Clinical Chemistry
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Clinical Chemistry 48: 140-149, 2002;
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Tables
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Young, D. S.
Right arrow Articles by Jefferies, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Young, D. S.
Right arrow Articles by Jefferies, L. C.
Related Collections
Right arrow Laboratory Management
(Clinical Chemistry. 2002;48:140-149.)
© 2002 American Association for Clinical Chemistry, Inc.

Effect of Disease Complications on Hospital Costs

Donald S. Young1a, Bruce S. Sachais1 and Leigh C. Jefferies1

1 Department of Pathology and Laboratory Medicine, Hospital of the University of Pennsylvania, 3400 Spruce St., Philadelphia, PA 19104-4283.

aAuthor for correspondence. Fax 215-349-5090; e-mail donaldyo{at}mail.med.upenn.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: To test the hypothesis that complications increase the use of resources in managing patients in hospitals, we examined the costs of managing patients with the same disease with and without complications.

Methods: We used a database developed by the University HealthSystems Consortium that contains the costs of managing more than 1 million patients in 60 University hospitals. We created a simplified database of the costs of 457 445 patients in 111-paired diagnosis-related groups (DRGs) that were classified as either having or not having complications and/or comorbidities. Costs were calculated from the ratio of costs to charges within the individual hospitals.

Results: The median costs of managing patients with complications were higher than those for managing patients without complications, confirming the appropriateness of the dual classification. Notably, these extra costs were largely incurred through increased length of stay. Of note, the cost per day for DRGs with complications and/or comorbidities was most often less than that for the corresponding uncomplicated conditions. Although accommodation costs generally were the largest single component of total costs for both complicated and uncomplicated conditions, in only 31 DRGs (15 with complications, 16 without) did they account for more than one-half the total costs. Laboratory and drug costs were higher for complicated conditions, but as a proportion of total costs were comparable for complicated and uncomplicated conditions.

Conclusions: Complications in patients are associated with increased hospital costs, although the costs per day of hospitalization are often less than in patients without such complications.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
With the ongoing pressure to contain hospital costs, it is important to make every effort to avoid unnecessary complications because expenses increase with added problems. Certain diseases have been assigned by the Centers for Medicare and Medicaid Systems, formerly Healthcare Financing Administration, to different diagnosis-related groups (DRGs) 1 based on whether they are associated with complications. To identify the source of additional expenses that arise from complications that occur during patients’ admissions to hospital, we accessed a database developed by the University HealthSystems Consortium (UHC) that comprises the costs attributable to the individual non-physician sources of costs for all DRGs in 60 University hospitals in 1995. For the majority of diseases classified by DRG, a single number includes all degrees of severity of the disease, regardless of the presence or absence of complications or comorbidities. However, 222, or 45.6% of all DRGs, are separately classified as having or not having complications and/or comorbidities. It should be recognized, however, that the simple classification into DRGs with and without complications does not reflect the considerable variation in disease severity for hospital patients. Many of the more complicated medical conditions, such as organ transplantations, are not stratified into categories with and without complications and/or comorbidities. A complication in the context of this study does not differentiate between preventable and unavoidable complications.

Although we focused our efforts on diseases with and without complications, there are three additional conditions that have been classified as with and without "problems". Problems are essentially conditions that arise during the hospitalization of patients that cannot be ascribed to preexisting complications and/or comorbidities. The conditions with or without problems are vaginal delivery without problems (DRG 373) and with problems (DRG 372), prematurity without problems (DRG 388) and with major problems (DRG 387), and healthy newborns (DRG 391), neonates with major problems (DRG 389), and neonates with other significant problems (DRG 390). We separately studied these conditions.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
The database of 1 346 028 patients from 60 University hospitals used for this study has been described previously (1)(2). It is important to note here that patients who were extreme outliers, i.e., those who consumed the top 1% of resources, were excluded from the study. Patients who died during their hospitalization were also excluded from the database because they consume a disproportionate amount of resources (3). Our interest was in the typical situation rather than in the unusual patient. We focused our analysis on the median costs of the patients in the 111 pairs of DRGs with and without complications and/or comorbidities. A total of 286 906 patients were included in the 111 DRGs with complications and comorbidities and 170 539 patients in the conditions without complications. We elected to use median data, rather than mean, wherever possible, because the former are less affected by outliers than the latter. We evaluated median total costs and lengths of stay and costs attributable to accommodation, intensive care unit (ICU), surgery, radiology, laboratory services, and drugs as well as the relative proportion of the total costs for each of these components. The costs used in this study were derived in each of the participating hospitals from the ratio of costs to charges in 1995, calculated from the detailed Centers for Medicare and Medicaid Systems PPS UB-2 cost reports. These costs reflect hospital costs only and do not include those attributable to physicians’ services. We also contrasted the mean lengths of stay of patients in the UHC database with the Medicare data from all hospitals because the Medicare data are based on mean (arithmetic) lengths of stay.

All data were entered into a FileMaker (Ver. 4.0; Claris Corporation) file. The data were also exported to an Excel (Ver. 5.01; Microsoft Corporation) file, and these programs were used to derive relationships among components and perform the calculations. Statistical calculations were performed using GraphPad Prism (Ver. 2.0; GraphPad Software Inc.)


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
number of patients
There were 111 pairs of DRGs stated to be with and without complications or comorbidities. Forty-eight of these pairs were classified as surgical conditions, and 63 were medical conditions. Those conditions with the highest number of patients with complications are shown in Table 1 and are contrasted with the number of patients in the corresponding DRG without complications. When the paired conditions were ranked by numbers of those with complications or comorbidities, the highest number of discharges was for cesarean sections: 11 566 patients with and 12 376 patients without complications and/or comorbidities. Typically, for each of the pairs, more patients had the condition with complications or comorbidities than without. Of the paired conditions, there were more patients with complications in 81 of the conditions and fewer in 30. Of the 21 paired conditions with >4000 patients categorized as having complications and/or comorbidities, 17 had more patients with complications, whereas only 4 had more patients without complications.


View this table:
[in this window]
[in a new window]
  		Table 1. Largest numbers of patients in DRGs with or without complications and/or comorbidities for common DRGs.1

Respiratory infections and inflammations in adults was the DRG that showed the largest ratio between the number of patients with complications and those without. Strikingly, there were more than 10 times more patients with complications and/or comorbidities than without. Other examples of large ratios are atherosclerosis and digestive malignancies, for which there were 8.5 and 7.5 times, respectively, more patients with complications and/or comorbidities than without (data not shown). In contrast, there were 4.0 times more patients with uncomplicated knee conditions than with complications and 3.4 times as many patients with appendectomy without a complicated principal diagnosis without complications than those with complications.

The smallest numbers of patients in the paired disease categories were 46 with urethral stricture with complications in adults (DRG 328) and 28 patients with the same condition without complications (DRG 329). There were 95 patients who had benign prostatic hypertrophy with complications (DRG 348) and 54 patients with the condition without complications (DRG 349).

overall costs
For all paired conditions, the total cost of caring for patients with complications and/or comorbidities exceeded the cost of the condition without. Generally, the costs for diseases that are categorized with or without complications and/or comorbidities are relatively lower than those for diseases that are not separately categorized. Thus, the median costs of 41 conditions of the 264 DRGs (15.5%) not subclassified exceeded $10 000, whereas only 15 (6.8%) classified as having or not having complications exceeded $10 000 (Table 2 ). Of all the DRGs included in this study, the median costs of both the condition with and the condition without complications exceeded $10 000 in only 1 of the 15 pairs (major cardiovascular procedure), as shown in Table 2 . The costs of 14 conditions not subclassified with or without complications exceeded $20 000 (data not shown), but only the costs of malignant breast disorders with complications or comorbidities exceeded $20 000.


View this table:
[in this window]
[in a new window]
  		Table 2. Conditions with complications and/or comorbidities for which median total costs exceeded $10 000.1

The difference in costs between conditions with complications and/or comorbidities and those without was most marked for malignant breast disorders, for which the median cost of the disease with complications was $20 247, more than five times the cost without complications. The DRG with the next largest cost difference ($8745) attributable to complications was for pancreas, liver, and shunt procedures. The median costs of the top four conditions listed in Table 2 Up exceeded 2 SD of the mean for the 111 DRGs with complications and/or comorbidities. For the uncomplicated conditions, those six with costs of $8292 or more exceeded 2 SD of the mean of the median costs.

The mean ratio of the median total costs of conditions with complications and/or comorbidities to those without was 1.56 (data not shown). The differences ranged from lows of 1.03 for major shoulder or elbow procedures (DRGs 223 and 224), 1.04 for urethral strictures in patients older than 17 years (DRGs 328 and 329), and 1.08 for subtotal mastectomy for malignant disease of the breast (DRGs 259 and 260) to a high of 5.54 for malignant breast disorders (DRGs 274 and 275). The second highest ratio of costs for the disease with complications compared with that without was much less, at 2.97 for other endocrine, nutritional, or metabolic conditions with an operating room procedure. For these pairs of conditions, the actual difference in median costs was $7989, substantially less than the difference for breast malignancies.

The smallest differences in median costs for diseases with and without complications were $98 for urethral strictures in patients over age 17 years, $105 for major shoulder or elbow procedure, $229 for other antepartum diagnoses, $303 for subtotal mastectomy for malignant disease, $341 for kidney and urinary tract infections, and $344 for atherosclerosis. All of these conditions were associated with quite low total costs. The highest cost for any of this group of diseases was $4031 for major shoulder or elbow procedure with complications and/or comorbidities.

ratio of mean to median costs
The ratio of mean to median costs provides a simple indication of the heterogeneity of the conditions included in a single DRG. For all the conditions listed as either having or not having complications and/or comorbidities, the mean total costs exceeded the median. There were six conditions with complications and six without whose ratios of mean to median costs exceeded 1.50 [Table 2A; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. Generally, the ratio of mean to median total costs was smaller for conditions without complications than for those with complications and/or comorbidities. The ratio was higher in 89 of the conditions with complications and/or comorbidities than in those without and lower in 22. The higher ratios for diseases with complications and/or comorbidities are consistent with the expected greater variability in these conditions. The ratios of mean to median costs for uncomplicated malignant breast disorders and atherosclerosis (2.522 and 2.453, respectively) were the largest of all the ratios [Table 2B; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. The largest ratio among conditions with complications was 1.908 for malignancy of the male reproductive system, which substantially exceeded the ratio of 1.140 for the condition without complications.

length of stay
Typically, the median length of stay for the complicated condition was 1.75 times longer than, or almost 2 days longer than, the median length of stay for the comparable uncomplicated condition. The greatest differences in median lengths of stay between the complicated and uncomplicated conditions were 13 days for malignant breast disorders, 7 days for other endocrine nutritional or metabolic conditions with an operative procedure, and 6 days for stomach, esophageal, and duodenal surgical procedures. The greatest differences in mean lengths of stay between complicated and uncomplicated conditions were 8.50 days for lymphoma or leukemia with an operative procedure; 7.10 days for other endocrine, nutritional, metabolic disease with an operative procedure; 6.90 days for myelodysplasia with a major operative procedure; 6.60 days for skin grafts for ulcers; 6.40 days for other skin and breast procedures; and 6.32 days for stomach, esophageal, or duodenal operative procedures. The average of the ratios of the mean lengths of stay of all complicated conditions compared with uncomplicated conditions was 1.58.

The ratios of mean to median length of stay were as high as 2.38 for malignant male reproductive system disorders without complications, 2.20 for kidney and urinary tract symptoms without complications, 2.09 for skin trauma with complications, 2.07 for concussion with complications and knee procedures without complications, 2.01 for laparoscopic cholecystectomy without complications, and 2.00 for traumatic injuries without complications (data not shown). In all of these situations with the largest ratios, the median length of stay was only 1 day. The UHC mean length of stay correlated positively with the Medicare arithmetic mean length of stay for diseases with complications or comorbidities (r2 = 0.8689) or for diseases without complications or comorbidities (r2 = 0.8631).

For all except eight conditions, the Medicare arithmetic mean length of stay exceeded the mean length of stay in the UHC hospitals [Table 2C; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. The largest differences were 5.71 days for malignant breast disorders with complications and 3.56 days for malignant breast disorders without complications. The difference of 1.09 days for complications of drug treatment without complications (DRG 453) was the only other condition for which the difference exceeded 1.00 days. The largest ratios of mean Medicare length of stay to those in the University hospitals were for other injuries, poisonings, and toxic conditions with complications, for which the ratio was 5.76, and complicated lymphoma or leukemia with an operative procedure, for which the ratio was 2.82 (data not shown). For both complicated and uncomplicated skin trauma, the ratios were 2.68 and 2.60, respectively, and for complicated and uncomplicated other injuries, poisonings, and toxic reactions, the ratios were 5.75 and 2.48, respectively.

The ratio of the Medicare mean length of stay to the UHC length of stay for biliary tract procedures was 0.99; for uncomplicated skin disease, the ratio was 0.87; for uncomplicated cesarean sections, the ratio was 0.99; for uncomplicated other antepartum diagnoses, it was 0.78; for uncomplicated nonacute lymphoma or leukemia, it was 0.97; and for uncomplicated complications of drug treatment, the ratio was 0.76. For malignant breast disorders with complications and nonacute lymphoma or leukemia with complications, the Medicare mean lengths of stay were 0.59 and 0.97, respectively, of the UHC lengths of stay (Table 2C Up in the on-line data supplement).

The average ratio of Medicare arithmetic mean lengths of stay of complicated conditions to uncomplicated conditions was 1.75. The greatest difference in ratio was 3.20 for nervous system procedures with and without complications, and the smallest ratio was 1.25 for both major shoulder or elbow procedures and for alcohol or drug dependence with detoxification.

accommodation costs
Both for diseases with complications and/or comorbidities and diseases without, there was a strong correlation between total median costs and median accommodation costs, r2 = 0.9447 and 0.8035, respectively. For 110 of the 111 paired conditions, accommodation costs were higher for the complicated condition. For the other paired condition, alcohol or drug dependence with detoxification, the accommodation costs were the same. The greatest difference in accommodation costs between conditions with complications and those without was $9901 for malignant breast procedures. For conditions with complications and/or comorbidities, accommodation costs as a percentage of total costs varied from 18.8% for hand procedures to 61.2% for alcohol or drug dependence with detoxification. For uncomplicated conditions, the proportion of total costs attributed to accommodation ranged from 15.2% for laparoscopic cholecystectomy to 76.4% for alcohol or drug dependence with detoxification. For 31 of the DRGs, 15 with complications and/or comorbidities and 16 without, the accommodation costs accounted for >50.0% of the total costs (Table 3 ). All of these conditions were medical conditions. The majority of conditions with accommodation costs <25.0% of total costs were surgical conditions. The four conditions with complications and/or comorbidities with the highest accommodation costs, all exceeding 2 SD of the mean, also had the highest total costs [Table 3A; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. For the uncomplicated conditions, the five with the highest accommodation costs were ranked among the six with the highest total costs.


View this table:
[in this window]
[in a new window]
  		Table 3. Conditions for which median accommodation costs were >50.0% of total costs.1

The majority of conditions with accommodation costs <25.0% of total costs were surgical conditions. Five conditions without complications had accommodation costs that were <20.0% of total costs, and one with complications, hand procedures, had accommodation costs <20.0% of the total costs [Table 3B; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)].

icu costs
Only four conditions with complications or comorbidities had identifiable median ICU costs. The only condition for which there were identifiable median ICU costs for both the complicated and uncomplicated condition was major cardiac reconstructive procedure, for which ICU costs were $2377 for the complicated condition and $1058 for the uncomplicated one. ICU costs comprised 36.9% of the accommodation costs (13.2% of total costs) for the complicated condition and 30.5% of the accommodation costs (10.0% of total costs) for the uncomplicated condition. Circulatory disease with complications had ICU costs of $1066, poisonings and toxic effects of drugs in patients over age 17 years had ICU costs of $1042, and pancreas and liver shunt procedures with complications had ICU costs of $542. None of the corresponding uncomplicated conditions had any identifiable ICU costs. For both circulatory disorders with acute myocardial infarction and major cardiovascular procedures, the ICU costs comprised 13.2% of total costs. The ICU costs of pancreas and liver shunt procedures with complications were 3.3% of the total costs.

surgical costs
When the surgical costs of DRGs 346 and 348, malignancy of the male reproductive tract and benign prostatic hypertrophy, respectively, were excluded because they were disproportionately low ($156 and $124, respectively), median surgical costs for the conditions with complications ranged from $4669 (major cardiovascular procedures) to $1166 (operative procedures on the respiratory system), and those without complications ranged from $3897 (biliary tract procedures) to $1356 [anal and stomal procedures; Table 3C; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. For all but seven of the paired DRGs, the surgical costs of the diseases with complications exceeded those for the disease without complications. Of the 43 conditions for which the surgical costs of the condition with complications exceeded that of the conditions without, the difference was <10% of the costs for complicated conditions in 18 and <20% in 36. For the two DRGs with surgical costs less than $1000, the number of patients both with and without complications was small. There were only 108 patients in DRG 293 (other endocrine, nutritional, and metabolic disorders with an operative procedure) and 406 patients with the complicated condition.

When DRGs 346 and 348 were excluded, the proportion of total costs of conditions with complications and/or comorbidities that were attributable to surgery ranged from 13.4% for operative procedure on the respiratory system (DRG 076) to 62.5% for hand procedures (DRG 228). In four conditions with complications, more than one-half of the total median costs were attributable to surgery. When we also excluded DRGs 347 and 349 (malignancy of the male reproductive system without complications and benign prostatic hypertrophy without complications, respectively), the proportion of total costs attributable to surgery for diseases without complications ranged from 29.5% for skin graft for ulcers (DRG 263) to 67.2% for laparoscopic cholecystectomy (DRG 493). For 21 uncomplicated conditions, surgical costs comprised more than one-half the total costs.

When the proportion of total costs attributable to surgery in conditions with and without complications or comorbidities were contrasted, the data showed that the presence of complications and/or comorbidities caused other costs to increase disproportionately in comparison with surgical costs. For conditions with complications, the largest proportion of total costs attributable to surgery was 48.4% for laparoscopic cholecystectomy compared with 67.2% for uncomplicated laparoscopic cholecystectomy (Table 3C in the on-line data supplement). The largest differences in proportion of total costs attributable to surgery between the uncomplicated and complicated condition were 27.6% for stomach, esophageal, and duodenal procedures and 26.2% for nervous system procedures.

radiology costs
For 30 diseases without complications, no median radiology costs were identified, whereas only 7 diseases with complications had no identifiable median radiology costs. Where both conditions with and without complications incurred radiology costs, the mean ratio of the median costs for diseases with complications to those without was 2.51. The largest cost difference was $819 for lymphoma or leukemia when compared with the condition without complications. The radiology costs for the complicated condition ($887) were 13.0 times higher than those for the uncomplicated condition ($68). For three pairs of conditions, atherosclerosis, major shoulder or elbow procedure, and fever of unknown origin, the radiology costs for the uncomplicated condition exceeded those of the complicated condition by $5, $68, and $64, respectively. However, for none of these conditions did radiology costs exceed $200.

Radiology costs were typically a small part of the total costs, but for 10 paired conditions they comprised >10.0% of the total costs [Table 3D; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. Radiology costs comprised 35.6% of total costs for concussion with complications and 19.4% of the total costs for the condition without complications. For trauma or stupor with coma, the radiology costs for both the complicated and uncomplicated conditions exceeded 19% of the total costs: 19.2% and 19.8%, respectively. For other injuries, poisonings, and toxicology, radiology costs were 16.5% of total costs for the condition with complications and 16.7% of the total costs for the condition without complications. For bone fractures of the arm, hand, or foot with complications and/or comorbidities, radiology costs were 18.8% of the total costs for the complicated condition but only 12.8% for the uncomplicated condition.

laboratory costs
Median laboratory costs varied from a low of $6 for uncomplicated knee procedures ($203 for the complicated condition) to $1620 for pancreas and liver shunt procedures with complications ($783 in the absence of complications; data not shown). For all 111 paired conditions, the laboratory costs for the condition with complications were higher than those for the condition without complications, although the differences of $6 for traumatic injury ($255 with complications and $249 without) and $24 for subtotal mastectomy for malignancy ($337 with complications $313 and without) were very small. The three complicated conditions with the highest laboratory costs—malignant breast procedures ($2221), pancreas and liver shunt procedures ($1620), and major cardiovascular procedures ($1400)—were among the four conditions with the highest overall costs (Table 3A in the on-line data supplement). The three uncomplicated conditions with the highest laboratory costs—pancreas and liver shunt procedures ($783), biliary tract procedures ($676), and major cardiovascular procedures ($577)—were also the conditions with highest total costs, although not ranked in the same order.

On average, median laboratory costs for diseases with complications were 2.46 times higher than the costs for those without complications. The laboratory costs for knee procedures with complications ($203) were 33.8 times higher than the $6 for the uncomplicated condition. The largest absolute difference in laboratory costs for conditions with and without complications was $1806 for malignant breast disorders (data not shown). The next largest difference ($1036 for stomach, esophageal, and duodenal procedures) was 7.1 times higher than the laboratory costs of the condition without complications ($145). Although median laboratory costs for diseases with complications or comorbidities were well correlated with total median costs (r2 = 0.8494), the correlation of the median laboratory costs of the same diseases without complications was considerably weaker (r2 = 0.5567).

The largest proportions of total disease costs attributable to laboratory costs were 17.9% for liver disease attributable to malignancy or cirrhosis with complications (DRG 205) and 17.6% for the condition without complications (DRG 206). The proportion of total costs attributable to the laboratory was only slightly less for fever of unknown origin: 16.6% with complications and 16.2% without. The conditions for which laboratory costs comprised >10.0% of total costs are listed in Table 4 . For 98 of the conditions, laboratory costs comprised a larger proportion of the total costs of the condition with complications than for that without. By far the largest ratio was 23.07, for knee procedures with complications compared with the uncomplicated situation. When this very different ratio was excluded, the median ratio was 1.33 and the mean ratio was 1.40 (SD, 0.45). The minimum ratio (0.23) was for minor bladder procedures, and the largest ratio, after 23.07 for knee procedures, was 3.25 for stomach, esophageal, and duodenal procedures. Eleven of the conditions for which the proportion of total costs attributable to the laboratory were less for the complicated condition than for the uncomplicated condition were medical. The two surgical conditions were minor bladder procedures and subtotal mastectomy for malignancy.


View this table:
[in this window]
[in a new window]
  		Table 4. Conditions for which median laboratory costs were >10.0% of median total costs.1

drug costs
The highest median drug costs were $6935 for malignant breast disorders with complications. For the uncomplicated condition, median drug costs were only $523 [Table 4A; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. This 13.3-fold difference was the greatest for all of the paired conditions. The second highest drug costs, $1540 for pancreas, liver, and shunt procedures with complications, were only $849 more for the same condition without complications. The lowest drug costs were $24 for uncomplicated syncope. The drug costs for syncope with complications were also low, only $59. The median drug costs for complicated conditions were, on average, 2.20 times higher than the costs for uncomplicated conditions. For all conditions except transurethral prostatectomy, for which the drug costs for the condition without complications ($260) exceeded those with ($212), the drug costs of the complicated condition exceeded those of the uncomplicated condition.

Median drug costs comprised 10.0% or more of the median total costs for either or both of 16 conditions with or without complications [Table 4B; available as a data supplement through the on-line version of this article at Clinical Chemistry Online (http://www.clinchem.org/content/vol48/issue1/)]. The median drug costs, as a proportion of total median costs, were higher for 14 conditions with complications than for those without. For malignant breast disorders with complications, drug costs comprised 27.8% of total costs, >11.1% higher than the costs for the condition without complications. Drug costs were 18.3% of total costs for kidney or urinary tract infections with complications and 10.5% without complications. There was a strong correlation between median drug costs and median total costs for both conditions with complications (r2 = 0.8126) and without (r2 = 0.8908).

blood costs
Of the 111 paired conditions, 10 of those with complications had identifiable median blood costs, but no condition without complications had such costs. One hundred seven pairs of conditions had identifiable mean blood costs, but in four of the paired conditions only the complicated condition had identifiable mean blood costs. For all except one of the pairs, the blood costs for the condition with complications were higher than the costs for the condition without complications. On average, the mean blood costs for patients with conditions with complications and/or comorbidities were 7.4 times higher than those for the uncomplicated conditions. There was a remarkable difference between the mean and median blood costs for patients with malignant breast disorders. The mean costs for the conditions with and without complications were $1798 and $790, respectively, whereas there were no apparent median costs for either condition (data not shown).

per diem costs
The difference in the median length of stay for all conditions with complications and/or comorbidities and those without is shown in Table 5 . Table 5 also shows that, although daily accommodation costs did not change materially with the presence of complications, the daily costs of radiology, laboratory, pharmacy, and other services did increase. Nevertheless, the median overall cost per day for patients with complications was less than that for uncomplicated conditions. Thus, the increased costs incurred by patients with complicated conditions were largely attributable to the increased length of stay of these patients.


View this table:
[in this window]
[in a new window]
  		Table 5. Comparison of median lengths of stay and costs for conditions with complications and/or comorbidities vs those without.

conditions with and without problems
Although DRG 387 (prematurity with major problems) is not identified as a disease with complications or comorbidities, the median cost ($11 214) for the condition was $10 030 higher than prematurity without major problems, for which the median cost was $1184. The median cost for treating a full-term neonate with major problems ($2408) was substantially higher, and that of a neonate with other significant problems ($699) was also higher than the $486 cost for a healthy newborn. The costs of the two conditions with problems were thus 9.47 and 1.14 times higher than the costs for a healthy newborn. The median costs of the conditions with and without problems are shown in Table 6 .


View this table:
[in this window]
[in a new window]
  		Table 6. Median lengths of stay and costs of conditions with and without problems.

The overall costs for vaginal deliveries with and without problems were quite low, $2569 and $2018, respectively. Thus, differences in costs of component services were small. Median accommodation costs for the condition with problems were 1.4 times higher than that without, although the median length of stay was the same. Drug costs were almost twice as high, $142 vs $73, and laboratory costs were 1.70 times higher, although the median laboratory costs for deliveries with problems were only $167.

In contrast, there were large differences in costs, complexity, and length of stay between premature babies with and without problems. The median length of stay of the infants with problems was more than four times longer than that of premature infants without problems. The median total costs for premature infants with problems were almost 10 times higher than those for premature infants without problems. The 9.6-fold increase in drug costs was comparable. The laboratory costs for prematurity without problems were only $99. Median laboratory costs were more than fivefold higher for infants with problems than for those without and comprised 4.6% of total costs for infants with problems but only 2.2% for premature infants without problems. The largest ratio of drug costs for a condition with complications to that without was 9.02 for unspecified signs and symptoms. For prematurity with problems, the ratio was 9.61 times the cost of prematurity without problems.

There was considerable difference between the mean (9.85 days) and median (3 days) length of stay in neonates with major problems. The median total, accommodation, laboratory, drug, and surgical costs for neonates with major problems were four to six times higher than for healthy newborns. For neonates with other significant problems, the corresponding costs were only 1.4–2.2 times higher than those for healthy newborns.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Overall there were more patients in DRGs with complications and/or comorbidities than in the corresponding DRGs without complications. This may reflect the tertiary care mission of most university hospitals, which might induce physicians to refer more complicated patients to university, rather than community, hospitals. In some cases the difference in numbers was striking, as for respiratory procedures, simple pneumonia, and pleural effusions. For all of these conditions, there were at least five times more patients with complications than without. There were three diseases of the gastrointestinal tract for which there were five times more patients with complications than without: digestive tract malignancy; gastrointestinal hemorrhage; and pancreas, liver, and shunt procedures. The only other conditions with a fivefold difference were fever of unknown origin (1646 patients with complications and/or comorbidities and 319 without) and other circulatory disorders (8041 patients with complications and/or comorbidities and 1299 without).

The diseases with the smallest number of patients, e.g., urethral stricture, benign prostate hypertrophy, and diseases of the male reproductive tract, are now most commonly treated on an outpatient basis, which accounts for the low numbers of patients with these quite common conditions included in the database. The database does not provide the information to determine whether the admissions were warranted.

The diseases that have been subclassified by the presence or absence of complications and/or comorbidities are typically of low cost. The majority of the diseases classified by DRG do not have such separation. Factors prompting classification with complications and/or comorbidities include complications of surgery and infections arising during the admission. Many studies have demonstrated considerable differences in use of resources by patients classified with the same DRG. For example, Horn et al. (4) showed that variability in the severity of the underlying condition can account for 61% of the variability of resource consumption. Regardless of the variability of severity of the condition, however, a hospital is reimbursed the same amount for patients with the same DRG.

As the data show, the costs of managing patients with complications and/or comorbidities are higher than those of managing patients with the corresponding uncomplicated conditions. Across all 111 paired DRGs with and without complications and/or comorbidities, the mean length of stay for the complicated condition was 1.75 times longer than that of the uncomplicated condition. Increased costs were associated with increased length of stay. For malignant breast disorders, the costs of the complicated condition were more than five times higher than those of the uncomplicated condition. For other endocrine, nutrition, and metabolic conditions with an operative procedure, the costs of the complicated condition were almost three times higher than those of the uncomplicated condition. Surprisingly, the mean of the median cost per day for all DRGs with complications was $150 less than that for the uncomplicated conditions. There was little difference in daily accommodation costs between the two populations, but the mean daily laboratory costs and drug costs were, respectively, $17 and $24 higher for complicated conditions than for uncomplicated conditions, reflecting the added costs to manage and monitor the more severely ill patients.

In spite of the higher costs typically associated with teaching hospitals and the presumed more complicated patients admitted to such hospitals (5), our data show that the mean length of stay in the 60 University hospitals in this study was frequently less than that in all hospitals, as indicated by the national Medicare data. The shorter length of stay in teaching hospitals may be attributable to a more intense focus on costs or, conceivably, better management of patients than in community hospitals. It should also be noted that the cited Medicare lengths of stay refer exclusively to patients whose costs are paid by Medicare and Medicaid. These patients are typically 65 years or older, whereas the data from the 60 University hospitals include a larger population of younger patients as well, who do not necessarily require as much care as older patients.

For surgical conditions, the costs of surgery were comparable for the same condition with or without complications and comorbidities. However, the proportion of total costs attributable to surgery was substantially higher for the conditions without complications because of the added costs incurred through increased length of stay and the treatment of the complications in the patients with complications and/or comorbidities. Accommodation costs as a proportion of total costs for conditions without complications ranged from a low of 15.2% for laparoscopic cholecystectomy to 76.4% for alcohol or drug dependency with detoxification. For conditions with complications, the proportion ranged from a low of 18.8% for hand procedures to 61.2% for alcohol or drug dependency with detoxification. The conditions with the largest proportion of costs attributable to room and board are largely medical and, because of the added costs of the complications, are usually higher for the DRGs of the uncomplicated conditions.

Radiology costs as a proportion of total costs were more than one-third of the total costs for patients with concussion and complications, but were <20% of the total costs in the absence of complications. Nevertheless, for 11 conditions with complications and 12 without, radiology costs exceeded 10.0% of the total costs. For many more conditions, the laboratory costs as a proportion of total costs exceeded the proportion incurred by radiology. For 38 conditions with complications and 19 conditions without, laboratory costs accounted for >10.0% of the total costs. Generally, the proportion of total costs attributable to laboratory testing were comparable for conditions with and without complications. Drug costs were markedly higher for conditions with complications and/or comorbidities than for those without. As a proportion of total costs, drug costs were also higher for conditions with complications.

Münoz et al. (6), using a simple approach of adding multiple diagnoses as an approximation for complications and comorbidities, demonstrated for patients with medical DRGs that accommodation, laboratory, radiology, and pharmaceutical costs increased with increasing number of complications and comorbidities. The mean length of stay also increased in the same way. Freeman et al. (7) identified those complications and comorbidities that are strong predictors of hospital resource use. These include hypertension, bronchitis, septicemia, and cellulitis for both medical and surgical DRGs.

This study clearly identifies the impact of complications and comorbid conditions on hospital resource use. We have not attempted to assess the effect of specific complications, such as infections, that arise during hospitalizations, but the experience of most hospitals is that these complications add substantially to a patient’s length of stay and use of resources such as radiology, the laboratory, and the pharmacy. Further studies are needed to quantify the real impact of individual hospital-derived complications. The results reported here also were obtained at a time when relatively few hospitals had developed critical (clinical) pathways for many of the conditions with large numbers of admissions. It would be interesting to study impact of critical pathways on overall costs.

This study confirms the appropriateness of having separate DRGs for those conditions that are associated with complications and/or comorbidities because of the higher costs incurred through managing these patients. Because this study was concerned only with the existing separate DRGs for conditions with and without complications, it is not possible to determine whether such a subclassification would be appropriate for other conditions. Our earlier study (2) showed that the median and mean costs for many diseases can be quite different, implying considerable variability in the consumption of resources, which in turn is probably attributable to differences in the severity of the patients’ underlying conditions. Much of the difference in costs between patients with and without complications is attributable to length of stay. Interestingly, the costs per day tend to be higher for patients with uncomplicated conditions. This is probably related to certain fixed costs, such as surgery, that must be apportioned over a smaller number of days.


   Acknowledgments
 
This work was supported in part by Grant PO1-HL-54516-02 from the National Heart, Lung and Blood Institute, National Institutes of Health. We are grateful to Kelley Griffin for initial processing and manipulation of the data for this study.


   Footnotes
 
1 Nonstandard abbreviations: DRG, diagnosis-related group; UHC, University HealthSystems Consortium; and ICU, intensive care unit.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Bankowitz RA, Neikirk MA, Juris A. UHC clinical data base analyst reference manual March 2000 University HealthSystems Consortium Oak Brook, IL. .
  2. Young DS, Sachais BS, Jefferies LC. The costs of disease. Clin Chem 2000;46:955-966.[Abstract/Free Full Text]
  3. Lubitz JD, Riley GF. Trends in Medicare payments in the last year of life. N Engl J Med 1993;328:1092-1096.[Abstract/Free Full Text]
  4. Horn SD, Horn RA, Sharkey PD, Chambers AF. Severity of illness within DRGs: homogeneity study. Med Care 1986;24:225-235.[Web of Science][Medline] [Order article via Infotrieve]
  5. Iezzoni LI, Shwartz M, Moskowitz MA, Ash AS, Sawitz E, Burnside S. Illness severity and costs of admissions at teaching and nonteaching hospitals. JAMA 1990;264:1426-1431.[Abstract/Free Full Text]
  6. Münoz E, Rosner F, Friedman R, Sterman H, Goldstein J, Wise L. Financial risk, hospital cost, and complications and co-morbidities in medical non-complications and co-morbidity-stratified diagnosis-related groups. Am J Med 1988;84:933-939.[Web of Science][Medline] [Order article via Infotrieve]
  7. Freeman JL, Fetter RB, Park H, Schneider KC, Lichtenstein JL, Hughes JS. Diagnosis-related group refinement with diagnosis- and procedure-specific co-morbidities and complications. Med Care 1995;33:806-827.[Web of Science][Medline] [Order article via Infotrieve]



The following articles in journals at HighWire Press have cited this article:


Home page
 Clin. Chem.Home page
D. S. Young, B. S. Sachais, and L. C. Jefferies
Comparative Costs of Treating Adults and Children within Selected Diagnosis-related Groups
Clin. Chem., January 1, 2002; 48(1): 150 - 160.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Tables
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Young, D. S.
Right arrow Articles by Jefferies, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Young, D. S.
Right arrow Articles by Jefferies, L. C.
Related Collections
Right arrow Laboratory Management

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS