Length of Stay in Emergency Departments: Variation Across Classifications of Clinical Condition and Patient Discharge Disposition Ernest Moy MD, MPH, Rosanna M. Coffey PhD, Brian J. Moore PhD, Marguerite L. Barrett MS, Kendall K. Hall MD, MS PII: DOI: Reference:
S0735-6757(15)00807-4 doi: 10.1016/j.ajem.2015.09.031 YAJEM 55295
To appear in:
American Journal of Emergency Medicine
Received date: Revised date: Accepted date:
16 December 2014 15 September 2015 21 September 2015
Please cite this article as: Moy Ernest, Coffey Rosanna M., Moore Brian J., Barrett Marguerite L., Hall Kendall K., Length of Stay in Emergency Departments: Variation Across Classifications of Clinical Condition and Patient Discharge Disposition, American Journal of Emergency Medicine (2015), doi: 10.1016/j.ajem.2015.09.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Title: Length of Stay in Emergency Departments: Variation Across Classifications of Clinical Condition and Patient Discharge Disposition a
b
c*
T
Authors: Ernest Moy, MD, MPH ; Rosanna M. Coffey, PhD ; Brian J. Moore, PhD ; Marguerite L. d e Barrett, MS ; Kendall K. Hall, MD, MS
NU
MA
* Corresponding Author: Brian J. Moore, PhD Truven Health Analytics 777. E. Eisenhower Parkway Ann Arbor, MI 48108 USA 734.913.3412;
[email protected]
SC
RI P
Affiliations: a Agency for Healthcare Research and Quality, Center for Quality Improvement and Patient Safety, Rockville, MD, USA b Truven Health Analytics, Bethesda, MD, USA c Truven Health Analytics, Ann Arbor, MI, USA d M.L. Barrett, Inc., Del Mar, CA, USA e MedStar Health, Washington, DC, USA
ED
Meetings: This work was presented at the Society for Academic Emergency Medicine 2014 Annual Meeting as an oral presentation by BM and at the Academy Health 2014 Annual Research Meeting as a poster presentation by EM.
CE
PT
Funding: This study was funded by the Agency for Healthcare Research and Quality (AHRQ) under a contract to Truven Health Analytics to develop and support the Healthcare Cost and Utilization Project (HCUP) (Contract No. HHSA-290-2013-00002-C). The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
AC
Conflicts of Interest: The authors have no conflicts of interest to report. Keywords: emergency department; length of stay; duration; quality measurement; case mix
ACCEPTED MANUSCRIPT ABSTRACT Study objective: Duration of a stay in an emergency department (ED) is considered a measure of quality, but current measures average lengths of stay across all conditions. Research on
T
condition-specific length of stay has been limited to a single condition or a few hospitals. We
RI P
use a census of one state’s data to measure length of ED stays by patients’ conditions and dispositions and explore differences between means and medians as quality metrics.
SC
Methods: The data source was the Healthcare Cost and Utilization Project 2011 State
NU
Emergency Department Databases and State Inpatient Databases for Florida. Florida is unique in collecting ED length of stay for both released and admitted patients. Clinical Classifications
MA
Software was used to group visits based on first-listed ICD-9-CM diagnoses. Results: For the 10 most common diagnoses, patients with relatively minor injuries typically
ED
required the shortest mean stay (3 hours or less); conditions resulting in admission or transfer tended to be more serious, resulting in longer stays. Patients requiring the longest stays, by
PT
disposition, had discharge diagnoses of nonspecific chest pain (mean 7.4 hours among
CE
discharged patients), urinary tract infections (4.8 hours among admissions), and schizophrenia (9.6 hours among transfers) among the top 10 diagnoses.
AC
Conclusion: ED length of stay as a measure of ED quality should take into account the considerable variation by condition and disposition of the patient. ED length of stay measurement could be improved in the U.S. by standardizing its definition; distinguishing visits involving treatment, observation, and boarding; and incorporating more distributional information.
ACCEPTED MANUSCRIPT 1. INTRODUCTION
T
1.1. Background and Importance
RI P
The amount of time that a patient spends in the emergency department (ED) has become an increasingly discussed quality measure, as length of stay and ED crowding have
SC
been linked to quality of care, patient safety, and treatment outcomes [1-3]. The Centers for Medicare & Medicaid Services (CMS) reports publicly several measures of ED throughput
NU
including the median time from ED arrival to ED departure for patients who are admitted or treated and released [4]. The National Quality Forum (NQF) has endorsed these measures as
MA
scientifically valid [5].
ED
The length of the patient’s stay in the ED may be influenced by several primary factors: the clinical condition of the patient and the need for specialized treatment [6,7]; the size and
PT
composition of the hospital staff [8]; the hospital triage and clinical decision-making processes [9]; capacity levels, particularly during peak times [10,11]; access to technologies for advanced
CE
diagnostic and treatment procedures [12,13]; and hospital disposition policies and ability to
AC
admit patients for treatment [14,15]. Various factors can contribute to the problems of extended stays in the ED due to ED boarding and observation status. ED boarding occurs when a patient has been admitted to the hospital, but inpatient capacity delays the transfer of the patient out of the ED. ED observation occurs when a patient needs to be monitored for an extensive period to rule in a serious condition that leads to admission or to rule out a serious diagnosis so that a routine discharge from the ED is safe. Many of these factors are potentially modifiable, but hospitals facing financial constraints must prioritize their responses. Previous studies have helped guide decisions about where to focus resources to improve ED services [16-18]. Unfortunately, many studies about ED length of
ACCEPTED MANUSCRIPT stay use data for a single condition, a single or few hospitals, or both, which limits generalization of the results for comprehensive ED services and policies [19-24].
T
The National Hospital Ambulatory Medical Care Survey (NHAMCS) contains ED length
RI P
of stay, defined as ED arrival (hour and minutes when the patient first arrived) minus ED discharge (hour and minutes when the patient left the ED) [25,26]. The National Center for
SC
Health Statistics (NCHS) used NHAMCS data from 2009 to analyze how ED wait times (from
NU
arrival at the ED to physician contact) vary by measures of ED crowding [27]. The results showed excessive wait times for patients with high-acuity conditions. These nationally
MA
representative data have not been used to analyze ED visits and their length by a complete list of clinical conditions seen in the ED; the NHAMCS sample is not designed for studying the
ED
thousands of types of clinical conditions seen in the ED, which are more suitably studied using a census of visits.
PT
1.2. Goals of This Investigation
CE
Our primary goal for this study was to describe the length of ED stay by clinical condition
AC
and disposition of patients. Such information might be useful to hospital managers who want to compare their own ED operational length of stay to a state-wide average by clinical condition and by what happened to the patient—admission, transfer, or release. Such information also should be useful to policymakers who aim to stimulate greater efficiencies in health care delivery. Our secondary goal was to explore two different statistics for measuring length of stay—mean and median. The former is a common statistic; the latter has been used in recent ED length of stay measures. We highlight the pros and cons of each statistic for reporting quality metrics for ED performance.
2. METHODS
ACCEPTED MANUSCRIPT 2.1. Study Design and Data Sources This is a retrospective study using data from large databases. The data were primarily
T
from the Healthcare Cost and Utilization Project (HCUP) State Emergency Department
RI P
Databases (SEDD) [28] and State Inpatient Databases (SID) [29] for the State of Florida in calendar year 2011. The SEDD represent the universe of ED visits that do not end in a hospital
SC
admission for all community hospitals (defined as all nonfederal, short term, general and other
NU
specialty hospitals, excluding hospital units of institutions). The SID represent the universe of inpatient admissions for community hospitals, including visits that started in the ED. Data are
MA
available for all payer types, including the privately-insured, Medicare, Medicaid, and uninsured populations. Payer types are derived from the expected payment source on the discharge
ED
abstract. Florida is the only HCUP Partner to provide data on the timing of ED arrival and departure for all patients, including those who are admitted from the ED for an inpatient stay.
PT
We restricted the sample to community nonrehabilitation hospitals as defined with the
CE
2011 AHA Annual Survey of Hospitals. The Florida SID and SEDD data have the power of a census of ED visits, containing 100% of all ED visits made in that state in 2011, with diagnoses
AC
coded according to the International Classification of Disease, Ninth Edition, Clinical Modification (ICD-9-CM) for the United States. The Florida SID and SEDD data used in this study contained over 8.2 million ED visits among 188 community nonrehabilitation hospitals in 2011. 2.2. Methods and Measurements In Florida, ED length of stay is measured as the total time from ED arrival to ED departure. The Florida Agency for Health Care Administration, Patient Data Submission Guide defines ED arrival as the hour on a 24-hour clock during which a patient registers in the ED for services and ED departure as the hour the patient was discharged from the ED. Times are to be
ACCEPTED MANUSCRIPT recorded as integers of hours rather than hours and minutes, and the hour is to be recorded without rounding for minutes. Hospital personnel are instructed to record ED departure as the
T
hour the patient physically left the ED, but we are unable assess compliance with those
RI P
instructions; it is possible that ED personnel apply these instructions differently such as the hour the ED physician issued the discharge directive, rather than when the patient left the facility.
SC
In light of the described method for recording admission and departure hours, we
NU
decided to keep records with zero lengths of stay in the analysis (4.33% of records), as they represent visits lasting less than one hour. For these records, we manually adjusted the length
MA
of stay time to be 0.5 hours in our data, reflecting the midpoint of the possible times. ED boarding and observation stays among ED records would skew the ED LOS
ED
estimates beyond the typical processing of patients through the ED. Ideally, we would stratify records by such situations. However, Florida ED records do not identify boarding and
PT
observation status. To reduce the influence of these records and of obvious errors in coding of
CE
length of stay, we trimmed records with an ED length of stay greater than 24 hours (3.25% of records). We considered lengths of stay of 24 hours or less as reasonable based on a
AC
reviewer’s suggestion, assuming that 24 hours might be needed by clinicians to rule out serious problems underlying presenting symptoms such as unspecified pain. The distribution of length of stay was relatively smooth with a few very high values. To describe clinical groups, we used the principal diagnoses from inpatient stays and the first-listed ED diagnoses from treat and release visits; the 2011 Florida SEDD include up to 10 diagnoses for each record. We classified these into 285 mutually exclusive clinical groups, using the Agency for Healthcare Research and Quality, Clinical Classifications Software (CCS) [30], which organizes ICD-9-CM diagnosis codes into clinically homogeneous groups. For each CCS category, we calculated the mean and median ED length of stay for three possible
ACCEPTED MANUSCRIPT dispositions: ED discharge, inpatient admission, or transfer to another acute care community,
RI P
T
nonrehabilitation hospital.
3. RESULTS
SC
The number of ED visits and length of stay for the 10 most common first-listed conditions among all ED visits in Florida are presented in Table 1. (Results for all first-listed conditions are
NU
available from the corresponding author.) Conditions resulting from relatively minor injuries
MA
(e.g., sprains and strains, superficial injuries and contusions, skin and subcutaneous tissue infections, open wounds of the extremities) had the shortest stays in the ED; these conditions
ED
resulted in a mean stay of 3 hours or less and a median stay of 2 hours. Conditions involving pain with nonspecific or unclear etiologies (e.g., chest, abdomen, or back pain; headache,
PT
including migraine), generally resulted in mean stays of 4 hours or more and median stays of 3 hours or longer. Patients with nonspecific chest pain incurred the longest mean stays (7 hours)
CE
and median stays (4 hours) among the ten most frequent conditions released or admitted.
AC
While Table 1 shows results for all ED visits, Figure 1 presents stratified results for patients discharged, admitted, or transferred. There were substantial clinical differences among patients released, admitted, and transferred. The 10 most frequent conditions admitted to hospitals’ EDs were more life threatening than those resulting in discharge (for one example, heart failure versus sprains and strains were the most frequent admitted versus released, respectively). Also, the mental condition of mood disorders was in the 10 most frequent conditions resulting in admission, but not the top 10 conditions resulting in discharge from the ED. The most frequently admitted conditions spent longer times in the ED (4 to 5 mean hours) than the most frequently released conditions (3 to 4 mean hours), except for nonspecific chest pain.
ACCEPTED MANUSCRIPT ED visits resulting in transfers to another hospital represented just 1% of all ED visits in Florida. Several of the most common conditions transferred included diseases not in the list of
T
the top 10 for admission. For example, abdominal pain, psychoses, intracranial injury,
RI P
convulsions, acute myocardial infarction, and upper limb fractures were common cases transferred. Most of the conditions transferred were more on par with the seriousness of the
SC
conditions admitted from the ED to the same hospital. The 10 most frequent conditions transferred included two mental conditions—(1) mood disorders and (2) schizophrenia and other
NU
psychoses—with extremely long stays (9.2 and 9.6 mean hours, respectively; and medians of 8
MA
hours, for both). Regarding length of stay generally, transferred conditions showed greater variability in length of ED stays compared with those admitted to the same hospital, ranging from 2.9 to 9.6 mean hours for those transferred and from 3.5 to 4.8 mean hours for those
ED
admitted.
PT
The overwhelming majority of patients seen in the ED are treated and released, the 10
CE
most common conditions among those treated and released in Figure 1 were the same as those identified among all ED visits in Table 1, although their order of frequency differed slightly. Their
AC
lengths of stay also were similar. The exception was for nonspecific chest pain, for which the mean ED stay for those eventually released was about 40 minutes longer than for all ED patients; the median times were the same. 4. LIMITATIONS This study has a number of limitations. The analyses used data from one state, so results cannot be generalized to the nation. The ED length of stay statistics are descriptive and do not control for potential confounding patient and hospital factors. For example, the severity of the patients’ condition (e.g., age, risk of mortality, comorbidities, presence of multiple chronic conditions) could influence the length of stay in the ED. Also, hospital differences, such as
ACCEPTED MANUSCRIPT volume of visits, ED capacity, and policies about observing or boarding patients, also could influence the ED length of stay. Diagnosis is generally more accurate in inpatient than ED
T
settings: the inpatient “principal diagnosis” is established after study to be chiefly responsible for
RI P
a patient’s admission; the ED “first-listed diagnosis” is assembled in a pressured environment and may not reflect the primary reason for the visit. Moreover, ED discharge diagnosis does not
SC
always correspond directly to the chief presenting complaint at arrival [31]. In general, all types
NU
of administrative data, whether billing or medical records, are subject to recording errors. It is possible that the times recorded by ED personnel were imprecise. The Florida
MA
Agency for Health Care Administration has a clear definition of total length of stay in the ED and selected a simple approach to recording time. The definition specifies the patient’s time of entry
ED
and time of departure from the ED facility as the point at which times are to be collected. Times are to be recorded as the hour on a 24-hour clock without minutes or rounding of minutes.
PT
Because of this clarity and simplicity, we expect little variation in approaches of ED personnel to
CE
the recording of ED times. We also expect the approaches used by individuals not to vary by condition or disposition within an ED facility. Nevertheless, although it is easier to record arrival
AC
and departure times in integer hours, this method does create a less precise estimate of actual length of stay than would recording both hours and minutes. The recorded time could be off by nearly 1 hour across individual cases. Consequently, our results should be used only as a general guideline rather than the true length of stay in this population. 5. DISCUSSION Previous studies have investigated ED length of stay but have not been able to provide a broad accounting across all types of hospitals (as in Table 1) and for a comprehensive list of conditions (available here upon request). Our very simple analysis of the time that patients spend in the ED in one state for all types of hospitals suggests that a few lessons could be
ACCEPTED MANUSCRIPT learned from measuring ED length of stay systematically in the U.S.—lessons for improving local care and health-system-wide resource allocation. We now know that ED length of stay
T
varies considerably by condition and disposition of the patient. The availability of ED-length-of-
RI P
stay measures and of some standard for comparison would enable local hospitals to assess ED management and to improve ED quality of care and patient satisfaction with ED services. Such
SC
measures should incorporate clinical condition mix of patients seen in the ED. Information for consumers on ED length of stay would enable them to make choices among hospitals, at least
NU
for non-life-threatening situations, and focus hospitals on patient-centered ED improvements.
MA
Currently, comparative data on ED waiting and treatment times are not readily available. Our results show wide differences in length of stay metrics for some conditions, most
ED
notably among patients diagnosed with nonspecific chest pain. While the contrast in times may be striking, there may be a systematic protocol for evaluating patients with nonspecific chest
PT
pain more intensively than other conditions in the ED. It makes clinical sense that a symptom
CE
that can harbinger a life-threatening myocardial infarction be evaluated carefully in the ED before the patient is discharged. These long pain-related stays may reflect the additional time
AC
necessary for repeated testing to rule out a serious condition or to provide time for observation to allow for an unclear diagnosis to declare itself to avoid premature closure of a visit. The limited ED length of stay data among state-wide partners of the Healthcare Cost and Utilization Project—only Florida collects ED length of stay for released, admitted and transferred patients—indicates that work is needed to promote standards for this data collection among U.S. hospitals. ED length of stay, even among visits ending in hospital admissions, is a measure that can be collected and reported on a large scale. Better standards should be important to hospitals because CMS publicly reports ED length of stay and because the Joint Commission uses it in evaluating hospitals. At a minimum, our work suggests that a general measure of ED length of stay should be adjusted by clinical condition mix and by a facility’s
ACCEPTED MANUSCRIPT propensity to admit, transfer, or discharge ED visitors. Work is needed to promulgate better collection of data elements in administrative data that contribute to ED-length-of-stay measures
T
and better risk adjustment in constructing these measures.
RI P
The choice of mean or median as a preferred measure of the central tendency of a distribution also deserves additional discussion. The mean accounts for the values throughout
SC
the distribution, whereas the median is not influenced by the actual values above and below the
NU
midpoint. We analyzed the distribution with and without the longest stays recorded in Florida (10 days in one case) but reported in the body of the manuscript only the trimmed analysis. The
MA
mean was almost double the median in both analyses. This raises questions in devising ED length of stay metrics for performance monitoring. Should ED length of stay be trimmed when
ED
presenting length of stay by condition or other characteristics? Previous research has suggested that percentile trimming of length of stay within hospital, as opposed to trimming
PT
above an absolute value, may be more appropriate for hospital-level comparisons [32].
valuable insights.
CE
Trimming high end values arbitrarily can throw out valid information, which could lead to
AC
Parallel arguments apply to the question of whether mean or median should be used for hospital performance measurement. Median is recommended by the National Quality Forum for ED-length-of-stay measures. Median statistics avoid the issue of capacity constraints that result in boarding of patients awaiting admission or transfer—waits that are beyond the control of the ED unit. Use of medians also avoids the issue of data trimming for unreasonable values. But use of medians, because they are not influenced by very high values, reduces the information about performance. However, because observation and boarding practices have changed over time, it may be important for fair comparisons to stratify length of stay by these situations.
ACCEPTED MANUSCRIPT Considerable work remains before standards for collecting such data are clear and facilities collect such data consistently. Promoting such standards can lead to patient
T
centeredness, transparency, and care improvement. Public information on ED efficiency, by
RI P
clinical condition mix and patient disposition, would empower patients to select hospital EDs in their communities based on how well patient care is managed, especially for visits where
NU
hospitals to improve the ED services they provide.
SC
patients have the time to make deliberate decisions. This would create powerful incentives for
6. ACKNOWLEDGEMENTS
MA
The authors would like to acknowledge the Florida Agency for Health Care Administration for contributing data to the HCUP State Inpatient Databases (SID) and State
ED
Emergency Department Databases (SEDD) and the HCUP Partner organizations that participated in the HCUP Nationwide Emergency Department Sample (NEDS). A full list of
PT
HCUP Data Partners can be found at www.hcup-us.ahrq.gov/hcupdatapartners.jsp. The
CE
authors also acknowledge Linda Lee, PhD, for her editorial contributions. This study was funded by the Agency for Healthcare Research and Quality (AHRQ) under a contract to Truven Health
AC
Analytics to develop and support the Healthcare Cost and Utilization Project (HCUP) (Contract No. HHSA-290-2013-00002-C). 7. DISCLAIMER The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
ACCEPTED MANUSCRIPT REFERENCES [1]
Bernstein SL, Aronsky D, Duseja R, et al. The effect of emergency department crowding
Moskop JC, Sklar DP, Geiderman JM, et al. Emergency department crowding, part 1—
RI P
[2]
T
on clinically oriented outcomes. Acad Emerg Med. 2009;16(1):1-10.
concept, causes, and moral consequences. Ann Emerg Med. 2009;53(5):605-611. Sun BC, Hsia RY, Weiss RE, et al. Effect of emergency department crowding on
SC
[3]
[4]
NU
outcomes of admitted patients. Ann Emerg Med. 2013;61(6):605-611. Centers for Medicare & Medicaid Services. Official Hospital Compare Data [database
MA
online]. https://data.medicare.gov/data/hospital-compare. Updated July 17, 2014. Accessed July 30, 2014.
National Quality Forum (NQF). NQF-Endorsed Measures. NQF Web site.
ED
[5]
http://www.qualityforum.org/QPS/QPSTool.aspx. Accessed July 31, 2014. Chakravarthy B, Tenny M, Anderson CL, et al. Analysis of mental health substance abuse-
PT
[6]
[7]
CE
related emergency department visits from 2002 to 2008. Subst Abus. 2013;34(3):292-297. Herring AA, Ginde AA, Fahimi J, et al. Increasing critical care admissions from U.S.
[8]
AC
emergency departments, 2001-2009. Crit Care Med. 2013;41(5):1197-1204. Bucheli B, Martina B. Reduced length of stay in medical emergency department patients: a prospective controlled study on emergency physician staffing. Eur J Emerg Med. 2004;11(1):29-34. [9]
Arya R, Dossantos F, Ohman-Strickland P, et al. Impact of trauma activation on the ED length of stay for nontraumatic patients. Am J Emerg Med. 2010;30(2):311-316.
[10] Wiler JL, Handel DA, Ginde AA, et al. Predictors of patient length of stay in 9 emergency departments. Am J Emerg Med. 2012;30(9):1860-1864.
ACCEPTED MANUSCRIPT [11] McCarthy ML, Zeger SL, Ding R, et al. Crowding delays treatment and lengthens emergency department length of stay, even among high-acuity patients. Ann Emerg Med.
T
2009;54(4):492-503.
RI P
[12] Herring A, Wilper A, Himmelstein DU, et al. Increasing length of stay among adult visits to U.S. Emergency departments, 2001-2005. Acad Emerg Med. 2009;16(7):609-616.
SC
[13] Kocher KE, Meurer WJ, Desmond JS, et al. Effect of testing and treatment on emergency department length of stay using a national database. Acad Emerg Med. 2012;19(5):525-
NU
534.
MA
[14] Lucas R, Farley, H, Twanmoh, J, et al. Emergency department patient flow: the influence of hospital census variables on emergency department length of stay. Acad Emerg Med. 2009;16:597-602.
ED
[15] Pitts SR, Pines JM, Handrigan MT, et al. National trends in emergency department
PT
occupancy, 2001 to 2008: effect of inpatient admissions versus emergency department practice intensity. Ann Emerg Med. 2012;60(6):679-686.
CE
[16] Pines JM, Newman D, Pilgrim R, et al. Strategies for integrating cost-consciousness into acute care should focus on rewarding high-value care. Health Aff. 2013;32(12):2157-2165.
AC
[17] Rabin E, Kocher K, McClelland M, et al. Solutions to emergency department “boarding” and crowding are underused and may need to be legislated. Health Aff. 2012;31(8):17571766. [18] McHugh M, Vandyke K, McClelland M, et al. Improving patient flow and reducing emergency department crowding: a guide for hospitals. http://www.ahrq.gov/research/findings/final-reports/ptflow/section1.html. Published October 2011. Accessed December 2, 2013. [19] Carr BG, Kaye AJ, Wiebe DJ, et al. Emergency department length of stay: a major risk factor for pneumonia in intubated blunt trauma patients. J Trauma. 2007;63(1):9-12.
ACCEPTED MANUSCRIPT [20] Edmondson D, Newman JD, Chang MJ, et al. Depression is associated with longer emergency department length of stay in acute coronary syndrome patients. BMC Emerg
T
Med. 2012;12:14.
RI P
[21] Forster AJ, Stiell I, Wells G, et al. The effect of hospital occupancy on emergency department length of stay and patient disposition. Acad Emerg Med. 2003;10(2):127-133.
SC
[22] Funkenstein A, Malowney M, Boyd JW. Insurance prior authorization approval does not substantially lengthen the emergency department length of stay for patients with
NU
psychiatric conditions. Ann Emergency Med. 2013;61(5):596-597.
MA
[23] Lucas R, Farley H, Twanmoh J, et al. Emergency department patient flow: the influence of hospital census variables on emergency department length of stay. Acad Emerg Med. 2009;16:597-602.
ED
[24] Weiss AP, Chang G, Rauch SL, et al. Patient- and practice-related determinants of
2012;60(2):162-171.
PT
emergency department length of stay for patients with psychiatric illness. Ann Emerg Med.
CE
[25] Centers for Disease Control and Prevention (CDC). National hospital ambulatory medical care survey: emergency department factsheet. CDC Web site.
AC
http://www.cdc.gov/nchs/data/ahcd/NHAMCS_Factsheet_ED_2009.pdf. Accessed August 26, 2014.
[26] Centers for Disease Control and Prevention (CDC). National hospital ambulatory medical care survey: 2010 emergency department summary tables. CDC Web site. http://www.cdc.gov/nchs/data/ahcd/nhamcs_emergency/2010_ed_web_tables.pdf. Accessed August 26, 2014. [27] Hing E and Bhuiya F. Wait time for treatment in hospital emergency departments: 2009. NCHS data brief, no 102. Hyattsville, MD: National Center for Health Statistics. 2012.
ACCEPTED MANUSCRIPT [28] Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project, State Emergency Department Databases [database online]. www.hcup-
T
us.ahrq.gov/seddoverview.jsp. Updated June 27, 2014. Accessed August 26, 2014.
RI P
[29] Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project, State Inpatient Databases [database online]. http://www.hcup-us.ahrq.gov/sidoverview.jsp
SC
Updated June 27, 2014. Accessed August 26, 2014.
[30] Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project,
NU
Clinical Classifications Software (CCS) for ICD-9-CM [database online]. www.hcup-
MA
us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Updated July 29, 2014. Accessed August 26, 2014. [31] Raven MC, Lowe RA, Maselli J, et al. Comparison of presenting complaint vs. discharge
2013;309(11):1145-1153.
ED
diagnosis for identifying “nonemergency” emergency department visits. JAMA.
PT
[32] Lee AH, Fung WK, Fu B. Analyzing hospital length of stay: mean or median regression?
AC
CE
Med Care. 2003;41(5):681-686.
ACCEPTED MANUSCRIPT a
Figure 1. Number of emergency department visits in Florida and length of stay for the 10 most common first-listed conditions that resulted in discharge, admission, and transfer.
NU
SC
RI P
T
Emergency department discharge
AC
CE
Transfer to another hospital
PT
ED
MA
Admission to same hospital
a
ED arrival and departure times are to be reported in integer hours in Florida, corresponding to the hour on the clock at the time of the event and removing the minutes without rounding. For this reason, and because observations are at the patient level and not the hospital level of observation, mean lengths of stay show decimals and medians do not. Abbreviations: COPD, chronic obstructive pulmonary disease; STD, sexually transmitted disease; TB, tuberculosis Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project, State Inpatient Database and State Emergency Department Database, Florida 2011
ACCEPTED MANUSCRIPT Table 1. Number of emergency department visits and length of stay for the 10 most common first-listed a b conditions among all Florida ED visits, excluding transfers and outliers.
232: Sprains and strains
374,597
2.64
239: Superficial injury; contusion
344,636
2.73
251: Abdominal pain
305,785
197: Skin and subcutaneous tissue infections
240,613
205: Spondylosis; disc; and other back problems
217,623
159: Urinary tract infections
RI P
Clinical Classifications Software Category
Median Length of c Stay (hours)
T
Number of Visits
Mean Length of Stay (hours)
2 2 4
2.84
2
3.31
3
211,073
3.87
3
102: Nonspecific chest pain
196,513
6.72
4
236: Open wounds of extremities
173,115
2.62
2
162,805
3.83
3
133,429
2.83
2
NU
SC
4.82
84: Headache; including migraine
MA
235: Open wounds of head; neck; and trunk a
Including transfers had almost no effect on mean length of stay.
b
Outlier observations were defined as records with lengths of stay longer than 24 hours.
c
ED
ED arrival and departure times are to be reported in integer hours in Florida, corresponding to the hour on the clock at the time of the event and removing the minutes without rounding. For this reason, and because observations are at the encounter level and not the hospital level of observation, mean lengths of stay show decimals and medians do not.
AC
CE
PT
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project, State Inpatient Databases and State Emergency Department Databases, Florida 2011