Intensive Care Services in the Veterans Health Administration*

Transcription

1 Original Research CRITICAL CARE MEDICINE Intensive Care Services in the Veterans Health Administration* Peter Almenoff, MD, FCCP; Anne Sales, PhD, RN; Sharon Rounds, MD; Michael Miller, MD, PhD; Kelly Schroeder, BSN, MPH; Karen Lentz, RN; and Jonathan Perlin, MD, PhD, MSHA Objective: We describe the national organization and distribution of intensive care services within the Veterans Health Administration (VHA), the largest single integrated health-care system in the United States. Data Sources: Data come primarily from the 2004 Survey of Intensive Care Units in VHA, an electronically distributed survey of all ICUs in the VHA. Medical directors and nurse managers from all 213 ICUs in the VHA responded to the survey. In addition, we extracted data on the number of ICU admissions and unique veterans served from national VHA databases. Results: The VHA has a geographically dispersed, multilevel system of care with variation in geographic access for eligible veterans (varying from 3.1 to 3.5 ICU beds per 1,000 patient discharges) and variation in service provision (from 10 to 19 level 1 ICUs across four regions). Level 1 ICUs are the highest tertiary-level ICUs, with the full range of subspecialty care. The proportion of beds associated with VHA-developed ICU levels of care ranges from 55% level 1 beds in the Northeast to 73% in the South, while level 4 beds represent 4% of all ICU beds in the South and 10% in the Midwest. Conclusions: Overall, the VHA system has a fair amount of regional variation, but level 1 ICUs are available in all geographic regions, and there are regional clusters of all levels. Adopting a four-level system for rating ICUs may assist in monitoring and assessing the quality of care provided in the smallest, most rural facilities. (CHEST 2007; 132: ) Key words: critical care; hospitals, veterans; organization and administration Abbreviations: CCM critical care medicine; CCRN critical care-certified registered nurse; FTEE full-time equivalent employee; FY fiscal year; RN registered nurse; VA Veterans Affairs; VHA Veterans Health Administration; VISN Veterans Integrated Service Network The Veterans Health Administration (VHA) is the largest national health-care system in the United States. 1 As in the private sector, issues of regionalization, variation in access to care, and use of resources are very important to VHA. As part of ongoing evaluation of structure and function within the VHA, a survey of ICUs was conducted in Prior to this survey, the last survey of VHA ICUs was conducted in 1990, 2 which was a companion to a major survey of all nonfederal hospitals focused on critical care units. 3,4 In addition, a database of all nonfederal hospital critical care units was constructed covering the years 1985 through 2000, which was used to assess trends in ICU beds and costs of care. 5,6 These reports provide comparisons of the VHA approximately 15 years prior to the survey study we report here, as well as comparisons from outside the VHA. A growing body of literature 7 16 describes factors in the organization and provision of intensive care that appear to influence patient outcomes, particularly mortality and length of stay. Some of the most important factors have been used to describe levels of intensive care provided by units within hospitals, analogous to levels used for trauma centers. A three-level system has been proposed for classifying ICUs. 17 The level of intensive care is a hospital characteristic based on services available at each hospital, reflecting the range of specialty and subspecialty care available, as well as specialized ancil- CHEST / 132 / 5/ NOVEMBER,

2 lary services such as pharmacy and respiratory care. In this article, we propose a classification scheme that includes ICUs in very small general hospitals. Challenges faced by all hospitals and health-care systems providing intensive care services in the United States include shortages of intensivists, nursing shortages, and changes in the financing and organization of acute care services. No published studies of ICU services to date have assessed regional differences, particularly in terms of ICU level of care. This analysis was conducted to assess regional differences in access to VHA intensive care services, and in the intensity of services available. In this article, we describe ICU service availability in the VHA by ICU level. Materials and Methods Survey Administration and Content The 2004 ICU survey was conducted under the auspices of the VHA National Program Office for Pulmonary and Critical Care. It was distributed via electronic mail through regional offices to each of 126 Veterans Affairs (VA) medical centers providing acute inpatient care. Regional divisions of VHA are called Veterans Integrated Service Networks (VISNs). There are 21 VISNs within the VHA, responsible for budgeting and developing clinical policy and decision making, such as referral patterns. We aggregated the 21 VISNs into the four major census regions. Where a VISN crossed census regions, we assigned it to the region containing most of the facilities. Medical directors and nurse managers in ICUs were asked to complete the surveys, one survey per ICU. All 126 VA medical centers providing inpatient acute care services responded to the survey (100% response). Survey questions covered the existence of one or more ICUs by type (medical, surgical, cardiac care unit, mixed); numbers of beds; type and number of ICU staff; credentials and training; availability of specialists (eg, neurosurgery, cardiology); patientcare protocols in use; and location and type of support services. Survey questions were designed to elicit information to classify VHA ICUs into the levels of care. The survey instrument is *From the VA Heartland Network (Dr. Almenoff), Kansas City, MO; VA Puget Sound Health Care System (Dr. Sales), Seattle, WA; Providence VAMC (Dr. Rounds), Providence, RI; VISN 1 (Dr. Miller), Boston, MA; St. Louis VA Medical Center (Ms. Schroeder), St. Louis, MO; VA Healthcare Analysis and Information Group (Ms. Lentz), Milwaukee, WI; and Office of the Under Secretary (Dr. Perlin), Department of Veterans Affairs, Washington, DC. Members of the 2004 Survey of Intensive Care Units in VHA Technical Advisory Group are listed in the Appendix. The views expressed in this article are those of the authors and do not reflect the views of the Department of Veterans Affairs. The authors have no conflicts of interest to disclose. Manuscript received December 22, 2006; revision accepted July 24, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Anne E. Sales, PhD, RN, University of Alberta, Clinical Sciences Building, Edmonton, AB T6G 3G2, Canada; anne.sales@ualberta.ca DOI: /chest available as an online Appendix to this article. Data on number of patients discharged from acute care hospitalizations were obtained from national VHA databases. No human subjects were involved in data collection. Respondents provided information related to the performance of their duties in their roles as directors or managers of ICUs. We report findings by dividing the 21 VISNs in the VHA into the four major census regions. Levels of Intensive Care Levels of intensive care were defined as shown in Tables 1, 2, following Haupt et al 17 and Brilli et al. 25 Level 1 is the most complex, including the highest-level tertiary care centers. Levels 1 and 2 differ in whether there is a critical care fellowship program, whether full-time attending physicians are present in the ICU, and whether neurosurgery and cardiovascular surgery are available. When an ICU had some characteristics of a higher level, but not all, we classified it as the lower of the two levels. We calculated the number of level 1, and level 1 and level 2 (combined) beds per 1,000 patient discharges in fiscal year (FY) We used patient discharges from acute care as the relevant denominator because these are patients at risk for intensive care. We described the number and proportion of ICU beds in each region by ICU level, as well as ICU beds as a proportion of total beds by region and level. Organizational Factors Associated With Patient Outcomes How ICUs are organized and the type of physicians practicing in them have been shown to affect patient outcomes. 7,24,26 In particular, whether an ICU has a closed system with all care provided by a dedicated ICU team, or an open system allowing non-icu physicians to provide care, has been shown to be related to patient mortality rates. 7 If ICUs are closed but mandate a consultation with a dedicated physician or team, we treated them as closed units. We combined responses indicating closed units with those that are open but mandate consultation, calculating the proportion of ICUs in each region that are either closed or open with mandated consultation. We also looked at the proportion of ICUs with a critical care medicine (CCM) board-certified director. VHA inpatient care is organized somewhat differently from the care delivered in private-sector hospitals. In tertiary, teaching VHA medical centers, if a patient is admitted to inpatient care, they are admitted to a service; there are no private attending physicians. In nontertiary, nonteaching VHA medical centers, care is organized depending on whether there is a hospitalist or intensivist model of care. If an ICU is designated as closed, or open with mandatory consultation, all patients in that unit will receive care from an intensive care team. If it is designated as open, all patients receive care from their admitting physician or service. In addition, nurse staffing levels have been shown to be related to patient outcomes in ICU. 15,27 We summed the number of full-time equivalent employed (FTEE) registered nurses (RNs) reported for the year in each facility to the regional level. We calculated FTEE RNs per 1,000 patient discharges in each region to look at the geographic distribution of these important components of intensive care personnel. We also calculated the proportion of RNs with Critical Care Registered Nurse (CCRN) certification in each region. 28,29 Results Description of VHA ICUs A total of 213 ICUs were reported at 126 separate locations (Table 3). Seventy hospitals had one ICU, 1456 Original Research

3 Table 1 Criteria for Determination of ICU Level* Diagnostic and Therapeutic Radiology Procedures Laboratory Services Pharmacy Services Cardiac Catheterization Laboratory Operating Room Emergency Department Medicine and Surgery MDs Subspecialists Available Dedicated ICU Attending Physician Intensivist Unit Type Specialty Fellowship Training Program Services Provided ICU Level No Basic Basic Basic Yes, limited scope Yes, limited scope Basic In house or available Level 4 Basic None Open No None, single MD model Yes/no Moderate Moderate Moderate Moderate Yes Yes Yes, limited scope Open No None, multiple consultants, medical doctors with with ACLS or FCCS Moderate Yes/no, no core fellowships (eg, cardiology) Level 3, exceeds criteria for level 4 plus Complex Complex Complex Complex Yes Yes Yes Yes, interventional Yes Intensivistconsultant Open or closed Complex Yes with core fellowships Level 2, exceeds criteria for level 3 4 plus Complex Complex Complex Yes Yes Yes Yes, interventional All plus neurology and/or cardiovascular surgery Yes Full-time attending MD Open or closed Complex Yes, plus critical care fellowship Level 1, exceeds criteria for levels 2 4 plus *MD medical doctor; ACLS advanced cardiac life support; FCCS fundamental critical care support. See Complex column in Table 2 for more information. with a range from one to four ICUs. The South had the largest number of ICUs (n 90). The mix of ICU types varies by region of the country, with single-specialty ICUs dominating in the South, and mixed ICUs dominating in the other three regions. Half the ICUs were single specialty, mostly surgical. There was regional variation in the number of beds in each region (Fig 1, Table 3), and in the proportion of beds in each ICU level. The Northeast had the lowest proportion of level 1 beds, followed by the Midwest, the West, and the South. Standardized by the number of patient discharges in a year, the Midwest had the lowest number of total beds per 1,000 patient discharges, followed by the Northeast, the West, and the South. It is notable that all four regions were similar in the proportion of total ICU beds per 1,000 patient discharges. There was variation in the proportion of ICU to total beds by level across regions, as well as differences in pattern. There was more variation in the proportion of level 1 beds per 1,000 patient discharges (Table 3). Summing level 1 and level 2 beds, however, there was less difference between highest and lowest, with the South still leading with 3.1 beds per 1,000 patient discharges. Organizational Factors Associated With Patient Outcomes Of the 213 ICUs, 126 were either closed or open with mandatory consultation (Table 4). The proportion of ICUs that were closed or open with mandatory consultation varied somewhat across the system, from 41% in the West to 69% in the South. There was little regional variation in the proportion of ICUs in which the director has CCM board certification. RN staffing varied in the proportion of RN FTEEs per 1,000 patient discharges, from a high of 9.2 in the West to 8.2 in the Midwest. Overall, the number and proportion of CCRNs was low, without much variation. Unit managers were also asked to report the RN/patient ratio; on average, this was 1:2, with little variation by region or ICU level. Discussion This study updates information from the 1990 survey of VHA ICUs 2 and provides the first assessment of regional variation in levels of ICU care in the largest integrated health-care system in the United States. In the two articles 3,4 reporting on ICU organization in the United States from the early 1990s, results were aggregated to the entire United States. The 1991 survey reported in these studies preceded development of ICU levels, 17 so results were not presented by level of care. Subsequent work using CHEST / 132 / 5/ NOVEMBER,

4 Criteria Table 2 Expanded Criteria for Key Services* Variables Basic Moderate Complex Subspecialty availability Pharmacy services Laboratory services Diagnostic and therapeutic radiologic procedures Continuous ECG; hourly vitals/ neurologic checks; central venous pressure monitoring; transcutaneous oxygen/pulse oximetry; complex airway management (ventilatorcontrolled respiration, PEEP, CPAP); emergency resuscitation equipment (ACLS algorithm); supplemental oxygen; suction; negative-pressure isolation rooms Very limited subset of specialist and subspecialist by referral Yes, limited; core hours or oncall support for portion of 24 h Limited laboratory services; urgent laboratory; point of care testing; cross and type; occasional blood product via community resource Limited; contractual; access/community resource; CT; MRI; echocardiography; fluoroscopy Basic plus: dialysis; transcutaneous and transvenous pacemaker; variceal tamponade; vasoactive/continuous antiarrythmic drug administration; continuous arterial pressure monitoring; pulmonary artery wedge pressure; cardiac output monitoring Intensivist; urologist; thoracic surgeon; vascular surgeon; anesthesiologist; pulmonologist; gastroenterologist; hematologist; infectious disease; nephrologist; interventional neuroradiologist; interventional radiologist; pathologist; neurologist; orthopedic surgeon ICU pharmacist available; availability of ready to administer unit dose drug distribution services; ability to supply immediate medication and admixtures On site 24/7 testing (hematologic, chemistry, blood gas); 24/7 blood bank Portable chest/interventional radiologic (invasive arterial/venous diagnosis and therapeutic techniques); CT/ angiography; duplex Doppler ultrasound; MRI/angiography; echocardiography; fluoroscopy Moderate plus: fiberoptic bronchoscopy; intracranial pressure monitoring; continuous EEG monitoring; intra-aortic balloon pump Moderate plus: neurosurgeon; cardiovascular surgeon Clinical pharmacist support as team member to support medical/surgical patient management; ICU pharmacist available; availability of ready to administer unit dose drug distribution services; ability to supply immediate medication and admixtures; may have full-time critical care pharmacy satellite On site 24/7 testing (hematologic, chemistry, blood gas, toxicology); 24/7 blood bank; in some instances, point-of-care technology for bedside; urgent laboratory All moderate diagnostic and therapeutic radiology 24/7 *PEEP positive end-expiratory pressure; CPAP continuous positive airway pressure; 24/7 around the clock. See Table 1 for expansion of abbreviation. databases from existing sources 5,6 also preceded development of ICU levels, and use of existing databases limits the ability to assess level of ICU. As a result of the 2004 study we report here, the VHA adopted a system defining four levels of intensive care, rather than the three proposed for use in the private sector in the United States. These levels are closely related to levels of academic affiliation, with the higher levels (levels 1 and 2) associated with tertiary academic medical centers. There is a relatively high proportion of small, often rural, nontertiary hospitals in the VHA that provide acute care services and have a lower level of intensive care, making it necessary to identify an additional level of intensive care with fewer services than those described in the model of Haupt et al. 17 These smaller, more rural VHA hospitals are analogous to non-vha critical-access hospitals, and the VHA fourth level may be useful for small, rural non-vha hospitals Original Research

5 Variables Table 3 ICUs Within the VHA Northeast Region South Region Midwest Region West Region VHA Total VISNs included in region 1, 2, 3, 4, 5 6, 7, 8, 9, 16, 17 10, 11, 12, 15, 23 18, 19, 20, 21, Facilities, No Facilities by ICU level, No. (% of total No. of facilities in region) Level 1 10 (38) 19 (46) 12 (39) 11 (39) 52 (41) Level 2 6 (23) 8 (20) 5 (16) 4 (14) 23 (18) Level 3 5 (19) 8 (19) 7 (22) 6 (21) 26 (21) Level 4 5 (19) 6 (15) 7 (22) 7 (25) 25 (20) ICUs, No ICU types Medical Surgical Cardiac care Mixed Other specialty Total operating ICU beds, No ,906 Operating ICU beds by level, No. (% of total ICU operating beds in region) Level (55) 577 (73) 239 (62) 248 (65) 1251 (66) Level 2 85 (25) 126 (16) 65 (17) 58 (15) 334 (17) Level 3 45 (13) 60 (8) 46 (12) 50 (13) 201 (10) Level 4 22 (6) 32 (4) 38 (10) 28 (7) 120 (6) ICU beds as proportion of total facility beds by level, % Level Level Level Level Total patient discharges in FY 2003, 99, , , , ,178 No. ICU beds per 1,000 patient discharges FY 2003, No. Level 1 ICU beds per 1,000 patient discharges in FY 2003, No. Percentage of national average Level 1 and 2 ICU beds per 1,000 patient discharges in FY 2003, No Concordant with the level of care available, the VHA mandates relatively short lengths of stay where lower levels of intensity are available. VHA policy, adopted in part as a result of this study, recommends that patients should be transferred to an ICU with a higher level of care (level 2 or greater) if they require intensive care 72 h in a level 4 hospital, and 5 days in a level 3 hospital. Transfer may be to a VHA hospital with a level 2 or level 1 ICU but may also be to a non-vha hospital with a higher level ICU if VHA ICU care is not available. In the latter case, the referring VHA hospital is responsible for paying for care for these patients. This latter point would not be relevant outside the VHA because hospitals do not pay for care if the patient is transferred. Within the VHA, however, payment for services outside the VHA can be very costly for smaller hospitals. Variation by region in the proportion of total beds that are ICU beds represents an opportunity to evaluate the type of beds based on patient acuity, and may be a metric worth tracking over time, particularly in relation to issues associated with hospital throughput. Looking at regional differences, we found variation in the proportion of ICUs that are closed or require intensivist consultation, which may be associated with variation in ICU type. In regions with relatively high proportions of single-specialty ICUs, the prevalence of closed units was somewhat higher, while in regions with relatively high proportions of mixed ICUs, the prevalence of closed units was somewhat lower. In addition, the number and proportion of surgical ICUs may affect the proportion of closed units because surgical ICUs are less likely to be closed. Despite variation in unit types, and in open or closed status, a relatively high proportion of ICU directors are board certified in CCM. We note CHEST / 132 / 5/ NOVEMBER,

6 Figure operational (staffed) ICU beds per 100,000 unique patients. that the proportion of RNs with critical care certification is quite low, well 20% in all regions, and does not vary widely across regions. However, the VHA does not appear very different from non-vha hospitals overall, as reported in prior studies, 3 6,33 although none of these studies report regional variation. Compared with the data reported in the 1990 VA study 2 of ICUs, the overall number of ICU beds across VHA has decreased: 2,589 beds in 1990 and 1,906 beds in 2004, from a total of 251 ICUs in 1990 and 213 ICUs in In 1990, 91% of the ICUs in the study had physician directors, compared with 95% in Thirty-nine percent of these were Table 4 Factors Associated With Improved Patient Outcomes Variables Northeast Region South Region Midwest Region West Region VHA Total VISNs included in region 1, 2, 3, 4, 5 6, 7, 8, 9, 16, 17 10, 11, 12, 15, 23 18, 19, 20, 21, Facilities, No ICUs, No Total patient discharges in FY 2003, No. 99, , , , ,178 Total closed or mandatory consult open ICUs, No. Percentage of closed or mandatory consult open ICUs Percentage of ICUs with CCM boardcertified directors RN FTEEs in ICU, No ,020 1,009 1,022 4,898 RN FTEEs per 1,000 patient discharges FY , No. RNs with CCRN certification, No Percentage of RNs with CCRN certification Original Research

7 CCM board certified or eligible in 1990, compared with 62% in Finally, in 1990, in 90% of ICUs, 50% of the RN staff were CCRN certified, while in 2004, 13% of the RNs working in ICU were CCRN certified. Although these figures are not directly comparable, they suggest that as the proportion of certified physicians in VA ICU care has been increasing, the proportion of CCRN-prepared RNs may be decreasing. Limitations The data reported here come from a single crosssectional survey and are self-reported by key individuals. As a result, while the information is as accurate as the data available to the individual reporting it, we were unable to check data validity and reliability, except for the patient discharge data from national databases; their reliability has been examined. 34,35 Although most of the data were self-reported, we know of no reason why individuals would have reported inaccurate data for any systematic reason. We note that we are not able to address issues of occupancy, mortality, length of stay, or other patient outcomes as part of this article because they were not a primary focus of the survey. In addition, we do not discuss reasons for admission to the ICU, nor severity of illness. Conclusions This study was unique in describing intensive care services by ICU level and by region. We describe variation across the VHA in key aspects, including availability of the highest ICU level, and in key factors described in the literature as associated with outcomes. Overall, the picture of the VHA system shows a fair amount of regional variation, but level 1 ICUs are available in all geographic regions, and there are regional clusters of all levels. This suggests that while the VHA should continue to assess regional variation in access to different levels of ICU care, system adjustments may be appropriate rather than major system overhaul. Adopting a four-level system for rating ICUs both within and outside the VHA may assist in monitoring and assessing the quality of care provided in the smallest, most rural facilities. ACKNOWLEDGMENTS: Special thanks is given to the 2004 Survey of Intensive Care Units in VHA Technical Advisory Group for their guidance and support of this project: Peter Almenoff, MD, FCCP, National Program Director Pulmonary and Critical Care Medicine, Network Director VISN 15, Kansas City, MO; Thomas Craig, MD, Medical Services Officer, VAHQ, Washington, DC; Ralph G. Depalma, MD, National Director of Surgery, VAHQ, Washington, DC; Rhonda Eisenzimmer, RN, MSN Clinical Manager, VAMC, Fresno, CA; Robert Jesse, MD, PhD, National Program Director, Cardiology, Director CCU, VAMC Richmond, VA; Wendell Jones, MD, Chief Medical Officer VISN 17, Grand Prairie, TX; Paul Low, MD, MBA, FCCP, VISN 12 Chief Medical Officer, Hines, IL; Michael J. Miller, MD, PhD, Chief Medical Officer VISN 1, Bedford, MA; Marthe J. Moseley, PhD, RN, CCRN, CCNS, Critical Nurse Specialist, Critical Care, VAMC San Antonio, TX; Pamela Reeves, MD, Chief Medical Officer VISN 11, VAMC Ann Arbor, MI; Cathy Rick, RN, Chief Consultant, VAHQ, Washington, DC; Sharon Rounds, MD, Chief, Pulmonary/Critical Care, VAMC Providence, RI; Anne Sales, RN, Health Systems Specialist, VAMC, Seattle, WA; Kelly Schroeder, RN, Program Analyst, VAMC St. Louis, MO. Healthcare Analysis and Information Group staff include Karen Lentz, RN; Kel Schelach, management analyst; Beth Babick, management analyst; and MaryBeth Matthews, computer specialist. References 1 Organizational briefing book. Washington, DC: Department of Veterans Affairs, Halpern NA, Wang JK, Alicea M, et al. Critical care medicine: observations from the Department of Veterans Affairs intensive care units. Crit Care Med 1994; 22: Groeger JS, Strosberg MA, Halpern NA, et al. Descriptive analysis of critical care units in the United States. Crit Care Med 1992; 20: Groeger JS, Guntupalli KK, Strosberg M, et al. Descriptive analysis of critical care units in the United States: patient characteristics and intensive care unit utilization. Crit Care Med 1993; 21: Halpern NA, Pastores SM, Greenstein RJ. Critical care medicine in the United States : an analysis of bed numbers, use, and costs. Crit Care Med 2004; 32: Halpern NA, Pastores SM, Thaler HT, et al. Changes in critical care beds and occupancy in the United States : differences attributable to hospital size. Crit Care Med 2006; 34: Pronovost PJ, Angus DC, Dorman T, et al. Physician staffing patterns and clinical outcomes in critically ill patients: a systematic review. JAMA 2002; 288: Burchardi H, Moerer O. Twenty-four hour presence of physicians in the ICU. Crit Care 2001; 5: Chang SY, Multz AS, Hall JB. Critical care organization. Crit Care Clin 2005; 21: Pronovost P, Sexton B. Assessing safety culture: guidelines and recommendations. Qual Saf Health Care 2005; 14: Pronovost PJ, Rinke ML, Emery K, et al. Interventions to reduce mortality among patients treated in intensive care units. J Crit Care 2004; 19: Dang D, Johantgen ME, Pronovost PJ, et al. Postoperative complications: does intensive care unit staff nursing make a difference? Heart Lung 2002; 31: Dimick JB, Swoboda SM, Pronovost PJ, et al. Effect of nurse-to-patient ratio in the intensive care unit on pulmonary complications and resource use after hepatectomy. Am J Crit Care 2001; 10: Dimick JB, Pronovost PJ, Heitmiller RF, et al. Intensive care unit physician staffing is associated with decreased length of stay, hospital cost, and complications after esophageal resection. Crit Care Med 2001; 29: Amaravadi RK, Dimick JB, Pronovost PJ, et al. ICU nurse-to-patient ratio is associated with complications and resource use after esophagectomy. Intensive Care Med 2000; 26: CHEST / 132 / 5/ NOVEMBER,

8 16 Pronovost PJ, Jenckes MW, Dorman T, et al. Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA 1999; 281: Haupt MT, Bekes CE, Brilli RJ, et al. Task Force of the American College of Critical Care Medicine. Society of Critical Care Medicine guidelines on critical care services and personnel: recommendations based on a system of categorization of three levels of care. Crit Care Med 2003; 31: Irwin RS, Marcus L, Lever A. The critical care professional societies address the critical care crisis in the United States. Chest 2004; 125: Ewart GW, Marcus L, Gaba MM, et al. The critical care medicine crisis: a call for federal action: a white paper from the critical care professional societies. Chest 2004; 125: Kelley MA, Angus D, Chalfin DB, et al. The critical care crisis in the United States: a report from the profession. Chest 2004; 125: Angus DC, Kelley MA, Schmitz RJ, et al. Committee on Manpower for Pulmonary and Critical Care Societies (COMPACCS). Caring for the critically ill patient: current and projected workforce requirements for care of the critically ill and patients with pulmonary disease; can we meet the requirements of an aging population? JAMA 2000; 284: Aiken LH, Buchan J, Sochalski J, et al. Trends in international nurse migration. Health Aff (Millwood) 2004; 23: Rogers AE, Hwang WT, Scott LD, et al. The working hours of hospital staff nurses and patient safety. Health Aff (Millwood) 2004; 23: Aiken LH, Clarke SP, Sloane DM, et al. Cause for concern: nurses reports of hospital care in five countries. LDI Issue Brief 2001; 6: Brilli RJ, Spevetz A, Branson RD, et al. Critical care delivery in the intensive care unit: defining clinical roles and the best practice model. Crit Care Med 2001; 29: Berenholtz S, Pronovost PJ. Barriers to translating evidence into practice. Curr Opin Crit Care 2003; 9: Pronovost PJ, Dang D, Dorman T, et al. Intensive care unit nurse staffing and the risk for complications after abdominal aortic surgery. Eff Clin Pract 2001; 4: Mitchell PH, Armstrong S, Simpson TF, et al. American Association of Critical-Care Nurses Demonstration Project: profile of excellence in critical care nursing. Heart Lung 1989; 18: Newhouse RP, Johantgen M, Pronovost PJ, et al. Perioperative nurses and patient outcomes mortality, complications, and length of stay. AORN J 2005; 81: Freeman VA, Walsh J, Rudolf M, et al. Intensive care in critical access hospitals. J Rural Health 2007; 23: Moscovice I, Wellever A, Sales A, et al. A clinically based service limitation option for alternative model rural hospitals. Health Care Financ Rev 1993; 15: Wellever AL, Weisgrau S, Wagner M. A model in Montana: ensuring access to care. Patient Acc 1991;14:2, 7 33 Halpern NA, Bettes L, Greenstein R. Federal and nationwide intensive care units and healthcare costs: Crit Care Med 1994; 22: Kashner TM. Agreement between administrative files and written medical records: a case of the Department of Veterans Affairs. Med Care 1998; 36: Rosen AK, Loveland S, Anderson JJ, et al. Evaluating diagnosis-based case-mix measures: how well do they apply to the VA population? Med Care 2001; 39: Original Research