P<.001) for patients with at least 1 chronic disease. Assuming 75% of patients with chronic disease are admitted to a children's hospital vs 25% to a general hospital, overall loss per patient at the children's hospital ranged between 1.5% and 2.9%, depending on assumptions regarding cost-to-charge ratios and the treatment of charge outliers. Pneumonia cases were associated with a 4.0% to 5.85% loss.ConclusionsLength of stay and charges are higher for everyday pediatric conditions or procedures when patients also have a chronic disease. If insurance contracts fail to account for chronic disease, then children's hospitals will realize significant financial losses, and over time this will lead to a decline in their financial viability, a reduction in quality, or a change in their mission.Editor's Note:Data provided in this study should be of little surprise to those who work in hospitals dedicated to children. I would be (pleasantly) surprised if it helps to convince payors. For children with chronic illnesses, DRG seems to mean "Da Revenue's Gone."—Catherine D. DeAngelis, MDMAJOR CHANGES have been occurring in the systems used to finance health care in the United States. Fee-for-service reimbursement rapidly is being replaced by a prospective, often capitated system.This new financial structure, combined with lower bed occupancy rates, has forced hospitals to compete not only along the traditional dimension of quality, but also along the dimensions of price and differentiation of service offerings.Since the introduction of the diagnosis related group (DRG) classification system for prospective payment from Medicare in 1983, adult medical care has been changing and adapting to these new financial realities.Reimbursement patterns in pediatric medicine have been slower to change. It has been only in the last 5 years that a notable portion of the pediatric market began to shift from a fee-for-service to a prospective payment system.This shift has caught many pediatric hospitals off guard, as little data exist to guide the development of prospective payment rates for pediatric diagnoses,although there is some evidence of improved accuracy in refinements of the DRG classification system for children.Our study explores 2 fundamental issues confronting pediatric tertiary care centers that enter competitive prospective contracts for common pediatric procedures or conditions: (1) how does the presence of patient chronic disease influence resource utilization and length of stay (LOS) for common pediatric conditions? and (2) if children's hospitals (defined as either children's hospitals or academic pediatric tertiary care centers) provide care for common pediatric conditions, but treat a disproportionate number of these common conditions in complicated patients with 1 or more chronic diseases, what effect will this have on their financial solvency, when health care is financed prospectively in a competitive market?To determine the financial influence of chronic disease on total charges and LOS, we estimated the influence of chronic diseases on a number of common pediatric conditions and procedures. The first objective was to test the hypothesis that the presence of a chronic disease increases the hospital LOS and resource utilization in selected routine pediatric admissions. Confirmation of this hypothesis would suggest that chronic disease adjusted case-mix should be accounted for when negotiating prospective payment contracts involving common conditions in pediatric populations, an item not usually included in pediatric contracts.The second objective was to use these results to determine the financial influence of prospective payment schemes when differences in chronic disease rates across hospitals are not included in negotiated contract prices, a situation that may reflect the predicament in which many specialty children's hospitals throughout the United States presently find themselves.MATERIALS AND METHODSHOSPITAL AND PATIENT SELECTIONThe 1991 and 1992 MedisGroups National Comparative Data Base (MediQual Inc, Westborough, Mass), was used for this study. It contained 259,358 pediatric admissions (patients aged 0-18 years). The MedisGroups' database included pediatric admissions to 163 hospitals during consecutive periods in 1990 and 1991. At least 4 months of data were available for each hospital. The hospitals chosen are a representative sample of MedisGroups clients, but are not a random sample of hospitals in the United States. The method of data collection has been described in detail elsewhereand consisted of detailed medical record abstraction for up to 3 hospital reviews occurring on admission, and possibly at mid stay and at discharge from the hospital or at death. Each review included up to a potential of 700 data elements called Key Clinical Findings that reflect the physical examination, medical history, and diagnostic laboratory test results that occurred during the hospitalization. MedisGroups is the predecessor to MediQual's current performance measurement system, Atlas, that was introduced in 1996. The reliability and validity of MedisGroups data collection compare favorably with other commonly used data collection systems,although the validation of the pediatric experience is more limited.CASE SELECTIONWe selected all pediatric patients in the study database admitted to general or children's hospitals. Admissions were grouped by DRG, and the DRGs (using DRGs both with and without complication in a paired grouping) were ranked in decreasing order of frequency. Of the 30 most common DRGs, we selected 9 DRGs that generally represented simple, common pediatric problems. This selection resulted in a study data set of 30,379 patient admissions across 163 hospitals. A list of each procedure/condition grouping, and the associated DRG code is provided in Table 1.Table 1. MedisGroups Patients 18 Years Old or YoungerProcedure/Condition Group NameDiagnosis Related Group CodeDiagnosis Related GroupConcussion33Concussion, age 0-17 yCroup71LaryngotracheitisPneumonia91Simple pneumonia and pleurisy, age 0-17 yAppendicitis164Appendectomy, complicated principal diagnosis, with chronic condition (CC)165Appendectomy, complicated principal diagnosis, without CC166Appendectomy, without complicated principal diagnosis, with CC167Appendectomy, without complicated principal diagnosis, without CCGastroenteritis184Esophagitis, gastroenteritis and miscellaneous digestive disorders, age 0-17 yFractures235Fractures of the femur252Fracture, sprain, strain and dislocation of forearm, hand, foot, age 0-17 y255Fracture, sprain, strain and dislocation of upper arm, lower leg excluding foot, age 0-17 yCellulitis279Cellulitis, age 0-17 yUrinary tract infection322Kidney and urinary tract infections, age 0-17 yViral illness&ast;422Viral illness and fever of unknown origin, age 0-17 y&ast;Excluded all patients with principle diagnosis 7806 (fever of unknown origin).DEFINITIONSNotable chronic diseases were defined as those concurrent diseases that impose additional clinical vulnerability in the patient, and that may, in turn, influence resource use. As an independent arbiter of clinical vulnerability, we selected those chronic conditions or chronic diseases that were recommended by the American Academy of Pediatrics Committee on Infectious Diseases for influenza immunization.This list serves as a proxy for clinical vulnerability, and therefore, we believe it is a good starting point for listing important chronic diseases that might adversely influence resource use. This list was supplemented with common neurologic diseases (seizures, cerebral palsy, and mental retardation), common birth defects (trisomy 21), and autoimmune diseases. Rare conditions were excluded, as the intent of the analysis was to identify important chronic disease categories for inclusion in prospective payment contracts, not to define the effects of all potentially important chronic diseases.Patients were identified as having a specific chronic disease on the basis of either an appropriate International Classification of Diseases, Ninth Revision, Clinical Modification(ICD-9-CM) code listed as 1 of the in-hospital discharge diagnoses, or a Key Clinical Finding that provides detailed clinical elements on the patient's hospital stay from initial review to discharge from the hospital. The MedisGroups Key Clinical Finding codes were used to supplement ICD-9-CMcodes in the event that limited numbers of fields for the ICD-9-CMcodes may have lead to an underreporting of chronic disease for some cases. Using both the ICD-9-CMand Key Clinical Finding codes, we increased our likelihood of identifying patients with the chronic diseases we selected. A list of each chronic disease group with the associated ICD-9-CMcode or MedisGroups Key Clinical Finding is provided in Table 2.Table 2. Definition of Chronic Diseases&ast;Chronic Disease GroupICD-9-CMCodesMedisGroups Key Clinical Findings Codes and DefinitionsDiabetes250-250.91805 (diabetes); 894 (current medication insulin)Sickle cell anemia282.6-282.69804 (chronic anemia)Cerebral palsy342-343.9821 (cerebral palsy)Seizures345-345.91820 (seizures); 780300 (witnessed seizure)Asthma493-493.91840 (chronic lung disease)Congenital heart disease745-747.49829 (congenital heart disease); 746700 (hypoplastic heart); 746890 (valve atresia); 746900 (major cardiac anomaly); 747210 (aortic arch abnormality)Trisomy 21758.0834 (Trisomy)Cancer140-239.9810 (cancer); 260 (preadmit malignant tumor); 199101 (malignant tumor)Immunocompromised279-279.9807 (HIV positive); 819 (immunocompromised); 892 (current medication immunosuppression)Autoimmune disease710-710.9, 714-714.9, 720.0803 (autoimmune disease); 279400 (autoimmune disease)Major organ disease277-277.01, 571-572.8, 582-583.9, 585-587806 (cystic fibrosis); 809 (chronic liver disease); 833 (chronic renal disease); 571500 (cirrhosis); 573305 (hepatitis)Mental retardation315, 317-319825 (mental retardation); 835 (developmental delay)Congenital anomaly740-742.9, 748-751.9, 756-756.9, 758-758.9836 (congenital anomaly)&ast;ICD-9-CM indicates International Classification of Diseases, Ninth Revision, Clinical Modification; HIV, human immunodeficiency virus.STATISTICAL ANALYSISSince distributions for LOS and hospital charge data are skewed, all LOS and charge data were transformed to the natural logarithm scale to create distributions that are closer to a normal distribution. For all statistical analyses, mean (±SD) and 95% confidence intervals were based on log scale values and transformed back to original units corresponding to the geometric mean for each category. The Student ttest was used to compare differences in LOS and charges between chronic disease groups and those without chronic disease. When reporting average LOS and charges, we have chosen to report both the geometric and arithmetic mean. The arithmetic mean has the advantage in that it reflects the actual resources and charges experienced in this data set. The geometric mean has the advantage in that outlier observations are deemphasized in the calculation; the resulting distributions are normally distributed and produce more correct statistical tests. Characteristics of the study hospitals were compared with those of the general and children's hospitals as represented in the 1991 American Hospital Association Annual Surveyusing the Student ttest or the χ2test.FINANCIAL ANALYSISWe compared the estimated profit margins of children's hospitals with those of general hospitals when contracts do not differentiate between patients with or without chronic diseases. It was assumed that the insurer's contract price (π) for a given condition or procedure (with and without chronic diseases) reflects the lower of either the children's hospital or the general hospital's average cost for treating the acute condition or performing the acute procedure. This pricing may occur if, for example, the insurer knew the costs at both hospitals and had the market power to enforce the minimum price. This assumption may help approximate the situation where general hospitals may offer bids close to their marginal cost in fear that another general hospital may enter the market as a close substitute, and children's hospitals may offer bids close to their marginal costs since they are forced to engage in competitive contracts to maintain the large volume of patients needed for their size of operation. Since each hospital has a mix of both patients with no chronic illness (N) and patients with chronic illness (C), we would expect to observe:πNΠ<πC, where πNand πCrepresent the total cost of care for patients with a specified disease or procedure with no chronic disease (πN) and with at least 1 chronic disease (πC). The number of cases at the children's hospital equals the number of cases with chronic disease (CC) plus the number of cases without chronic disease (CN), and similarly, the number of cases at the general hospital is GC+GN. We also assume that the children's hospital and the general hospital face identical costs when treating identical patients, and that the cost of care is a function of the procedure/condition and the presence of a chronic disease. Hence, the price that the children's hospital would offer to treat a given disease would be the average cost of treating all patients with that disease (with and without chronic diseases) at that hospital, or &lcub;[CCπC+CNπN]/(CC+CN)&rcub;, and the price offered by the general hospital to treat the same disease would be &lcub;[GCπC+GNπN]/(GC+GN)&rcub;.Revenue at each hospital is defined as the number of patients receiving care with and without chronic diseases multiplied by the contract price per case. The cost to hospitals is assumed to be the price of caring for patients with and without chronic diseases (πCand πN) multiplied by the numbers of patients with and without chronic diseases. Profit for the hypothetical children's hospitals or general hospitals is defined as the difference between revenue and cost.In the analysis, it was assumed that the cost-to-charge ratio across hospitals and conditions was uniform, so charges were used as a proxy for costs. The use of charges, rather than costs, was deemed appropriate as long as the cost-to-charge ratio did not vary systematically within disease categories across groups of patients with and without chronic diseases. Since the primary analytical focus concerned the relative costs of caring for a similar patient with and without chronic diseases, costs could be substituted for charges without biasing the analysis, under the assumption of a constant cost-to-charge ratio; however, it is possible that charges may sometimes overestimate costs for longer hospitalizations, resulting in overestimates of costs for patients with chronic illness. We also report an analysis that subtracted 40% of the average costs per day when LOS in the chronically ill group exceeded the LOS in the group without a chronic disease.In such an analysis the estimate of increased cost for caring for the chronically ill is adjusted downward, since longer LOS may not always imply proportionally greater costs if little treatment is performed, or few resources used, in the last days of a hospitalization.We consider a situation in which only 2 hospitals are competing in a single market, one is a children's hospital, and one a general hospital. All the pediatric cases are shared between these hospitals, and it is assumed that equal numbers of pediatric patients are admitted to each hospital for the condition or procedure of interest. The overall percentage of patients with chronic disease in the common patient population analyzed in our study is assumed to be 15.6%, as empirically determined from our database. Based on these simplifying assumptions, we model the per-patient profit or loss in children's and general hospitals as a function of each hospital's share of patients with chronic illnesses. For a more detailed description of the model, see Appendix.RESULTSHOSPITAL CHARACTERISTICSThe analysis included 163 hospitals. There were 186±202 study cases per hospital. The characteristics of these study hospitals are compared with all other acute care hospitals that provided pediatric inpatient care listed in the American Hospital Association's 1991 Annual Survey in Table 3. The proportion of children's hospitals and the percentage of pediatric admissions in our sample were similar to the American Hospital Association sample of hospitals excluded from the study; however, the characteristics of the general and children's hospitals in our study showed many differences that were statistically significant. Among them, study hospitals tended to be larger, concentrated in more urban areas, were less likely to be for-profit, had higher nurse-to-bed ratios, had a higher technological mix, and engaged in more teaching than nonstudy hospitals in the American Hospital Association survey.Table 3. Comparison of Study Hospitals vs All Other American Hospital Association General and Children's Hospitals&ast;CharacteristicStudy Hospitals (n=163)American Hospital Association Hospitals Not in Study (n=5470)No. of beds296 (221)175 (180)†Metropolitan area and population >250,00066.342.5†Nongovernment, for profit1.213.3†Full-time physicians5.69 (11.59)7.79 (34.83)‡Nurse-to-bed ratio0.75 (.26)0.55 (.34)†Burn care service3.12.5Trauma center24.112.3†Magnetic resonance imaging facility32.117.6†Bone marrow transplant unit10.53.4†Approved residency training program36.818.2†Member of the American Council of Teaching Hospitals12.36.5†Children's hospital0.610.77Pediatric-to-total medical staff ratio0.061 (0.045)0.047 (0.056)†Primarily pediatric admissions0.620.82&ast;Values are presented as either mean (SD) or percentage.†P<.0005.‡P<.05.DESCRIPTION OF PATIENTSThe analysis included 30,379 patient admissions, of which 54.8% were males. The mean age was 7.2±5.2 years. Of the 30,379 admissions, 4737 patients (15.6%) had at least 1 of the selected chronic diseases. In the tables to follow, a description of specific patient conditions and procedures is presented with their associated LOS and charge data.LENGTH OF STAYTable 4provides the geometric mean LOS for all procedures and conditions according to each chronic disease category. Of note, patients with cerebral palsy displayed a 48% increase in LOS compared with those without a chronic disease (P<.001). Patients with major organ disease (cystic fibrosis, kidney or liver disease) displayed a 54% increase in LOS (P<.001) compared with those without chronic disease.Table 4. Summary of Length of Stay by Diagnosis Related Group and Chronic DiseaseChronic Disease&ast;Total†No Chronic DiseaseDiabetes MellitusSickle Cell AnemiaCerebral PalsySeizuresAsthmaConditionNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSConcussion11711.3710751.3631.590. . .11.00431.55361.38Croup20881.7217951.7031.260. . .152.42522.061361.79Pneumonia77773.3355153.23134.15433.981394.533973.8214333.40Appendicitis34133.2531653.24103.9253.1023.00333.741192.97Gastroenteritis86242.2976312.26722.19133.20802.922702.593202.14Fractures20771.9819421.9611.0021.4153.35292.19772.37Cellulitis9843.258693.2342.6322.6553.52283.16443.40Urinary tract infection13523.5211573.4493.3013.00155.70614.29303.27Viral illness28932.6424932.60142.69123.67122.911292.59843.08Total30,3792.6125,6422.531292.52‡783.58§2743.75§10423.03§22792.95§&ast;Chronic diseases are not mutually exclusive. LOS indicates length of stay; ellipses, not applicable.†Mean based on log scale values transformed back to length of stay expressed as days.‡Not significant for the comparison with no chronic disease.§P=.001 for the comparison with no chronic disease.&par;P=.005 for the comparison with no chronic disease.Table 5compares the arithmetic and geometric mean LOS among all patients with a specific condition or procedure with values calculated for patients without chronic disease and with at least 1 chronic disease. Across all conditions, patients with chronic diseases had a longer LOS than patients without a chronic disease, although LOS differences for concussion, croup, appendicitis, and cellulitis failed to reach statistical significance.Table 5. Comparison of Length of Stay for Various Diagnosis Related Groups of Patients, According to Chronic DiseaseTotalProcedure/ConditionNo. of PatientsArithmetic Mean (SD)Geometric Mean (95% CI&ast;)Concussion11711.57 (1.15)1.37 (1.33, 1.40)Croup20882.01 (1.32)1.72 (1.68, 1.76)Pneumonia77773.86 (2.47)3.33 (3.29, 3.37)Appendicitis34133.86 (2.69)3.25 (3.19, 3.31)Gastroenteritis86242.77 (2.46)2.29 (2.26, 2.32)Fractures20773.61 (5.92)1.98 (1.91, 2.06)Cellulitis9843.85 (2.49)3.25 (3.14, 3.37)Urinary tract infections13524.01 (2.17)3.52 (3.42, 3.62)Viral illness28933.05 (1.81)2.64 (2.59, 2.69)Total30,3793.25 (2.80)2.61 (2.59, 2.63)&ast;Geometric mean and 95% confidence interval (95% CI) based on log scale values transformed back to length of stay expressed as days.†Student t test, compared with no chronic disease groups, using log-transformed data.CHARGESTable 6provides the arithmetic mean charge for all procedures and conditions according to each chronic disease group. The most expensive chronic disease group, major organ disease (cystic fibrosis, kidney or liver disease), showed a 79% increase over those patients without a chronic disease (P<.001). Patients with cancer showed a 71% increase over those without a chronic disease (P<.001).Table 6. Summary of Total Charge Dollars by Diagnosis Related Groups and Chronic DiseaseChronic Disease&ast;Total†No Chronic DiseaseDiabetes MellitusSickle Cell AnemiaCerebral PalsySeizuresAsthmaConditionNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSConcussion1171200510751956315640. . .11018433555361951Croup2088169817951620319310. . .1522555224211361799Pneumonia77773173551527791343484341191395939397508714333660Appendicitis341347703165474110552054296256443352271194581Gastroenteritis862419607631183572267513637480273127026783201946Fractures207731891942316016472159655373293587773894Cellulitis98429408692877423782224652640283012443310Urinary tract infection13523089115729279313312506154858614127303324Viral illness28932396249322731428661226631228971292516843173Total30,379277825,64226141293049‡784149§2744517§10423799§22793308§&ast;Chronic diseases are not mutually exclusive. LOS indicates length of stay; ellipses, not applicable.†Arithmetic mean.‡Not significant for the comparison with no chronic disease.§P=0.0001 for the comparison with no chronic disease.The relationship observed between chronic disease and hospital charges was similar to that observed between chronic disease and LOS. The mean hospital charges for each procedure/condition group are displayed in Table 7. The presence of chronic disease increased hospital charges for all conditions studied. We report both arithmetic and geometric mean values, the former reflecting all patients equally, the latter allowing for a decreased influence from patient charge outliers. Use of the geometric mean, through reducing the influence of outliers, may better reflect a hospital's ability to pass on extreme charge outliers through reinsurance contracts. The geometric mean hospital charge for patients without chronic diseases was $2041 per admission compared with $2632 for patients with at least 1 chronic disease, a difference of $591 per patient (P<.001). The arithmetic mean hospital charge for patients without chronic diseases was $2614 per admission compared with $3663 for patients with at least 1 chronic disease, a difference of $1049 per patient (P<.001).Table 7. Comparison of Total Charge for Various Diagnosis Related Groups of Patients, According to Chronic Disease&ast;TotalProcedure/ConditionNo. of PatientsArithmetic Mean (SD)Geometric Mean (95% CI)Concussion11712004 (2021)1526 (1467, 1588)Croup20881698 (1699)1384 (1350, 1419)Pneumonia77773173 (3597)2497 (2463, 2532)Appendicitis34134770 (3402)4108 (4039, 4179)Gastroenteritis86241960 (2463)1564 (1544, 1584)Fractures20773189 (3168)2376 (2303, 2452)Cellulitis9842940 (2513)2347 (2256, 2442)Urinary tract infection13523098 (2064)2617 (2539, 2697)Viral illness28932396 (1978)1979 (1936, 2022)Total30,3792778(2992)2124(2108, 2140)&ast;Geometric mean and 95% confidence interval (95% CI) based on log scale values transformed back to total charge in dollars.†Student t test, compared with no chronic disease groups, using log-transformed data.PATIENTS WITH MULTIPLE CHRONIC DISEASESOf the 4737 patients with chronic disease, our study identified 1072 patients (22.6%) with 2 or more chronic diseases, occurring predominantly in the gastroenteritis (n=261) and pneumonia (n=536) groups. Overall, these patients had 31.2% longer LOSs than those with just 1 chronic disease and 58.5% higher charges. Admissions associated with more than 1 chronic disease accounted for 27.8% of hospital days for the chronically ill, and 31.7% of the total hospital expenditures on chronically ill children with common acute conditions. Patients with multiple chronic diseases experienced greater charges than did those with a single chronic disease. Patients with 1 chronic disease had a mean (arithmetic) charge of $3235 compared with $5126 in patients with 2 or more chronic diseases (P<.001).FINANCIAL ANALYSISChange in per-patient profit or loss at the children's hospital and general hospital was a function of the relative percentage of patients with chronic diseases treated at the children's and general hospitals. Table 8displays profit and loss for the children's hospital by the percentage of the population of patients with chronic disease treated at the children's hospital rather than the general hospital. If children's hospitals care for 75% of the patients with chronic illnesses and acute conditions, the loss per acute condition is 2.13% per case, based on the geometric mean and up to 2.86% per patient when using the arithmetic mean (fully counting outlier charges). Adjusting for the possible bias in cost-to-charge due to long LOS, the loss per patient is 1.5% per case (geometric) and 2.0% (arithmetic).Table 8. Children's Hospital per-Patient Profit (Loss) by Share of Chronically Ill and Outlier Assumptions65% Share of Chronically Ill Cases75% Share of Chronically Ill Cases85% Share of Chronically Ill CasesProcedure/ConditionGeometric Analysis&ast;Arithmetic Analysis†Geometric AnalysisArithmetic AnalysisGeometric AnalysisArithmetic AnalysisConcussion−0.44−0.73−0.74−1.20−1.04−1.68Croup−0.80−1.37−1.32−2.26−1.84−3.13Pneumonia−2.50−3.59−4.10−5.85−5.65−8.00Appendicitis−0.01−0.02−0.03−0.03−0.04−0.04Gastroenteritis−0.99−1.88−1.65−3.10−2.29−4.28Fractures−0.11−0.28−0.19−0.47−0.28−0.65Cellulitis−0.63−0.64−1.04−1.06−1.46−1.48Urinary tract infection−1.08−1.55−1.79−2.55−2.49−3.54Viral illness−1.13−1.52−1.86−2.50−2.59−3.47Total−1.29−1.73−2.13−2.86−2.96−3.96&ast;This method reduces the influence of patient outliers under an assumption that hospitals may obtain other funding for extreme outlier patients. All values are expressed as percentages of profit (loss).†This method assumes that the full cost of all patient outliers is paid by the hospital.Figure 1focuses on an especially costly group: patients admitted for pneumonia. Since it is assumed that the insurer pays the lower of the 2 average cost figures from either hospital, and that the cost structures of the 2 hospitals are a function only of their percentage of patients with and without chronic disease, there is no "excess" profit for either the children's or general hospital. Loss occurs in treating pneumonia when either hospital acquires a disproportionate share of patients with chronic illness and pneumonia, since treatment costs for pneumonia are higher in patients with a chronic illness than in those without a chronic illness. If children's hospitals care for 75% of the patients with chronic illness and acute pneumonia, the loss per case of acute pneumonia is 4.1% (geometric) and 5.85% (arithmetic). When we adjust for the possible bias in cost-to-charge ratios based on formulas relating to LOS, we get a loss of 4.0% per case (geometric) and 5.0% (arithmetic).The percentage of profit and loss per patient with pneumonia for the children's and general hospital as a function of case-mix fraction, assuming 29% of pneumonia cases have at least 1 chronic disease and an equal sample size across hospitals.COMMENTProviding hospital care to pediatric patients admitted for common procedures and conditions is more costly when a child has a chronic disease. In a recent study of pediatric Medicaid expenditures in Washington State, Ireys et alfound that patients with chronic illnesses (defined in a manner similar to our study) represented 5.9% of this population, and that patients with 1 or more chronic conditions had, on average, more than a 4-fold increase in total yearly Medicaid expenditures, of which approximately half of these expenditures were for inpatient services. Similar trends have been observed in adults with chronic disease.Our figure of 15.6% of pediatric hospital admissions being composed of patients with chronic illnesses is consistent with the Washington State experience. Not surprisingly, our study identified both meaningful and statistically significant increases in LOS and total hospitalization cost, as reflected through charges,when hospitalization for a routine pediatric condition or procedure is accompanied by a chronic medical condition. The size of the effect of chronic disease varied, depending on the specific underlying procedure or condition, but was found to be important for most of the conditions noted in the study.Newacheck and Taylor,using data from the 1988 National Health Interview Survey, reported that approximately 31% of children have some chronic disease. Only 5% of these patients had severe conditions; however, these 5% accounted for 33% of hospital expenditures for patients with chronic illnesses as a whole. Newacheck and Stoddardhave reported approximately 5% of patients had 2 or more chronic diseases, similar to the figure we found in our study. Our study shows that the 3.5% of admissions associated with more than 1 chronic disease accounted for 27.8% of hospital days for patients with chronic illnesses, and 31.7% of the total hospital expenditures on children with chronic illnesses and common acute conditions. One may conclude that patients with chronic illness, either in the hospitalized population (as in our study), or in the population as a whole (as in the Ireysor Newacheckstudies) represent a nonhomogeneous group of considerable size, with notable effect on total pediatric expenditures.Market changes are creating new difficulties in securing adequate funds for the care of chronic illness. Neff and Anderson,citing data from the National Association of Children's Hospitals and Related Institutions, point out that children's hospitals or hospitals that have postgraduate pediatric residency programs account for only 5% of hospitals, yet these hospitals admit 70% to 80% of the more complex chronic pediatric health conditions. A major distinguishing feature between pediatrics and adult medicine is the high concentration of pediatric specialists at children's hospitals, where for some specialties there are almost no practitioners at community hospitals.Therefore, the complex patients of specialists, often with chronic diseases, tend to be admitted to children's hospitals. To prevent market failure, Neff and Anderson have argued for"carve-out arrangements" and risk adjusters that take into account the higher than average costs of caring for children with chronic diseases. This will make possible adequate payments for health care plans that care for these patients and reduce the possibility that children with chronic illnesses will receive inappropriate care.(p1869)The model developed in this analysis is simple, yet, we believe useful in helping one understand the problems faced by children's hospitals or tertiary care centers specializing in pediatrics. By providing examples of per-patient loss, we believe the model helps us understand the magnitude of the problem outlined by Neff and Anderson from the perspective of the hospital. Faced with a loss per patient, the model suggests that hospitals may improve their economic situation by discouraging the admission of patients with chronic illnesses. This may be financially the right course to take, but would subvert the mission of children's hospitals. A preferred solution would be to ensure that the market price for acute conditions and procedures properly reflects the cost of providing such services, so that there is no disincentive to care for patients with chronic illnesses.Whenever contract prices do not reflect the full cost of care, market forces will cause some undesirable effects. The problem of unaccounted costs of chronic diseases occurs in adult medicine, but on a slightly smaller scale since chronic conditions are more common in adult medicine and have demanded more attention.Medicare's DRGs theoretically account for the major comorbidities of elderly patients. These reflect some of the increased resource utilization needs of patients with chronic disease; however, asymmetries in severity and case-mix continue to be a problem,and as competition increases throughout the hospital sector, such asymmetries in disease burden will cause problems similar to those described here, for all tertiary care hospitals.The market forces resulting from an asymmetric financial burden of disease may act to correct this disparity among hospitals in ways that are not necessarily in the best interest of society, and certainly not in the best interests of patients with chronic illnesses. Health economist Mark V. Pauly, PhD, has recently described an analogous problem concerning care for the uninsured. As he points out, " . . . hospitals with a heavy charity care . . . load may therefore find it impossible to provide care that they cannot avoid while still covering their costs. These firms will then cease to exist. The nonpaying patients who used to use these hospitals will go to other hospitals. . . ."(p359) Pauly goes on to predict that the uninsured will be distributed more evenly over the remaining hospitals, and that the care provided at these hospitals will be of lower quality. A similar argument could be made concerning the consequences of uncompensated care associated with treating patients with chronic illness.While children's hospitals need not exit the marketplace, one may expect they would either move to reduce the number of patients with chronic illness that they care for or lower their expenditures on these patients through lower quality of care. There are many ways to lower the quality of care provided to patients with chronic illness. Ireys et alhave outlined several major areas of concern regarding the adequacy of the current health service system for children with disabilities and chronic illnesses. Indicators of quality erosion may be found not only by studying the content of care, but also changes in the goals for care. Downward revisions of expectations may be an insidious method of cost reduction. Areas of concern can be manipulated by hospitals to save resources, yet in the long run, erosion of related services may lead to lower quality and worse outcomes.These observations should be significant for children's hospitals negotiating changes in their third-party contractual relationships. In the new capitated environment, children's hospitals, by necessity, require a far larger population of households to support their infrastructure than most adult hospitals, since the need for hospital services is far less frequent in childhood than in old age. Given the financial necessity at the children's hospitals for a sufficient volume of patients to maintain viable tertiary care and teaching programs, children's hospitals must enter competitive contracts with insurers, often offering the same price as the general hospitals for services performed at both hospitals.If market contracts for common procedures and conditions do not include chronic disease adjustments, children's hospitals will be placed at a financial disadvantage, precisely because the mission of most children's hospitals and pediatric tertiary care facilities is to provide care for the most complicated patients.Our analysis likely underestimated the problems facing children's hospitals. Charges in this analysis reflect resource utilization at general hospitals since most of the admissions in our study were to general hospitals. Thus, the analysis did not fully account for increases in service intensity at the children's hospital, that would increase the resource utilization, and hence, cause even greater loss for children's hospitals than those described in Table 8.Our analysis also did not address quality of care, and provided no premium for increased technological investment that may be required to treat specific pediatric populations.This analysis was limited to common pediatric problems. The observed effect of chronic disease on resource utilization may actually be stronger for conditions with lower incidence, which are even more likely to be treated at specialized children's hospitals. Children with rare conditions and chronic diseases are, with high probability, referred to, and remain in, the children's hospital systems. By concentrating on common procedures, we have likely underestimated the potential difficulties facing children's hospitals. Furthermore, by not defining all chronic diseases in our study population, we have overestimated the charges associated with the patients "without" chronic disease because this group reflects some patients who do have chronic disease. The difference between costs is likely greater than we estimated and hence children's hospital losses are also likely greater than we estimated.CONCLUSIONSChildren's hospitals should see a disproportionate share of patients with chronic disease. Our study has shown that for common acute conditions, children with chronic disease are more expensive to care for in the hospital than children without a chronic disease. Children's hospitals will not succeed in solving their financial problems by entering into high-volume contracts simply defined by the acute condition or procedure. Hospitals that care for a greater percentage of children with chronic conditions must seek contracts that better reflect the increased resources needed to treat such patients—even for common acute conditions and procedures. If not, market forces will act to change the mission of children's hospitals, either by reducing access of the chronically ill to children's hospitals or by lowering the quality of care provided at these hospitals.APPENDIXWe assume that there are 2 hospitals, a children's hospital (C) and a general hospital (G). The total number of pediatric patients seen at the children's hospital is a mixture of patients with chronic disease (CC) and those with no chronic disease (CN), while the total number of pediatric patients seen at the general hospital is a mixture of patients with chronic disease (GC) and those with no chronic disease (GN). We will refer to CC, CN, GC, and GNas fractions of 100% of the pediatric admissions in the town. Hence: CC+CN+GC+GN=1.0.Furthermore, we will assume that the total number of patients at the children's hospital is equal to the total number of patients at the general hospital, so that: CC+CN=GC+GN=0.5.The overall rate of chronic disease in the population (for a given disease or group of diseases) is defined as k, where k=CC+GCand the percentage of patients with chronic diseases who are treated at the children's hospital is defined as M, where: M=CC/(CC+GC)= CC/k.We can solve for the value of CC, CN, GC, and GNif values of k and M are specified. The above equations yield the following solutions: CC=kM; CN=0.5-kM; GC=k(1-M); and GN=0.5-k(1-M).We assume 1 insurer pays for all pediatric visits, and each hospital will contract for specific diagnoses. The final contract price offered by the insurer (π) will be the minimum average cost of treating all patients with a specified acute condition for either the general or children's hospital.Profit equations, where profit at the children's hospital is defined as PCand profit at the general hospital is defined as PGare as follows: PC=&lcub;[(π-πC)CC]+[(π-πN)CN]&rcub; and PG=&lcub;[(π-πC)GC]+[(π-πN)GN]&rcub;, where π is defined as the contract price, which is the minimum of either the price bid by the children's hospital or general hospital. The average cost at the children's hospital is: &lcub;(πCCC+πNCN)/(CC+CN)&rcub;.The average cost at the general hospital is: &lcub;(πCGC+πNGN)/(GC+GN)&rcub;.MGGoldfarbRMCoffeyChange in the Medicare case-mix index in the 1980s and the effect of the prospective payment system.Health Serv Res.1992;27:385-415.JGabelTen ways HMOs have changed during the 1990s.Health Aff (Millwood).1997;16:134-145.EGGayJJKronenfeldRegulation, retrenchment—the DRG experience: problems from changing reimbursement practice.Soc Sci Med.1990;31:1103-1118.RCCoile JrAssessing healthcare market trends and capital needs: 1996-2000.Health Care Financial Manage.1995;49:60-65.JHMaxwellHMSapolskyThe first DRG: lessons from the end stage renal disease program for the prospective payment system.Inquiry.1987;24:57-67.HFroehlichWRJarvisEconomic impact of diagnosis-related groups and severity of illness on reimbursement for central nervous system infections.J Pediatr.1991;118:693-697.LKLichtigRAKnaufRHParrottJMuldoonRefining DRGs: the example of children's diagnosis-related groups.Med Care.1989;27:491-506.AASkolnickChildren's hospitals say about health care reform, ‘one size won't fit all.'JAMA.1993;270:2151-2152.PDPhelanKBaxterJBishopIHudsonDHindleDo diagnosis related groups separate the case-mix of a specialist children's hospital and a paediatric unit in a general hospital?J Paediatr Child Health.1993;29:266-269.JCGayJHMuldoonJMNeffLJWingProfiling the health service needs of populations: description and uses of the NACHRI classification of congenital and chronic health conditions.Pediatr Ann.1997;26:655-663.JLFreemanRBFetterHParkDiagnosis-related group refinement with diagnosis- and procedure-specific comorbidities and complications.Med Care.1995;33:806-827.HTIreysGFAndersonTJShafferJMNeffExpenditures for care of children with chronic illnesses enrolled in the Washington State Medicaid program, fiscal year 1993.Pediatrics.1997;100:197-204.MJLongJLDreachslinJFisherShould children's hospitals have special consideration in reimbursement policy?Health Care Finance Rev.1986;8:55-63.JMNeffGAndersonProtecting children with chronic illness in a competitive marketplace.JAMA.1995;274:1866-1869.ACBrewsterBGKarlinLAHydeCMJacobsRCBradburyYMChaeMedisGroups: a clinically based approach to classifying hospital patients at admission.Inquiry.1985;22:377-387.LIIezzoniMAMoskowitzA clinical assessment of MedisGroups.JAMA.1988;260:3159-3163.JWThomasMLFAshcraftMeasuring severity of illness: a comparison of interrater reliability among severity methodologies.Inquiry.1989;26:483-492.JWThomasMLFAshcraftMeasuring severity of illness: six severity systems and their ability to explain cost variations.Inquiry.1991;28:39-55.LIIezzoniEKHotchkinASAshMShwartzYMackiernanMedisGroups data bases: the impact of data collection guidelines on predicting in-hospital mortality.Med Care.1993;31:277-283.PMSteenACBrewsterRCBradburyEEstabrookJAYoungPredicted probabilities of hospital death as a measure of admission severity of illness.Inquiry.1993;30:128-141.DRDurbinAPGiardioKNShawMCHarrisJHSilberThe effect of insurance status on likelihood of neonatal interhospital transfer.Pediatrics(electronic journal). 1997;100:381-382.American Academy of Pediatrics Committee on Infectious Diseases1994 Redbook.Elk Grove Village, Ill: American Academy of Pediatrics; 1994:280.JHSilberPRRosenbaumJSSchwartzRNRossSVWilliamsEvaluation of the complication rate as a measure of quality of care in coronary artery bypass surgery.JAMA.1995;274:317-323.MShwartzDWYoungRSiegristThe ratio of costs to charges: how good a bias for estimating costs?Inquiry.1995/96;32:476-481.JRLaveLBLaveHospital cost functions.Am Econ Rev.1970;60:379-395.RMUngerleiderARBengurALKessenichRisk factors for higher cost in congenital heart operations.Ann Thorac Surg.1997;64:44-48.JRGButlerHospital cost analysis.In: Zweifel P, Frech HE, eds. Developments in Health Economics and Public Policy. Dordrecht, the Netherlands: Kluwer Academic Publishers; 1995;1-393.SLKaufmanDSShepardCosts of neonatal intensive care by day of stay.Inquiry.1982;19:167-178.PFishmanMVon KorffPLozanoJHechtChronic care costs in managed care.Health Aff (Millwood).1997;16:239-247.PWNewacheckWRTaylorChildhood chronic illness: prevalence, severity, and impact.Am J Public Health.1992;82:364-371.PWNewacheckJJStoddardPrevalence and impact of multiple childhood chronic illnesses.J Pediatr.1994;124:40-48.SFJencksADobsonRefining case-mix adjustment: the research evidence.N Engl J Med.1987;317:679-686.SFJencksDKWilliamsTLKayAssessing hospital-associated deaths from discharge data: the role of length of stay and chronic disease.JAMA.1988;260:2240-2246.JGreenNWintfeldPSharkeyLJPassmanThe importance of severity of illness in assessing hospital mortality.JAMA.1990;263:241-246.JGreenLJPassmanNWintfeldAnalyzing hospital mortality: the consequences of diversity in patient mix.JAMA.1991;265:1849-1853.SAJoyRWYurtAn all-payor prospective payment system (PPS) based on diagnosis-related groups (DRG): financial impact on reimbursement for trauma care and approaches to minimizing loss.J Trauma.1990;30:866-873.JFederJHadleySZuckermanHow did Medicare's prospective payment system affect hospitals?N Engl J Med.1987;317:867-873.SOzatalayRBroylesNet returns, fiscal risks, and the optimal patient mix for a profit-maximizing hospital.J Med Syst.1987;11:331-347.MVPaulyTrading cost, quality and coverage of the uninsured: what will we demand and what will we supply?In: Altman S, Reinhardt U, Shields A, eds. The Future U.S. Health Care System: Who Will Care for the Poor and Uninsured?Chicago, Ill: Health Administration Press; 1997:353-373.HTIreysHAGrasonBGuyerAssuring quality of care for children with special needs in managed care organizations: roles for pediatricians.Pediatrics.1996;98:178-185.SAGliedSGnanasekaranHospital financing and neonatal intensive care.Health Serv Res.1996;31:593-607.CAHinzPediatric oncology patients: no future profit center.Health Care Strategic Manage.1991;9:1,19-22.Accepted for publication July 15, 1998.This research was funded in part by grant RO1-HS06560 from the Agency for Health Care Policy and Research, Rockville, Md, and the Ethel Brown Foerderer Fund for Excellence, Philadelphia, Pa (Dr Silber).Presented in part at the 15th Annual Association for Health Services Research National Meeting, Washington, DC, June 23, 1998.We thank William P. Greeley, MD, MBA, Mark V. Pauly, PhD, and Orit Even-Shoshan, MS for their helpful suggestions. We thank MediQual Systems Inc for their assistance in this research. Errors or omissions are solely our responsibility.Corresponding author: Jeffrey H. Silber, MD, PhD, The Children's Hospital of Philadelphia, Abramson Bldg, Room 702, 34th and Civic Center Blvd, Philadelphia, PA 19104-4318 (e-mail: Silberj@wharton.upenn.edu). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JAMA Pediatrics American Medical Association

The Influence of Chronic Disease on Resource Utilization in Common Acute Pediatric Conditions

Download PDF
11 pages

Loading next page...
 
/lp/american-medical-association/the-influence-of-chronic-disease-on-resource-utilization-in-common-nLJrt8W3Xb

References (42)

Publisher
American Medical Association
Copyright
Copyright 1999 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
ISSN
2168-6203
eISSN
2168-6211
DOI
10.1001/archpedi.153.2.169
Publisher site
See Article on Publisher Site

Abstract

ObjectivesTo estimate the resource utilization in hospitalizations for common pediatric conditions or procedures involving patients with chronic disease vs those with no chronic disease and to develop an economic model of hospital per-patient profit (or loss) when insurance contracts fail to account for the presence of chronic disease.Setting and DesignA retrospective analysis of selected acute pediatric conditions found in the 1991 and 1992 MedisGroups National Comparative Data Base.PatientsWe studied 30,379 pediatric admissions for common acute conditions, including concussion, croup, pneumonia, appendicitis, gastroenteritis, fractures, cellulitis, urinary tract infection, and viral illness.Main Outcome MeasuresHospital length of stay and total hospital charges.ResultsFor patients without chronic disease, mean (geometric) length of stay was 2.53 vs 3.05 days (P<.001) for patients with at least 1 chronic disease. For patients without chronic disease, mean (arithmetic) total hospital charge was $2614 vs $3663 (P<.001) for patients with at least 1 chronic disease. Assuming 75% of patients with chronic disease are admitted to a children's hospital vs 25% to a general hospital, overall loss per patient at the children's hospital ranged between 1.5% and 2.9%, depending on assumptions regarding cost-to-charge ratios and the treatment of charge outliers. Pneumonia cases were associated with a 4.0% to 5.85% loss.ConclusionsLength of stay and charges are higher for everyday pediatric conditions or procedures when patients also have a chronic disease. If insurance contracts fail to account for chronic disease, then children's hospitals will realize significant financial losses, and over time this will lead to a decline in their financial viability, a reduction in quality, or a change in their mission.Editor's Note:Data provided in this study should be of little surprise to those who work in hospitals dedicated to children. I would be (pleasantly) surprised if it helps to convince payors. For children with chronic illnesses, DRG seems to mean "Da Revenue's Gone."—Catherine D. DeAngelis, MDMAJOR CHANGES have been occurring in the systems used to finance health care in the United States. Fee-for-service reimbursement rapidly is being replaced by a prospective, often capitated system.This new financial structure, combined with lower bed occupancy rates, has forced hospitals to compete not only along the traditional dimension of quality, but also along the dimensions of price and differentiation of service offerings.Since the introduction of the diagnosis related group (DRG) classification system for prospective payment from Medicare in 1983, adult medical care has been changing and adapting to these new financial realities.Reimbursement patterns in pediatric medicine have been slower to change. It has been only in the last 5 years that a notable portion of the pediatric market began to shift from a fee-for-service to a prospective payment system.This shift has caught many pediatric hospitals off guard, as little data exist to guide the development of prospective payment rates for pediatric diagnoses,although there is some evidence of improved accuracy in refinements of the DRG classification system for children.Our study explores 2 fundamental issues confronting pediatric tertiary care centers that enter competitive prospective contracts for common pediatric procedures or conditions: (1) how does the presence of patient chronic disease influence resource utilization and length of stay (LOS) for common pediatric conditions? and (2) if children's hospitals (defined as either children's hospitals or academic pediatric tertiary care centers) provide care for common pediatric conditions, but treat a disproportionate number of these common conditions in complicated patients with 1 or more chronic diseases, what effect will this have on their financial solvency, when health care is financed prospectively in a competitive market?To determine the financial influence of chronic disease on total charges and LOS, we estimated the influence of chronic diseases on a number of common pediatric conditions and procedures. The first objective was to test the hypothesis that the presence of a chronic disease increases the hospital LOS and resource utilization in selected routine pediatric admissions. Confirmation of this hypothesis would suggest that chronic disease adjusted case-mix should be accounted for when negotiating prospective payment contracts involving common conditions in pediatric populations, an item not usually included in pediatric contracts.The second objective was to use these results to determine the financial influence of prospective payment schemes when differences in chronic disease rates across hospitals are not included in negotiated contract prices, a situation that may reflect the predicament in which many specialty children's hospitals throughout the United States presently find themselves.MATERIALS AND METHODSHOSPITAL AND PATIENT SELECTIONThe 1991 and 1992 MedisGroups National Comparative Data Base (MediQual Inc, Westborough, Mass), was used for this study. It contained 259,358 pediatric admissions (patients aged 0-18 years). The MedisGroups' database included pediatric admissions to 163 hospitals during consecutive periods in 1990 and 1991. At least 4 months of data were available for each hospital. The hospitals chosen are a representative sample of MedisGroups clients, but are not a random sample of hospitals in the United States. The method of data collection has been described in detail elsewhereand consisted of detailed medical record abstraction for up to 3 hospital reviews occurring on admission, and possibly at mid stay and at discharge from the hospital or at death. Each review included up to a potential of 700 data elements called Key Clinical Findings that reflect the physical examination, medical history, and diagnostic laboratory test results that occurred during the hospitalization. MedisGroups is the predecessor to MediQual's current performance measurement system, Atlas, that was introduced in 1996. The reliability and validity of MedisGroups data collection compare favorably with other commonly used data collection systems,although the validation of the pediatric experience is more limited.CASE SELECTIONWe selected all pediatric patients in the study database admitted to general or children's hospitals. Admissions were grouped by DRG, and the DRGs (using DRGs both with and without complication in a paired grouping) were ranked in decreasing order of frequency. Of the 30 most common DRGs, we selected 9 DRGs that generally represented simple, common pediatric problems. This selection resulted in a study data set of 30,379 patient admissions across 163 hospitals. A list of each procedure/condition grouping, and the associated DRG code is provided in Table 1.Table 1. MedisGroups Patients 18 Years Old or YoungerProcedure/Condition Group NameDiagnosis Related Group CodeDiagnosis Related GroupConcussion33Concussion, age 0-17 yCroup71LaryngotracheitisPneumonia91Simple pneumonia and pleurisy, age 0-17 yAppendicitis164Appendectomy, complicated principal diagnosis, with chronic condition (CC)165Appendectomy, complicated principal diagnosis, without CC166Appendectomy, without complicated principal diagnosis, with CC167Appendectomy, without complicated principal diagnosis, without CCGastroenteritis184Esophagitis, gastroenteritis and miscellaneous digestive disorders, age 0-17 yFractures235Fractures of the femur252Fracture, sprain, strain and dislocation of forearm, hand, foot, age 0-17 y255Fracture, sprain, strain and dislocation of upper arm, lower leg excluding foot, age 0-17 yCellulitis279Cellulitis, age 0-17 yUrinary tract infection322Kidney and urinary tract infections, age 0-17 yViral illness&ast;422Viral illness and fever of unknown origin, age 0-17 y&ast;Excluded all patients with principle diagnosis 7806 (fever of unknown origin).DEFINITIONSNotable chronic diseases were defined as those concurrent diseases that impose additional clinical vulnerability in the patient, and that may, in turn, influence resource use. As an independent arbiter of clinical vulnerability, we selected those chronic conditions or chronic diseases that were recommended by the American Academy of Pediatrics Committee on Infectious Diseases for influenza immunization.This list serves as a proxy for clinical vulnerability, and therefore, we believe it is a good starting point for listing important chronic diseases that might adversely influence resource use. This list was supplemented with common neurologic diseases (seizures, cerebral palsy, and mental retardation), common birth defects (trisomy 21), and autoimmune diseases. Rare conditions were excluded, as the intent of the analysis was to identify important chronic disease categories for inclusion in prospective payment contracts, not to define the effects of all potentially important chronic diseases.Patients were identified as having a specific chronic disease on the basis of either an appropriate International Classification of Diseases, Ninth Revision, Clinical Modification(ICD-9-CM) code listed as 1 of the in-hospital discharge diagnoses, or a Key Clinical Finding that provides detailed clinical elements on the patient's hospital stay from initial review to discharge from the hospital. The MedisGroups Key Clinical Finding codes were used to supplement ICD-9-CMcodes in the event that limited numbers of fields for the ICD-9-CMcodes may have lead to an underreporting of chronic disease for some cases. Using both the ICD-9-CMand Key Clinical Finding codes, we increased our likelihood of identifying patients with the chronic diseases we selected. A list of each chronic disease group with the associated ICD-9-CMcode or MedisGroups Key Clinical Finding is provided in Table 2.Table 2. Definition of Chronic Diseases&ast;Chronic Disease GroupICD-9-CMCodesMedisGroups Key Clinical Findings Codes and DefinitionsDiabetes250-250.91805 (diabetes); 894 (current medication insulin)Sickle cell anemia282.6-282.69804 (chronic anemia)Cerebral palsy342-343.9821 (cerebral palsy)Seizures345-345.91820 (seizures); 780300 (witnessed seizure)Asthma493-493.91840 (chronic lung disease)Congenital heart disease745-747.49829 (congenital heart disease); 746700 (hypoplastic heart); 746890 (valve atresia); 746900 (major cardiac anomaly); 747210 (aortic arch abnormality)Trisomy 21758.0834 (Trisomy)Cancer140-239.9810 (cancer); 260 (preadmit malignant tumor); 199101 (malignant tumor)Immunocompromised279-279.9807 (HIV positive); 819 (immunocompromised); 892 (current medication immunosuppression)Autoimmune disease710-710.9, 714-714.9, 720.0803 (autoimmune disease); 279400 (autoimmune disease)Major organ disease277-277.01, 571-572.8, 582-583.9, 585-587806 (cystic fibrosis); 809 (chronic liver disease); 833 (chronic renal disease); 571500 (cirrhosis); 573305 (hepatitis)Mental retardation315, 317-319825 (mental retardation); 835 (developmental delay)Congenital anomaly740-742.9, 748-751.9, 756-756.9, 758-758.9836 (congenital anomaly)&ast;ICD-9-CM indicates International Classification of Diseases, Ninth Revision, Clinical Modification; HIV, human immunodeficiency virus.STATISTICAL ANALYSISSince distributions for LOS and hospital charge data are skewed, all LOS and charge data were transformed to the natural logarithm scale to create distributions that are closer to a normal distribution. For all statistical analyses, mean (±SD) and 95% confidence intervals were based on log scale values and transformed back to original units corresponding to the geometric mean for each category. The Student ttest was used to compare differences in LOS and charges between chronic disease groups and those without chronic disease. When reporting average LOS and charges, we have chosen to report both the geometric and arithmetic mean. The arithmetic mean has the advantage in that it reflects the actual resources and charges experienced in this data set. The geometric mean has the advantage in that outlier observations are deemphasized in the calculation; the resulting distributions are normally distributed and produce more correct statistical tests. Characteristics of the study hospitals were compared with those of the general and children's hospitals as represented in the 1991 American Hospital Association Annual Surveyusing the Student ttest or the χ2test.FINANCIAL ANALYSISWe compared the estimated profit margins of children's hospitals with those of general hospitals when contracts do not differentiate between patients with or without chronic diseases. It was assumed that the insurer's contract price (π) for a given condition or procedure (with and without chronic diseases) reflects the lower of either the children's hospital or the general hospital's average cost for treating the acute condition or performing the acute procedure. This pricing may occur if, for example, the insurer knew the costs at both hospitals and had the market power to enforce the minimum price. This assumption may help approximate the situation where general hospitals may offer bids close to their marginal cost in fear that another general hospital may enter the market as a close substitute, and children's hospitals may offer bids close to their marginal costs since they are forced to engage in competitive contracts to maintain the large volume of patients needed for their size of operation. Since each hospital has a mix of both patients with no chronic illness (N) and patients with chronic illness (C), we would expect to observe:πNΠ<πC, where πNand πCrepresent the total cost of care for patients with a specified disease or procedure with no chronic disease (πN) and with at least 1 chronic disease (πC). The number of cases at the children's hospital equals the number of cases with chronic disease (CC) plus the number of cases without chronic disease (CN), and similarly, the number of cases at the general hospital is GC+GN. We also assume that the children's hospital and the general hospital face identical costs when treating identical patients, and that the cost of care is a function of the procedure/condition and the presence of a chronic disease. Hence, the price that the children's hospital would offer to treat a given disease would be the average cost of treating all patients with that disease (with and without chronic diseases) at that hospital, or &lcub;[CCπC+CNπN]/(CC+CN)&rcub;, and the price offered by the general hospital to treat the same disease would be &lcub;[GCπC+GNπN]/(GC+GN)&rcub;.Revenue at each hospital is defined as the number of patients receiving care with and without chronic diseases multiplied by the contract price per case. The cost to hospitals is assumed to be the price of caring for patients with and without chronic diseases (πCand πN) multiplied by the numbers of patients with and without chronic diseases. Profit for the hypothetical children's hospitals or general hospitals is defined as the difference between revenue and cost.In the analysis, it was assumed that the cost-to-charge ratio across hospitals and conditions was uniform, so charges were used as a proxy for costs. The use of charges, rather than costs, was deemed appropriate as long as the cost-to-charge ratio did not vary systematically within disease categories across groups of patients with and without chronic diseases. Since the primary analytical focus concerned the relative costs of caring for a similar patient with and without chronic diseases, costs could be substituted for charges without biasing the analysis, under the assumption of a constant cost-to-charge ratio; however, it is possible that charges may sometimes overestimate costs for longer hospitalizations, resulting in overestimates of costs for patients with chronic illness. We also report an analysis that subtracted 40% of the average costs per day when LOS in the chronically ill group exceeded the LOS in the group without a chronic disease.In such an analysis the estimate of increased cost for caring for the chronically ill is adjusted downward, since longer LOS may not always imply proportionally greater costs if little treatment is performed, or few resources used, in the last days of a hospitalization.We consider a situation in which only 2 hospitals are competing in a single market, one is a children's hospital, and one a general hospital. All the pediatric cases are shared between these hospitals, and it is assumed that equal numbers of pediatric patients are admitted to each hospital for the condition or procedure of interest. The overall percentage of patients with chronic disease in the common patient population analyzed in our study is assumed to be 15.6%, as empirically determined from our database. Based on these simplifying assumptions, we model the per-patient profit or loss in children's and general hospitals as a function of each hospital's share of patients with chronic illnesses. For a more detailed description of the model, see Appendix.RESULTSHOSPITAL CHARACTERISTICSThe analysis included 163 hospitals. There were 186±202 study cases per hospital. The characteristics of these study hospitals are compared with all other acute care hospitals that provided pediatric inpatient care listed in the American Hospital Association's 1991 Annual Survey in Table 3. The proportion of children's hospitals and the percentage of pediatric admissions in our sample were similar to the American Hospital Association sample of hospitals excluded from the study; however, the characteristics of the general and children's hospitals in our study showed many differences that were statistically significant. Among them, study hospitals tended to be larger, concentrated in more urban areas, were less likely to be for-profit, had higher nurse-to-bed ratios, had a higher technological mix, and engaged in more teaching than nonstudy hospitals in the American Hospital Association survey.Table 3. Comparison of Study Hospitals vs All Other American Hospital Association General and Children's Hospitals&ast;CharacteristicStudy Hospitals (n=163)American Hospital Association Hospitals Not in Study (n=5470)No. of beds296 (221)175 (180)†Metropolitan area and population >250,00066.342.5†Nongovernment, for profit1.213.3†Full-time physicians5.69 (11.59)7.79 (34.83)‡Nurse-to-bed ratio0.75 (.26)0.55 (.34)†Burn care service3.12.5Trauma center24.112.3†Magnetic resonance imaging facility32.117.6†Bone marrow transplant unit10.53.4†Approved residency training program36.818.2†Member of the American Council of Teaching Hospitals12.36.5†Children's hospital0.610.77Pediatric-to-total medical staff ratio0.061 (0.045)0.047 (0.056)†Primarily pediatric admissions0.620.82&ast;Values are presented as either mean (SD) or percentage.†P<.0005.‡P<.05.DESCRIPTION OF PATIENTSThe analysis included 30,379 patient admissions, of which 54.8% were males. The mean age was 7.2±5.2 years. Of the 30,379 admissions, 4737 patients (15.6%) had at least 1 of the selected chronic diseases. In the tables to follow, a description of specific patient conditions and procedures is presented with their associated LOS and charge data.LENGTH OF STAYTable 4provides the geometric mean LOS for all procedures and conditions according to each chronic disease category. Of note, patients with cerebral palsy displayed a 48% increase in LOS compared with those without a chronic disease (P<.001). Patients with major organ disease (cystic fibrosis, kidney or liver disease) displayed a 54% increase in LOS (P<.001) compared with those without chronic disease.Table 4. Summary of Length of Stay by Diagnosis Related Group and Chronic DiseaseChronic Disease&ast;Total†No Chronic DiseaseDiabetes MellitusSickle Cell AnemiaCerebral PalsySeizuresAsthmaConditionNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSConcussion11711.3710751.3631.590. . .11.00431.55361.38Croup20881.7217951.7031.260. . .152.42522.061361.79Pneumonia77773.3355153.23134.15433.981394.533973.8214333.40Appendicitis34133.2531653.24103.9253.1023.00333.741192.97Gastroenteritis86242.2976312.26722.19133.20802.922702.593202.14Fractures20771.9819421.9611.0021.4153.35292.19772.37Cellulitis9843.258693.2342.6322.6553.52283.16443.40Urinary tract infection13523.5211573.4493.3013.00155.70614.29303.27Viral illness28932.6424932.60142.69123.67122.911292.59843.08Total30,3792.6125,6422.531292.52‡783.58§2743.75§10423.03§22792.95§&ast;Chronic diseases are not mutually exclusive. LOS indicates length of stay; ellipses, not applicable.†Mean based on log scale values transformed back to length of stay expressed as days.‡Not significant for the comparison with no chronic disease.§P=.001 for the comparison with no chronic disease.&par;P=.005 for the comparison with no chronic disease.Table 5compares the arithmetic and geometric mean LOS among all patients with a specific condition or procedure with values calculated for patients without chronic disease and with at least 1 chronic disease. Across all conditions, patients with chronic diseases had a longer LOS than patients without a chronic disease, although LOS differences for concussion, croup, appendicitis, and cellulitis failed to reach statistical significance.Table 5. Comparison of Length of Stay for Various Diagnosis Related Groups of Patients, According to Chronic DiseaseTotalProcedure/ConditionNo. of PatientsArithmetic Mean (SD)Geometric Mean (95% CI&ast;)Concussion11711.57 (1.15)1.37 (1.33, 1.40)Croup20882.01 (1.32)1.72 (1.68, 1.76)Pneumonia77773.86 (2.47)3.33 (3.29, 3.37)Appendicitis34133.86 (2.69)3.25 (3.19, 3.31)Gastroenteritis86242.77 (2.46)2.29 (2.26, 2.32)Fractures20773.61 (5.92)1.98 (1.91, 2.06)Cellulitis9843.85 (2.49)3.25 (3.14, 3.37)Urinary tract infections13524.01 (2.17)3.52 (3.42, 3.62)Viral illness28933.05 (1.81)2.64 (2.59, 2.69)Total30,3793.25 (2.80)2.61 (2.59, 2.63)&ast;Geometric mean and 95% confidence interval (95% CI) based on log scale values transformed back to length of stay expressed as days.†Student t test, compared with no chronic disease groups, using log-transformed data.CHARGESTable 6provides the arithmetic mean charge for all procedures and conditions according to each chronic disease group. The most expensive chronic disease group, major organ disease (cystic fibrosis, kidney or liver disease), showed a 79% increase over those patients without a chronic disease (P<.001). Patients with cancer showed a 71% increase over those without a chronic disease (P<.001).Table 6. Summary of Total Charge Dollars by Diagnosis Related Groups and Chronic DiseaseChronic Disease&ast;Total†No Chronic DiseaseDiabetes MellitusSickle Cell AnemiaCerebral PalsySeizuresAsthmaConditionNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSNo. of PatientsMean LOSConcussion1171200510751956315640. . .11018433555361951Croup2088169817951620319310. . .1522555224211361799Pneumonia77773173551527791343484341191395939397508714333660Appendicitis341347703165474110552054296256443352271194581Gastroenteritis862419607631183572267513637480273127026783201946Fractures207731891942316016472159655373293587773894Cellulitis98429408692877423782224652640283012443310Urinary tract infection13523089115729279313312506154858614127303324Viral illness28932396249322731428661226631228971292516843173Total30,379277825,64226141293049‡784149§2744517§10423799§22793308§&ast;Chronic diseases are not mutually exclusive. LOS indicates length of stay; ellipses, not applicable.†Arithmetic mean.‡Not significant for the comparison with no chronic disease.§P=0.0001 for the comparison with no chronic disease.The relationship observed between chronic disease and hospital charges was similar to that observed between chronic disease and LOS. The mean hospital charges for each procedure/condition group are displayed in Table 7. The presence of chronic disease increased hospital charges for all conditions studied. We report both arithmetic and geometric mean values, the former reflecting all patients equally, the latter allowing for a decreased influence from patient charge outliers. Use of the geometric mean, through reducing the influence of outliers, may better reflect a hospital's ability to pass on extreme charge outliers through reinsurance contracts. The geometric mean hospital charge for patients without chronic diseases was $2041 per admission compared with $2632 for patients with at least 1 chronic disease, a difference of $591 per patient (P<.001). The arithmetic mean hospital charge for patients without chronic diseases was $2614 per admission compared with $3663 for patients with at least 1 chronic disease, a difference of $1049 per patient (P<.001).Table 7. Comparison of Total Charge for Various Diagnosis Related Groups of Patients, According to Chronic Disease&ast;TotalProcedure/ConditionNo. of PatientsArithmetic Mean (SD)Geometric Mean (95% CI)Concussion11712004 (2021)1526 (1467, 1588)Croup20881698 (1699)1384 (1350, 1419)Pneumonia77773173 (3597)2497 (2463, 2532)Appendicitis34134770 (3402)4108 (4039, 4179)Gastroenteritis86241960 (2463)1564 (1544, 1584)Fractures20773189 (3168)2376 (2303, 2452)Cellulitis9842940 (2513)2347 (2256, 2442)Urinary tract infection13523098 (2064)2617 (2539, 2697)Viral illness28932396 (1978)1979 (1936, 2022)Total30,3792778(2992)2124(2108, 2140)&ast;Geometric mean and 95% confidence interval (95% CI) based on log scale values transformed back to total charge in dollars.†Student t test, compared with no chronic disease groups, using log-transformed data.PATIENTS WITH MULTIPLE CHRONIC DISEASESOf the 4737 patients with chronic disease, our study identified 1072 patients (22.6%) with 2 or more chronic diseases, occurring predominantly in the gastroenteritis (n=261) and pneumonia (n=536) groups. Overall, these patients had 31.2% longer LOSs than those with just 1 chronic disease and 58.5% higher charges. Admissions associated with more than 1 chronic disease accounted for 27.8% of hospital days for the chronically ill, and 31.7% of the total hospital expenditures on chronically ill children with common acute conditions. Patients with multiple chronic diseases experienced greater charges than did those with a single chronic disease. Patients with 1 chronic disease had a mean (arithmetic) charge of $3235 compared with $5126 in patients with 2 or more chronic diseases (P<.001).FINANCIAL ANALYSISChange in per-patient profit or loss at the children's hospital and general hospital was a function of the relative percentage of patients with chronic diseases treated at the children's and general hospitals. Table 8displays profit and loss for the children's hospital by the percentage of the population of patients with chronic disease treated at the children's hospital rather than the general hospital. If children's hospitals care for 75% of the patients with chronic illnesses and acute conditions, the loss per acute condition is 2.13% per case, based on the geometric mean and up to 2.86% per patient when using the arithmetic mean (fully counting outlier charges). Adjusting for the possible bias in cost-to-charge due to long LOS, the loss per patient is 1.5% per case (geometric) and 2.0% (arithmetic).Table 8. Children's Hospital per-Patient Profit (Loss) by Share of Chronically Ill and Outlier Assumptions65% Share of Chronically Ill Cases75% Share of Chronically Ill Cases85% Share of Chronically Ill CasesProcedure/ConditionGeometric Analysis&ast;Arithmetic Analysis†Geometric AnalysisArithmetic AnalysisGeometric AnalysisArithmetic AnalysisConcussion−0.44−0.73−0.74−1.20−1.04−1.68Croup−0.80−1.37−1.32−2.26−1.84−3.13Pneumonia−2.50−3.59−4.10−5.85−5.65−8.00Appendicitis−0.01−0.02−0.03−0.03−0.04−0.04Gastroenteritis−0.99−1.88−1.65−3.10−2.29−4.28Fractures−0.11−0.28−0.19−0.47−0.28−0.65Cellulitis−0.63−0.64−1.04−1.06−1.46−1.48Urinary tract infection−1.08−1.55−1.79−2.55−2.49−3.54Viral illness−1.13−1.52−1.86−2.50−2.59−3.47Total−1.29−1.73−2.13−2.86−2.96−3.96&ast;This method reduces the influence of patient outliers under an assumption that hospitals may obtain other funding for extreme outlier patients. All values are expressed as percentages of profit (loss).†This method assumes that the full cost of all patient outliers is paid by the hospital.Figure 1focuses on an especially costly group: patients admitted for pneumonia. Since it is assumed that the insurer pays the lower of the 2 average cost figures from either hospital, and that the cost structures of the 2 hospitals are a function only of their percentage of patients with and without chronic disease, there is no "excess" profit for either the children's or general hospital. Loss occurs in treating pneumonia when either hospital acquires a disproportionate share of patients with chronic illness and pneumonia, since treatment costs for pneumonia are higher in patients with a chronic illness than in those without a chronic illness. If children's hospitals care for 75% of the patients with chronic illness and acute pneumonia, the loss per case of acute pneumonia is 4.1% (geometric) and 5.85% (arithmetic). When we adjust for the possible bias in cost-to-charge ratios based on formulas relating to LOS, we get a loss of 4.0% per case (geometric) and 5.0% (arithmetic).The percentage of profit and loss per patient with pneumonia for the children's and general hospital as a function of case-mix fraction, assuming 29% of pneumonia cases have at least 1 chronic disease and an equal sample size across hospitals.COMMENTProviding hospital care to pediatric patients admitted for common procedures and conditions is more costly when a child has a chronic disease. In a recent study of pediatric Medicaid expenditures in Washington State, Ireys et alfound that patients with chronic illnesses (defined in a manner similar to our study) represented 5.9% of this population, and that patients with 1 or more chronic conditions had, on average, more than a 4-fold increase in total yearly Medicaid expenditures, of which approximately half of these expenditures were for inpatient services. Similar trends have been observed in adults with chronic disease.Our figure of 15.6% of pediatric hospital admissions being composed of patients with chronic illnesses is consistent with the Washington State experience. Not surprisingly, our study identified both meaningful and statistically significant increases in LOS and total hospitalization cost, as reflected through charges,when hospitalization for a routine pediatric condition or procedure is accompanied by a chronic medical condition. The size of the effect of chronic disease varied, depending on the specific underlying procedure or condition, but was found to be important for most of the conditions noted in the study.Newacheck and Taylor,using data from the 1988 National Health Interview Survey, reported that approximately 31% of children have some chronic disease. Only 5% of these patients had severe conditions; however, these 5% accounted for 33% of hospital expenditures for patients with chronic illnesses as a whole. Newacheck and Stoddardhave reported approximately 5% of patients had 2 or more chronic diseases, similar to the figure we found in our study. Our study shows that the 3.5% of admissions associated with more than 1 chronic disease accounted for 27.8% of hospital days for patients with chronic illnesses, and 31.7% of the total hospital expenditures on children with chronic illnesses and common acute conditions. One may conclude that patients with chronic illness, either in the hospitalized population (as in our study), or in the population as a whole (as in the Ireysor Newacheckstudies) represent a nonhomogeneous group of considerable size, with notable effect on total pediatric expenditures.Market changes are creating new difficulties in securing adequate funds for the care of chronic illness. Neff and Anderson,citing data from the National Association of Children's Hospitals and Related Institutions, point out that children's hospitals or hospitals that have postgraduate pediatric residency programs account for only 5% of hospitals, yet these hospitals admit 70% to 80% of the more complex chronic pediatric health conditions. A major distinguishing feature between pediatrics and adult medicine is the high concentration of pediatric specialists at children's hospitals, where for some specialties there are almost no practitioners at community hospitals.Therefore, the complex patients of specialists, often with chronic diseases, tend to be admitted to children's hospitals. To prevent market failure, Neff and Anderson have argued for"carve-out arrangements" and risk adjusters that take into account the higher than average costs of caring for children with chronic diseases. This will make possible adequate payments for health care plans that care for these patients and reduce the possibility that children with chronic illnesses will receive inappropriate care.(p1869)The model developed in this analysis is simple, yet, we believe useful in helping one understand the problems faced by children's hospitals or tertiary care centers specializing in pediatrics. By providing examples of per-patient loss, we believe the model helps us understand the magnitude of the problem outlined by Neff and Anderson from the perspective of the hospital. Faced with a loss per patient, the model suggests that hospitals may improve their economic situation by discouraging the admission of patients with chronic illnesses. This may be financially the right course to take, but would subvert the mission of children's hospitals. A preferred solution would be to ensure that the market price for acute conditions and procedures properly reflects the cost of providing such services, so that there is no disincentive to care for patients with chronic illnesses.Whenever contract prices do not reflect the full cost of care, market forces will cause some undesirable effects. The problem of unaccounted costs of chronic diseases occurs in adult medicine, but on a slightly smaller scale since chronic conditions are more common in adult medicine and have demanded more attention.Medicare's DRGs theoretically account for the major comorbidities of elderly patients. These reflect some of the increased resource utilization needs of patients with chronic disease; however, asymmetries in severity and case-mix continue to be a problem,and as competition increases throughout the hospital sector, such asymmetries in disease burden will cause problems similar to those described here, for all tertiary care hospitals.The market forces resulting from an asymmetric financial burden of disease may act to correct this disparity among hospitals in ways that are not necessarily in the best interest of society, and certainly not in the best interests of patients with chronic illnesses. Health economist Mark V. Pauly, PhD, has recently described an analogous problem concerning care for the uninsured. As he points out, " . . . hospitals with a heavy charity care . . . load may therefore find it impossible to provide care that they cannot avoid while still covering their costs. These firms will then cease to exist. The nonpaying patients who used to use these hospitals will go to other hospitals. . . ."(p359) Pauly goes on to predict that the uninsured will be distributed more evenly over the remaining hospitals, and that the care provided at these hospitals will be of lower quality. A similar argument could be made concerning the consequences of uncompensated care associated with treating patients with chronic illness.While children's hospitals need not exit the marketplace, one may expect they would either move to reduce the number of patients with chronic illness that they care for or lower their expenditures on these patients through lower quality of care. There are many ways to lower the quality of care provided to patients with chronic illness. Ireys et alhave outlined several major areas of concern regarding the adequacy of the current health service system for children with disabilities and chronic illnesses. Indicators of quality erosion may be found not only by studying the content of care, but also changes in the goals for care. Downward revisions of expectations may be an insidious method of cost reduction. Areas of concern can be manipulated by hospitals to save resources, yet in the long run, erosion of related services may lead to lower quality and worse outcomes.These observations should be significant for children's hospitals negotiating changes in their third-party contractual relationships. In the new capitated environment, children's hospitals, by necessity, require a far larger population of households to support their infrastructure than most adult hospitals, since the need for hospital services is far less frequent in childhood than in old age. Given the financial necessity at the children's hospitals for a sufficient volume of patients to maintain viable tertiary care and teaching programs, children's hospitals must enter competitive contracts with insurers, often offering the same price as the general hospitals for services performed at both hospitals.If market contracts for common procedures and conditions do not include chronic disease adjustments, children's hospitals will be placed at a financial disadvantage, precisely because the mission of most children's hospitals and pediatric tertiary care facilities is to provide care for the most complicated patients.Our analysis likely underestimated the problems facing children's hospitals. Charges in this analysis reflect resource utilization at general hospitals since most of the admissions in our study were to general hospitals. Thus, the analysis did not fully account for increases in service intensity at the children's hospital, that would increase the resource utilization, and hence, cause even greater loss for children's hospitals than those described in Table 8.Our analysis also did not address quality of care, and provided no premium for increased technological investment that may be required to treat specific pediatric populations.This analysis was limited to common pediatric problems. The observed effect of chronic disease on resource utilization may actually be stronger for conditions with lower incidence, which are even more likely to be treated at specialized children's hospitals. Children with rare conditions and chronic diseases are, with high probability, referred to, and remain in, the children's hospital systems. By concentrating on common procedures, we have likely underestimated the potential difficulties facing children's hospitals. Furthermore, by not defining all chronic diseases in our study population, we have overestimated the charges associated with the patients "without" chronic disease because this group reflects some patients who do have chronic disease. The difference between costs is likely greater than we estimated and hence children's hospital losses are also likely greater than we estimated.CONCLUSIONSChildren's hospitals should see a disproportionate share of patients with chronic disease. Our study has shown that for common acute conditions, children with chronic disease are more expensive to care for in the hospital than children without a chronic disease. Children's hospitals will not succeed in solving their financial problems by entering into high-volume contracts simply defined by the acute condition or procedure. Hospitals that care for a greater percentage of children with chronic conditions must seek contracts that better reflect the increased resources needed to treat such patients—even for common acute conditions and procedures. If not, market forces will act to change the mission of children's hospitals, either by reducing access of the chronically ill to children's hospitals or by lowering the quality of care provided at these hospitals.APPENDIXWe assume that there are 2 hospitals, a children's hospital (C) and a general hospital (G). The total number of pediatric patients seen at the children's hospital is a mixture of patients with chronic disease (CC) and those with no chronic disease (CN), while the total number of pediatric patients seen at the general hospital is a mixture of patients with chronic disease (GC) and those with no chronic disease (GN). We will refer to CC, CN, GC, and GNas fractions of 100% of the pediatric admissions in the town. Hence: CC+CN+GC+GN=1.0.Furthermore, we will assume that the total number of patients at the children's hospital is equal to the total number of patients at the general hospital, so that: CC+CN=GC+GN=0.5.The overall rate of chronic disease in the population (for a given disease or group of diseases) is defined as k, where k=CC+GCand the percentage of patients with chronic diseases who are treated at the children's hospital is defined as M, where: M=CC/(CC+GC)= CC/k.We can solve for the value of CC, CN, GC, and GNif values of k and M are specified. The above equations yield the following solutions: CC=kM; CN=0.5-kM; GC=k(1-M); and GN=0.5-k(1-M).We assume 1 insurer pays for all pediatric visits, and each hospital will contract for specific diagnoses. The final contract price offered by the insurer (π) will be the minimum average cost of treating all patients with a specified acute condition for either the general or children's hospital.Profit equations, where profit at the children's hospital is defined as PCand profit at the general hospital is defined as PGare as follows: PC=&lcub;[(π-πC)CC]+[(π-πN)CN]&rcub; and PG=&lcub;[(π-πC)GC]+[(π-πN)GN]&rcub;, where π is defined as the contract price, which is the minimum of either the price bid by the children's hospital or general hospital. The average cost at the children's hospital is: &lcub;(πCCC+πNCN)/(CC+CN)&rcub;.The average cost at the general hospital is: &lcub;(πCGC+πNGN)/(GC+GN)&rcub;.MGGoldfarbRMCoffeyChange in the Medicare case-mix index in the 1980s and the effect of the prospective payment system.Health Serv Res.1992;27:385-415.JGabelTen ways HMOs have changed during the 1990s.Health Aff (Millwood).1997;16:134-145.EGGayJJKronenfeldRegulation, retrenchment—the DRG experience: problems from changing reimbursement practice.Soc Sci Med.1990;31:1103-1118.RCCoile JrAssessing healthcare market trends and capital needs: 1996-2000.Health Care Financial Manage.1995;49:60-65.JHMaxwellHMSapolskyThe first DRG: lessons from the end stage renal disease program for the prospective payment system.Inquiry.1987;24:57-67.HFroehlichWRJarvisEconomic impact of diagnosis-related groups and severity of illness on reimbursement for central nervous system infections.J Pediatr.1991;118:693-697.LKLichtigRAKnaufRHParrottJMuldoonRefining DRGs: the example of children's diagnosis-related groups.Med Care.1989;27:491-506.AASkolnickChildren's hospitals say about health care reform, ‘one size won't fit all.'JAMA.1993;270:2151-2152.PDPhelanKBaxterJBishopIHudsonDHindleDo diagnosis related groups separate the case-mix of a specialist children's hospital and a paediatric unit in a general hospital?J Paediatr Child Health.1993;29:266-269.JCGayJHMuldoonJMNeffLJWingProfiling the health service needs of populations: description and uses of the NACHRI classification of congenital and chronic health conditions.Pediatr Ann.1997;26:655-663.JLFreemanRBFetterHParkDiagnosis-related group refinement with diagnosis- and procedure-specific comorbidities and complications.Med Care.1995;33:806-827.HTIreysGFAndersonTJShafferJMNeffExpenditures for care of children with chronic illnesses enrolled in the Washington State Medicaid program, fiscal year 1993.Pediatrics.1997;100:197-204.MJLongJLDreachslinJFisherShould children's hospitals have special consideration in reimbursement policy?Health Care Finance Rev.1986;8:55-63.JMNeffGAndersonProtecting children with chronic illness in a competitive marketplace.JAMA.1995;274:1866-1869.ACBrewsterBGKarlinLAHydeCMJacobsRCBradburyYMChaeMedisGroups: a clinically based approach to classifying hospital patients at admission.Inquiry.1985;22:377-387.LIIezzoniMAMoskowitzA clinical assessment of MedisGroups.JAMA.1988;260:3159-3163.JWThomasMLFAshcraftMeasuring severity of illness: a comparison of interrater reliability among severity methodologies.Inquiry.1989;26:483-492.JWThomasMLFAshcraftMeasuring severity of illness: six severity systems and their ability to explain cost variations.Inquiry.1991;28:39-55.LIIezzoniEKHotchkinASAshMShwartzYMackiernanMedisGroups data bases: the impact of data collection guidelines on predicting in-hospital mortality.Med Care.1993;31:277-283.PMSteenACBrewsterRCBradburyEEstabrookJAYoungPredicted probabilities of hospital death as a measure of admission severity of illness.Inquiry.1993;30:128-141.DRDurbinAPGiardioKNShawMCHarrisJHSilberThe effect of insurance status on likelihood of neonatal interhospital transfer.Pediatrics(electronic journal). 1997;100:381-382.American Academy of Pediatrics Committee on Infectious Diseases1994 Redbook.Elk Grove Village, Ill: American Academy of Pediatrics; 1994:280.JHSilberPRRosenbaumJSSchwartzRNRossSVWilliamsEvaluation of the complication rate as a measure of quality of care in coronary artery bypass surgery.JAMA.1995;274:317-323.MShwartzDWYoungRSiegristThe ratio of costs to charges: how good a bias for estimating costs?Inquiry.1995/96;32:476-481.JRLaveLBLaveHospital cost functions.Am Econ Rev.1970;60:379-395.RMUngerleiderARBengurALKessenichRisk factors for higher cost in congenital heart operations.Ann Thorac Surg.1997;64:44-48.JRGButlerHospital cost analysis.In: Zweifel P, Frech HE, eds. Developments in Health Economics and Public Policy. Dordrecht, the Netherlands: Kluwer Academic Publishers; 1995;1-393.SLKaufmanDSShepardCosts of neonatal intensive care by day of stay.Inquiry.1982;19:167-178.PFishmanMVon KorffPLozanoJHechtChronic care costs in managed care.Health Aff (Millwood).1997;16:239-247.PWNewacheckWRTaylorChildhood chronic illness: prevalence, severity, and impact.Am J Public Health.1992;82:364-371.PWNewacheckJJStoddardPrevalence and impact of multiple childhood chronic illnesses.J Pediatr.1994;124:40-48.SFJencksADobsonRefining case-mix adjustment: the research evidence.N Engl J Med.1987;317:679-686.SFJencksDKWilliamsTLKayAssessing hospital-associated deaths from discharge data: the role of length of stay and chronic disease.JAMA.1988;260:2240-2246.JGreenNWintfeldPSharkeyLJPassmanThe importance of severity of illness in assessing hospital mortality.JAMA.1990;263:241-246.JGreenLJPassmanNWintfeldAnalyzing hospital mortality: the consequences of diversity in patient mix.JAMA.1991;265:1849-1853.SAJoyRWYurtAn all-payor prospective payment system (PPS) based on diagnosis-related groups (DRG): financial impact on reimbursement for trauma care and approaches to minimizing loss.J Trauma.1990;30:866-873.JFederJHadleySZuckermanHow did Medicare's prospective payment system affect hospitals?N Engl J Med.1987;317:867-873.SOzatalayRBroylesNet returns, fiscal risks, and the optimal patient mix for a profit-maximizing hospital.J Med Syst.1987;11:331-347.MVPaulyTrading cost, quality and coverage of the uninsured: what will we demand and what will we supply?In: Altman S, Reinhardt U, Shields A, eds. The Future U.S. Health Care System: Who Will Care for the Poor and Uninsured?Chicago, Ill: Health Administration Press; 1997:353-373.HTIreysHAGrasonBGuyerAssuring quality of care for children with special needs in managed care organizations: roles for pediatricians.Pediatrics.1996;98:178-185.SAGliedSGnanasekaranHospital financing and neonatal intensive care.Health Serv Res.1996;31:593-607.CAHinzPediatric oncology patients: no future profit center.Health Care Strategic Manage.1991;9:1,19-22.Accepted for publication July 15, 1998.This research was funded in part by grant RO1-HS06560 from the Agency for Health Care Policy and Research, Rockville, Md, and the Ethel Brown Foerderer Fund for Excellence, Philadelphia, Pa (Dr Silber).Presented in part at the 15th Annual Association for Health Services Research National Meeting, Washington, DC, June 23, 1998.We thank William P. Greeley, MD, MBA, Mark V. Pauly, PhD, and Orit Even-Shoshan, MS for their helpful suggestions. We thank MediQual Systems Inc for their assistance in this research. Errors or omissions are solely our responsibility.Corresponding author: Jeffrey H. Silber, MD, PhD, The Children's Hospital of Philadelphia, Abramson Bldg, Room 702, 34th and Civic Center Blvd, Philadelphia, PA 19104-4318 (e-mail: Silberj@wharton.upenn.edu).

Journal

JAMA Pediatrics – American Medical Association

Published: Feb 1, 1999

There are no references for this article.