| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Medicine, Division of Critical Care, Stamford Hospital, Stamford, CT.
Correspondence to: James Stephen Krinsley, MD, Director of Critical Care, Stamford Hospital, 190 West Broad St, Stamford, CT 06902; e-mail: jkrinsley{at}stamhealth.org
Abstract
Study objectives: To assess the effect of an intensive glycemia management protocol on the cost of care of a heterogeneous population of critically ill adult patients.
Design: Economic analysis of a 1,600-patient "before-and-after" study of intensive glycemia management.
Setting: Fourteen-bed mixed medical-surgical adult ICU of a university-affiliated community teaching hospital.
Patients: Eight hundred consecutive admissions to the ICU prior to the institution of an intensive glucose management protocol were compared to the first 800 patients admitted to the ICU following institution of the protocol.
Interventions: Cost data were analyzed using the comprehensive database of the ICU as well as other hospital data repositories.
Measurements and results: The ICU database was used to quantify the major components of the cost of care. The analysis includes costs associated with ICU and non-ICU patient days, ventilator days, and laboratory, pharmacy, and radiology services. Comparing the baseline and treatment periods, there were decreases in patient days in the ICU; ventilator days; total laboratory, pharmacy and radiology costs; and post-ICU hospital length of stay. The net annualized decrease in costs during the treatment period was $1,339,500, or $1,580 per patient.
Conclusions: The institution of a program to intensively monitor glucose levels and treat even modest hyperglycemia in the ICU was associated with substantial cost savings. This finding, in conjunction with the previously demonstrated improvement in mortality and morbidity, strongly supports the adoption of this intervention as a standard of care in the ICU.
Key Words: cost • glycemia management • hyperglycemia • ICU • length of hospital stay
An emerging medical literature has described the adverse consequences of hyperglycemia in a variety of different clinical contexts. Hyperglycemia occurs commonly among patients with acute neurologic disease,12345 ischemic heart disease,678910 and a range of surgical and trauma diagnoses111213; its mechanism has been attributed to numerous stress, counterregulatory, and iatrogenic factors.14 A review15 of 1,826 patients admitted to an adult medical-surgical ICUs demonstrated a strong association between increasing glucose levels during ICU admission and the risk of hospital mortality. The lowest hospital mortality occurred among patients with mean glucose levels of 80 to 100 mg/dL during their ICU stay; the rate of hospital mortality increased dramatically for every 20 mg/dL increment above this normal level.
There are relatively few studies in the literature reporting improved outcomes with control of blood glucose in the acute setting. Observational studies1617 of a large group of cardiac surgery patients suggest that tight glycemic control improves mortality and postoperative morbidity, especially infection rates. A landmark randomized controlled study18 performed in a surgical ICU among 1,548 patients receiving mechanical ventilation, 63% of whom had undergone cardiac surgery, demonstrated 34% reduction in mortality and impressive changes in numerous organ system dysfunctions with an intensive regimen that targeted euglycemia. More recently, a before-and-after study19 of intensive glucose management among a heterogeneous population of adult medical and surgical patients, targeting a range of 80 to 140 mg/dL, revealed a 29% decrease in mortality as well as a reduction in the development of new renal insufficiency and the number of patients requiring RBC transfusions.
The costs associated with intensive glycemic management in a critical care population have not been well described. The purpose of this article is to delineate changes in the cost of care associated with this intervention, using actual data from the 1,600-patient study.19
Materials and Methods
Patients and ICU Organization
Stamford Hospital is a 305-bed, university-affiliated community teaching hospital that serves as a major teaching affiliate of Columbia University College of Physicians and Surgeons. The 14-bed ICU treats a heterogeneous population of adult medical, surgical, and cardiac patients. Cardiovascular surgery is not performed at the institution. Although the unit is an "open" unit that allows any credentialed physician to admit patients, the Director of Critical Care and the Director of Surgery closely supervise the unit-based medical and surgical house staff who write all patient care orders. The ICU is highly data and protocol driven.
The clinical outcomes of the 1,600 patients in this study have been reported.19 The 800 patients in the baseline population were admitted consecutively to the ICU between February 23, 2002, and January 31, 2003. The 800 patients in the treatment population were admitted consecutively between February 1, 2003, and January 10, 2004. The intervention promulgated in the second group was a protocol to maintain blood glucose levels in the 80 to 140 mg/dL range.
Data Analysis
A comprehensive ICU database was created in 1998 and has since tracked numerous demographic and clinical features for > 6,500 consecutive admissions. The database is linked to several hospital data repositories allowing detailed analysis of clinical and financial parameters. The database was queried to allow analysis of ICU length of stay (LOS) [measured in 0.1-day increments, rather than calendar days]; post-ICU LOS (measured in calendar days); duration of mechanical ventilation (measured in 0.1-day increments); APACHE (acute physiology and chronic health evaluation) II scores20; primary admitting diagnosis to the ICU (determined at the time of ICU admission by the Director of Critical Care or the Associate Director); final hospital discharge status of each patient; and all laboratory, pharmacy, and radiology costs.
Laboratory, pharmacy, and radiology costs were calculated by applying the appropriate fiscal year (FY) 2002, 2003, and 2004 Medicare Cost Report cost:charge ratios to the charge data captured in the database. The costs were calculated for each major diagnostic group (cardiac, respiratory, GI, septic shock, miscellaneous medical and surgical) and fully adjusted for any differences in the prevalence of mechanical ventilation. There were 292 patients during the baseline period and 230 patients during the treatment period who required mechanical ventilation at the onset of ICU stay. To adjust for this difference, costs in the latter period were recalculated assuming that the number of patients requiring ventilation had not changed, using the actual mean costs for ventilated and nonventilated patients. The "variable" cost of an ICU day and a non-ICU day were calculated by dividing the total cost for each type of bed during the fiscal year, after overhead allocations and adjustments were excluded, by the number of bed days for each category during that FY. This yielded a weighted average of $866.47 during the baseline period and $864.37 during the treatment period. The variable cost of a non-ICU day was $403 for the two periods. The cost of a ventilator day was calculated by reviewing charges and applying the appropriate cost:charge ratio. An adjustment for year-to-year charge inflation was made by reviewing the increase in charges from FY 2002 to 2003 and from FY 2003 to 2004. The reported costs are based on FY 2004, adjusted for inflation.
Some patients required readmission to the ICU after initial discharge from the ICU during their hospitalization. Cumulative data from all of the ICU admissions were analyzed. During the baseline period, 45 of the 800 patients were readmitted to the ICU once during their hospitalization and 6 were readmitted twice. During the treatment period, 35 of the 800 patients were readmitted to the ICU once, 4 were readmitted twice, 1 was readmitted three times, and 2 were readmitted four times to the ICU during their hospitalization.
The parameters were assessed for normality. For univariate analyses, continuous variables were compared using the Mann-Whitney rank-sum test (all comparisons were nonparametric) with the values expressed as a median and interquartile range (IQR), whereas categorical variables were compared using a 
2 test. A p value of < 0.05 was considered statistically significant. All analyses were two tailed and performed using statistical software (Version 7.4.4.0; MedCalc Software; Mariakerke, Belgium). The Stamford Hospital Institutional Review Board approved this study.
Results
Demographic and clinical characteristics of the baseline and treatment populations have been described previously.19 The two groups were well matched overall in regards to age, gender, race, distribution of admitting diagnoses, prevalence of diabetes and APACHE II scores. Table 1 lists the major diagnostic categories of the patients. The percentage of patients undergoing mechanical ventilation at any time during their ICU stay was 40.6% during the baseline period and 33.6% during the treatment period. The percentage of patients who required mechanical ventilation at the onset of their ICU stay was 36.5% and 28.8% for baseline and treatment periods, respectively. Table 2 reports the age and APACHE II scores of patients requiring mechanical ventilation and those not requiring mechanical ventilation during the two periods.
|
|
|
|
Resource Utilization
Table 5
reports the median resource (laboratory, pharmacy, and diagnostic imaging) costs per patient during the two periods, stratified by ventilation status. These inflation-adjusted data are derived from the captured charge data, with appropriate application of the Medicare cost:charge ratios for each category of charge and each FY. There was a nonsignificant decrease in resource costs among nonventilated patients in the treatment period and a large decrease in resource costs among the ventilated patients, driven by significant decreases in imaging and laboratory costs.
|
|
|
Close monitoring and intensive treatment of hyperglycemia has become an emerging standard of care among critically ill patients in the last several years. Observational studies1617 among a large cohort of cardiovascular surgery patients have demonstrated substantial decreases in wound infection rates and hospital mortality when patients are treated with continuous IV insulin with the goal of achieving euglycemia. Findings from a study by Van den Berghe et al,18 performed in 1,548 surgical ICUs patients requiring mechanical ventilation, most of whom had undergone cardiac surgery, provided evidence from a prospective, randomized controlled trial that tight glycemic control yielded impressive reductions in mortality and multiple morbidity. The decreases in the rates of blood stream infections, prolonged antibiotic use, renal replacement therapy, prolonged mechanical ventilation, and prolonged ICU stay likely translated into substantial cost savings to the institution. Most recently, a before-and-after trial19 of intensive glucose monitoring and management in 1,600 consecutive patients admitted to the mixed medical-surgical-cardiac ICU of a university-affiliated community hospital reported a 29.3% reduction in hospital mortality in the treated patients, as well as a decrease in the development of new renal insufficiency, the need for RBC transfusions, and a decrease in the ICU LOS. This report analyzes the major components of the cost of care for these 1,600 patients. To our knowledge, there has been no other economic analysis of the consequences of tight glycemic control in a critically ill population.
The main strength of this article is the comprehensive nature of the data reported, allowing the capture of all the major components of the cost of care. This is attributable to the robust and powerful nature of the ICU database, created in 1998 and now including detailed information regarding > 6,500 consecutive admissions to the unit. The database measures LOS and ventilator duration in 0.1-day increments, avoiding the inaccuracies inherent in calendar-day measurements. APACHE II scoring is performed by the director of critical care or his associate, ensuring accuracy and consistency. The data reported include information from any readmissions to the ICU that may have occurred during the hospitalization. Laboratory, pharmacy, and imaging charges are complete due to the linkages between the ICU database and the central billing database. Charge data are converted to costs using inflation adjustments and the appropriate cost:charge ratios generated by the Medicare cost reports for each FY.
There are some limitations to the analysis that must be explored. The study has a "before-and-after," not randomized, controlled prospective design. However, the 800 patients in the baseline group were well matched to the 800 patients in the treatment group, with no significant differences in age, gender, distribution of admitting diagnoses or severity of illness, as reflected by the APACHE II score. The number of patients requiring mechanical ventilation at the onset of ICU admission was slightly higher among the patients in the baseline group than among the patients in the treatment group; the cost analysis adjusted all the components of care for this difference. Finally, the resource utilization costs were derived by the indirect methods described above. Though this is standard methodology, it does not yield an exact measure of cost, only the best estimate.
The savings associated with the intensive glucose management program were not shared equally among the different groups of patients. The largest net savings occurred among surgical, cardiac, and GI patients. The largest net deficit occurred among the small group of patients (43 in the baseline group and 45 in the treatment group) with septic shock. Notably, however, the mortality rate among this group of the most severely ill patients was 60.5% among the baseline patients and 33.3% among the treatment patients (p = 0.020).19 Perhaps the higher costs among the septic patients in the treatment group occurred because a subset of new "expensive" survivors was created. The data and study design do not allow this hypothesis to be tested explicitly.
The quantity of savings associated with this intervention, $1,580 per patient, may understate the actual savings. The categories of resource utilization measured in this study—laboratory, pharmacy, and radiology—do not encompass all ICU costs. Indeed, a recent review21 of data from 253 hospitals that included 51,009 patients quantified the cost of intensive care admission using a large administrative database. The authors estimated the total cost of the first day of ICU admission as $10,794 for patients requiring mechanical ventilation and $6,667 for those patients not requiring mechanical ventilation.21 The corresponding costs for day 2 were $4,796 and $3,496, and for all subsequent days were $3,968 and $3,184, respectively. Using the assumptions detailed in Table 6, the mean daily cost of the components of care measured in this study was $2,708 for the 1,600 patients.
In summary, this study reports the cost implications of an intervention, intensive glycemic control in critically ill adult patients, that has substantial beneficial effects on patient morbidity and mortality. The substantial annualized savings that accrued to the institution, conservatively estimated at $1,339,500, or $1,580 per patient, were due to decreases in all major categories of resource utilization. Extension of these findings to large numbers of ICUs could have a major impact on national health-care expenditures. These findings provide further impetus for the emergence of tight glycemic control as a standard of care among the critically ill.
Acknowledgements
The authors thank Alex Kiss, PhD, for assistance with statistical analysis, and Paul Sachs, MD, for his ongoing efforts to help maintain the ICU database.
Footnotes
Abbreviations: APACHE = acute physiology and chronic health evaluation; FY = fiscal year; IQR = interquartile range; LOS = length of stay
Dr. Krinsley is Director of Critical Care, and Mr. Jones is Chief Financial Officer, Stamford Hospital.
Received for publication March 9, 2005. Accepted for publication September 22, 2005.
References
This article has been cited by other articles:
|
|
 |
|
  I. Vanhorebeek, L. Langouche, and G. Van den Berghe Tight Blood Glucose Control With Insulin in the ICU: Facts and Controversies Chest, July 1, 2007; 132(1): 268 - 278. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
