Context.—Emergency department (ED) overcrowding has reached crisis proportions in the United States. Many hospitals are seeking to identify process reengineering efforts to reduce crowding and ED patient length of stay (LOS).

Objectives.—To investigate the impact of a point-of-care testing (POCT) satellite laboratory in the ED of a large academic medical center.

Setting.—The ED of the Massachusetts General Hospital, Boston, Mass.

Design and Outcome Measures.—Evaluation of physician satisfaction, turnaround time (TAT), and ED LOS before and after implementation of a POCT laboratory. ED LOS was measured by patient chart audits. TAT was assessed by manual and computer audits. Clinician satisfaction surveys measured satisfaction with test TAT and test accuracy.

Results.—Blood glucose, urine human chorionic gonadotropin, urine dipstick, creatine kinase–MB, and troponin tests were performed in the ED POCT laboratory. Test TAT declined an average of 87% after the institution of POCT. The ED LOS decreased for patients who received pregnancy testing, urine dipstick, and cardiac markers. Although these differences were not significant for individual tests, when the tests were combined, the decreased LOS was, on average, 41.3 minutes (P = .006). Clinician satisfaction surveys documented equivalent satisfaction with test accuracy between the central laboratory and the POCT laboratory. These surveys also documented dissatisfaction with central laboratory TAT and increased satisfaction with TAT of the POCT program (P < .001).

Conclusions.—The POCT satellite laboratory decreased test TAT and decreased ED LOS. There was excellent satisfaction with test accuracy and TAT.

Emergency departments (EDs) in urban areas must adapt to overcrowding, high acuity, and the need to minimize ED patient length of stay (LOS). These problems have been highlighted at the national level and include reports in the lay press emphasizing the need for EDs to reinvent the way they approach their operations.1 Many factors contribute to these problems, including the closing of hospital beds, staff shortages, decreased reimbursements to hospitals, and the decreased access of patients without insurance to medical care. Acute overcrowding may force hospitals to request the diversion of ambulances away from their EDs during very busy times. Developing programs that increase the efficiency of emergency medical care may decrease ED LOS and overcrowding. Long laboratory test turnaround times (TATs) are presumed to contribute to patient LOS in the ED. Consequently, increasing the efficiency of laboratory testing could potentially improve ED operations and thus have a positive impact on patient care.

Many hospital laboratories have difficulty providing prompt service to the ED because of problems with logistics, staff shortages, and technical factors relating to the performance of some tests (particularly immunoassays). Simple methods for whole-blood and urine testing make it possible to perform some laboratory tests rapidly without complex instruments or highly trained personnel.2 These technologies have intuitive appeal, but there are relatively few studies in the peer-reviewed literature to document the impact of point-of-care testing (POCT) on ED operations or patient outcomes.2–4 One study used POCT for a limited menu including sodium, potassium, chloride, glucose, and blood urea nitrogen. In this case, POCT produced no change in ED LOS.3 A Canadian group studied POCT for electrolytes, blood urea nitrogen, creatinine, glucose, hematocrit, and qualitative creatine kinase–MB and myoglobin. They showed a significant decrease in LOS for the patients randomized to the POCT option.4 

Federal legislation governing hospital and outpatient laboratories (Clinical Laboratory Improvement Amendments of 1988 [CLIA '88]) and guidelines from hospital accreditation organizations, including the Joint Commission on Accreditation of Healthcare Organizations and the College of American Pathologists, have complicated the use of POCT.5 Requirements for quality control, quality assurance, documentation, and competency training of staff have made POCT more difficult to implement in the ED, where multiple staff including physicians and nurses may be required to perform patient testing.

We report a study of the implementation of a POCT program in our ED for rapid glucose, urine dipstick, pregnancy testing, and cardiac markers using dedicated personnel in a satellite laboratory. We measured test TAT, patient LOS, and clinician/nurse satisfaction with test accuracy and TAT before and after the initiation of the POCT program.

Study Setting and Population

This study was performed in a large university-associated urban ED (approximately 70 000 annual visits) and was approved by the hospital investigational review board.

Study Protocol

A process improvement team, including representatives from the ED, laboratory, and administration, was established to develop solutions to problems concerning laboratory services for the ED at our institution. The TAT for a selected menu of tests was believed to be contributing to problems with ED operations. The team developed TAT goals for different laboratory tests and assessed the options for meeting these goals. The team thought it impractical to train the 80 nurses, 20 attending physicians, and 48 emergency medicine residents to perform POCT testing and ensure that testing would be compliant with CLIA '88 and Joint Commission on Accreditation of Healthcare Organizations regulations. Therefore, the POCT program was set up in conjunction with the hospital laboratory.

The TAT required by clinicians includes the preanalytic, analytic, and postanalytic phases (so-called “vein-to-brain” time). Although the clinical laboratory is equipped with a pneumatic tube system, the average time to collect and transport specimens from the ED to the pneumatic tube transfer station was 92 minutes, and the average time from the reporting of results in the laboratory until physicians acknowledged these results in the ED information system was 62 minutes. In contrast, the average in-laboratory TAT was 66 minutes. Therefore, the preanalytic, analytic, and postanalytic phases accounted for 41.8%, 30%, and 28.2% of the total TAT, respectively. Meeting our goals for a TAT of 15 minutes (human chorionic gonadotropin), 5 minutes (glucose), 15 to 30 minutes (urinalysis), and 30 minutes (cardiac markers) would not be feasible using the central laboratory. For this reason, we chose to explore the point-of-care option. Having decided to pilot a POCT program, we were concerned with issues of quality control, documentation, competency assessment, and regulatory compliance. For this reason, we decided to use dedicated staff in an ED satellite laboratory as opposed to distributed testing performed by nurses and physicians directly at the bedside. A small area in the ED that had previously functioned as an onsite laboratory with a sink and counter space was used as a central area for testing. The POCT laboratory was staffed with a research nurse (S.M.) and laboratory technicians and was open from 8:00 am to 5:00 pm Monday through Friday. Extended hours were not available during the pilot program because of limitations on staffing, but they have since been implemented. The ED in our institution is divided into 5 distinct areas. For this reason, it was considered most convenient to have a transport staff make rounds every 15 minutes to collect specimens from clinical areas and bring them to the POCT laboratory. During the pilot phase of the POCT program, the results were reported directly to the clinician by telephone, or they were faxed to the area in the ED where the patient was located. Electronic reporting has subsequently been implemented.

Test TAT and patient LOS in the ED were measured for the patients who received glucose testing, urine dipstick, pregnancy testing, and cardiac markers using a combination of time and motion data obtained by human observation and data obtained directly from the hospital information systems. These tests were selected for a variety of reasons. The urine dipstick test and pregnancy test were selected because of physician dissatisfaction with the central laboratory TAT and the availability of simple, reliable POCT test kits from various manufacturers. Although the central laboratory TAT for plasma glucose in our institution is acceptable, it was felt that a rapid whole-blood method for measuring glucose should be available in the ED to facilitate the care of critical patients and to obviate specimen transport issues. The number of whole-blood glucose tests ordered is comparatively small because most glucose requests are ordered as part of an electrolyte panel (chem 7) on plasma specimens. Cardiac markers were added to the protocol in a trial format to determine whether rapid whole-blood results would aid in the triage of patients with chest pain presenting to the ED and to document negative markers in those patients likely to be sent home.

The within-laboratory test TAT was defined as the time the specimen was received in the central laboratory by the hospital pneumatic tube system or in the ED laboratory by the transport personnel until the results were posted in the hospital computer for the clinicians to view or called back to the care unit. The ED LOS was defined as the time from registration in triage to the time of discharge or transport to the floor for admitted patients. These measurements were made before and after the initiation of the POCT program. Clinician and nursing surveys were performed to measure satisfaction with test accuracy and TAT. The survey was administered before and after the start of the POCT program. The clinicians rated their subjective satisfaction with test TAT and test accuracy on a scale of 1 to 5. A rating of 1 indicated the least satisfaction, and a rating of 5 indicated the most satisfaction.

The urine dipstick test was performed with Bayer (Tarrytown, NY) test strips by a Bayer Clinitek 50 Reader. Pregnancy testing on urine samples was performed with Fisher (Houston, Tex) test kits. Glucose testing on blood samples was performed by the Abbott Medisense (Bedford, Mass) PCX system. Cardiac markers including creatine kinase–MB and troponin I were performed with the Spectral Diagnostics (Toronto, Ontario, Canada) Status test kit.

Data Analysis

Statistical analysis included both paired and unpaired t tests by the Statworks statistical program.

A total of 369 patients were evaluated before and during the POCT pilot program. The mean patient age was 51.2 years. There were 158 (42.8%) men and 211 (57.2%) women. Of the patients studied, 162 (43.9%) were admitted to the hospital.

Prior to the institution of the POCT program, we measured the central hospital laboratory test TAT for those tests that would be included in the POCT program. The TAT was measured for samples obtained from ED patients during the same hours of the day that the POCT program would be operational. During the POCT program, clinicians had the option of sending tests to the central laboratory or to the POCT satellite laboratory. The TATs for the selected tests were measured only on samples sent to the POCT laboratory. Table 1 shows the in-laboratory TATs for urinalysis, pregnancy testing, glucose, and cardiac markers. Except for the whole-blood glucose, there was a substantial reduction in the average TAT. The decrease in laboratory TAT for all tests combined was significant (P = .02, unpaired t test).

Table 1. 

In-laboratory Test Turnaround Time for Urinalysis, Pregnancy Testing, Glucose, and Cardiac Markers*

In-laboratory Test Turnaround Time for Urinalysis, Pregnancy Testing, Glucose, and Cardiac Markers*
In-laboratory Test Turnaround Time for Urinalysis, Pregnancy Testing, Glucose, and Cardiac Markers*

The ED LOS for patients who received the selected tests was measured before and during the POCT program, as shown in Table 2. Overall, there was a trend toward decreased ED LOS during the POCT program, except for the patients who received rapid glucose testing. The difference in ED LOS was not significant for any individual test (eg, P = .06 for cardiac markers, unpaired t test) but achieved statistical significance when pregnancy testing, cardiac markers, and urinalysis data were combined (P = .006, paired t test). In contrast, the LOS for patients who did not receive POCT tests increased slightly (11 minutes) during the study period. These numbers are consistent with the decrease in test TAT for all tests except glucose. The data shown in Table 2 are not valid for glucose testing because whole-blood glucose measurements were not available before POCT. Most glucose tests were ordered as part of electrolyte panels and not for the management of diabetic patients in the ED. For this reason, the glucose data before and after the initiation of the POCT pilot are not comparable, as the data reflect testing on different types of patients for different reasons. A P value for statistical significance is not provided in Table 2 for glucose testing, and the data were not included in the aggregate analysis for measuring LOS.

Table 2. 

Emergency Department Length of Stay for Patients Receiving Urinalysis, Pregnancy Testing, Glucose, or Cardiac Markers*

Emergency Department Length of Stay for Patients Receiving Urinalysis, Pregnancy Testing, Glucose, or Cardiac Markers*
Emergency Department Length of Stay for Patients Receiving Urinalysis, Pregnancy Testing, Glucose, or Cardiac Markers*

Quality assurance surveys were performed before and during the institution of the POCT program to measure clinician satisfaction with test accuracy and TAT. Both nursing staff and physicians were surveyed. Physicians order testing in the ED, whereas the nursing staff collects samples and sends them to the central laboratory. The survey results, although subjective, were important to document the success of the POCT program. POCT would be useful only if there was a perceived need for rapid test results (ie, dissatisfaction with the central laboratory TAT) and confidence in the results obtained from POCT (ie, satisfaction with test accuracy of POCT). The results of the survey concerning TAT (Table 3) demonstrated dissatisfaction with the TATs from the central laboratory. These results echoed the general dissatisfaction in the ED with the central laboratory TAT that initiated the effort to develop POCT. During the POCT program, there was a marked increase in satisfaction with the TAT (from a mean of 1.95 before POCT to a mean of 4.3 after POCT), and this difference was significant (P < .001, paired t test). There was an increase in satisfaction with the glucose TAT, even though there was only a 4-minute decrease in TAT. This may have been due to the clinician's having more rapidly received notification of the results (direct call or fax for POCT vs posting on the computer from the central laboratory), or it may reflect a general improvement in satisfaction with laboratory services after the initiation of POCT.

Table 3. 

Clinician Satisfaction With Test Turnaround Times*

Clinician Satisfaction With Test Turnaround Times*
Clinician Satisfaction With Test Turnaround Times*

Clinicians must have confidence in the results of POCT. If the results are perceived to be inaccurate, then the modality will not be used, or the tests will be repeated, reflecting duplicate testing. This would nullify the time savings and possible financial benefit of the POCT program. The results of the survey concerning test accuracy (Table 4) showed excellent satisfaction with the accuracy of both the central laboratory (mean, 4.25 for all tests) and POCT in the satellite laboratory (mean, 4.53 for all tests).

Table 4. 

Clinician Satisfaction With Test Accuracy*

Clinician Satisfaction With Test Accuracy*
Clinician Satisfaction With Test Accuracy*

Compared to central laboratory testing, this study documents a significant decrease in the in-laboratory TAT after the implementation of a POCT program that included urinalysis, urine pregnancy testing, whole-blood glucose, and cardiac markers. This study also demonstrates a decrease in patient LOS in the ED for the patients receiving POCT. Clinicians were satisfied with the accuracy of POCT testing and were significantly more satisfied with the test TAT using the POCT option.

The use of POCT in ED applications has yielded mixed results.1–4,6,7 For example, Kendall et al6 demonstrated a significant decrease in time to medical decision making for a number of patients that received hematologic, biochemical, and arterial blood gas testing, but there was no change in ED LOS. Kilgore et al7 also demonstrated a decrease in therapeutic TAT using the point-of-care option but reported no data on patient LOS in the ED. Their data showed that satellite laboratories received the highest satisfaction ratings from staff, followed by bedside POCT, with the central laboratory receiving the lowest satisfaction ratings. Other studies have confirmed the tendency of POCT to reduce TAT for laboratory testing, but other outcomes such as ED patient LOS were not affected.2 Collectively, the available literature indicates that the implementation of POCT in an ED setting may produce positive benefits for selected patient outcomes but that the POCT program should be targeted to a specific menu and integrated into the overall ED operation. Our results are among the few studies that show a positive outcome from POCT in the ED beyond simply reducing TAT and improving physician satisfaction.

The use of dedicated technical staff in the satellite laboratory allowed the ED to implement onsite testing that was compliant with CLIA '88 and Joint Commission on Accreditation of Healthcare Organizations regulations without the need to train and maintain the competency of multiple ED personnel. This program has also created an infrastructure for expanding the menu of POCT to include a larger menu of tests (eg, electrolytes, liver function, and rapid influenza screening) and to extend the hours of operation. As other rapid testing technologies become available, their introduction into the satellite laboratory may further streamline testing and improve operations in the ED. The staffing of satellite laboratories can be expensive and complex, especially when 7-d/wk, 24-h/d operations are required. Conceptually, it is more cost-effective to use existing ED staff to perform testing. However, issues concerning quality control, documentation, and regulatory compliance can be extremely challenging when CLIA '88–waived and moderately complex tests are performed by a large number of personnel as an add-on to their existing duties. We set up the satellite laboratory to reduce TAT and ED LOS. We wanted to avoid complications with regulatory compliance and felt that this could be best accomplished by using a limited number of dedicated, specially trained technical staff. Admittedly, this staffing arrangement is not necessarily the most cost-efficient model, but it served our purposes at the time. Currently, we are cross-training technologists in the chemistry laboratory and expanding the test menu of the satellite laboratory to increase the utility of the laboratory to the ED and to improve operating efficiency.

Although there was a perceived need for more rapid TAT for selected laboratory tests, we were concerned that clinicians would not trust POCT results and would either not use the service or send duplicate samples to the central laboratory. However, satisfaction with POCT TAT and test accuracy documented the utility of the program. On a national level, physician satisfaction with laboratory ED TAT is not optimal, and many physicians believe long TATs contribute to delayed treatment and increased ED LOS.8 A recent study by the College of American Pathologists reported a survey of 690 hospitals and concluded that the central laboratory ED TAT has remained unchanged for nearly a decade, that physicians are not satisfied with those laboratory services, and that, in many hospitals, laboratory TAT delays treatment and increases ED LOS.8 Our experience with the central laboratory versus POCT testing supports these conclusions.

The implementation of POCT usually involves incremental cost to the institution. Most studies have shown that the unit cost of the POCT option is greater than that of the testing performed in the central laboratory including both labor and consumables.9 However, previous studies did not usually take into account the total system cost of POCT versus central laboratory testing, including the impact of reduced TAT on the overall cost of patient care.9 The impact of POCT on revenue is even more complex. For example, if POCT reduces ED LOS, then presumably, more patients could be evaluated, and revenue would increase. This assumes that there is revenue associated with evaluating more patients, an assumption that may not be true in an urban ED setting. It is possible that treating more patients in the ED increases costs without equivalent revenue to offset the cost. In effect, the ED simply becomes more efficient at losing money. The unit cost of a POC test in our satellite laboratory is approximately $19.20/test including consumables and labor. In contrast, the comparable central laboratory cost per test for this menu and test mix is $2.94. Therefore, the POCT option is approximately 6.5 times more expensive on a unit cost basis than central laboratory testing. On average, each test performed in the satellite laboratory saves 41 minutes of bed time in the ED. An ED bed costs approximately $46/h to maintain. Assuming a testing volume of 25 812 billable tests per year, the total annual bed time saved by the satellite laboratory would be 17 638 hours (2.0 full-time beds). The value of saving bed hours in terms of cost avoidance and revenue gained from new admissions is extraordinarily difficult to calculate. This cost analysis is greatly oversimplified, as it does not account for potential revenue, physician and nurse time, or the cost of resources committed to the ED laboratory that could have been directed toward other services. Accounting for all of these variables is beyond the capabilities of our cost accounting system, leaving the issue of cost-effectiveness unresolved. In this situation, the decision to implement POCT in complex medical services remains largely intuitive. In the final analysis, the fundamental question for our operation is whether 41 bed minutes in the ED is worth $16.26 in incremental laboratory cost.

Our data showed a statistically significant change in the ED LOS for patients having POCT. Many factors contribute to the total LOS for ED patients. Specifically, the LOS may be prolonged in patients who are eventually admitted to the hospital because of delayed bed availability, the need for other non-POCT tests including radiology, wait times for specialty consults, and other factors. The number of patients in our study was too small to evaluate differences in LOS for admitted versus discharged patients or other subcategories.

After this study, other operational improvements were implemented in our ED as part of a multidepartment effort to reduce ED divert, including the addition of a triage physician and changes in radiology services. Collectively, these efforts have reduced ED divert hours substantially. In the current month (June 2001), ED divert hours are down 27% from a year ago despite a record number of ED patient visits. We do not know to what extent the satellite laboratory affected these changes, but our opinion is that the effort was a significant contribution to the overall effort and was one of the easier initiatives to implement.

Future studies should focus on more accurately defining the efficiencies achieved by POCT, including optimizing models for financial and quality management and evaluating the utility of new tests that become available for POCT.

After initiating a POCT program in the ED, we were able to show a significantly decreased TAT for selected tests and an overall decreased ED LOS for the patients who received testing. Clinician satisfaction with TAT and the accuracy of testing was documented.

This study was supported in part by grants from Abbott Laboratories (Medisense Division, Bedford, Mass) and Spectral Diagnostics (Toronto, Ontario, Canada).

Landro
,
L.
Emergency rooms hope new technology will ease their trauma.
Wall Street Journal
July 13, 2001
:
B1
.
Peredy
,
T.
and
R.
Powers
.
Bedside diagnostic testing of body fluids.
Am J Emerg Med
1997
.
4
:
400
407
.
Parvin
,
C.
,
S.
Lo
,
S.
Deuser
,
L.
Weaver
,
L.
Lewis
, and
M.
Scott
.
Impact of point-of-care testing on patients length of stay in a large emergency department.
Clin Chem
1996
.
42
:
711
717
.
Murray
,
R. P.
,
M.
Leroux
,
E.
Sabga
,
W.
Palatnick
, and
L.
Ludwig
.
Effect of point of care testing in an adult emergency department.
J Emerg Med
1999
.
17
:
811
814
.
Lee-Lewandrowski
,
E.
and
K.
Lewandrowski
.
Regulatory compliance for point-of-care testing: a perspective from the United States (circa 2000).
In: Lewandrowski K, ed. Point-of-care testing
.
Clin Lab Med
2001
.
21
:
241
254
.
Kendall
,
J.
,
B.
Reeves
, and
M.
Clancy
.
Point-of-care testing: randomized controlled trial of clinical outcome.
Br Med J
1998
.
316
:
1052
1057
.
Kilgore
,
M.
,
S.
Steven
, and
J.
Smith
.
Evaluating STAT testing options in an academic medical center: therapeutic turnaround time and staff satisfaction.
Clin Chem
1998
.
44
:
1597
1603
.
Steindel
,
S.
and
P.
Howanitz
.
Physician satisfaction and emergency department laboratory test turnaround time: observations based on College of American Pathologists Q-probes studies.
Arch Pathol Lab Med
2001
.
125
:
863
871
.
Lee-Lewandrowski
,
E.
and
K.
Lewandrowski
.
Point-of-care testing: an overview and a look to the future.
In: Lewandrowski K, ed. Point-of-care testing
.
Clin Lab Med
2001
.
21
:
241
254
.

Results were presented as a poster at the SAEM Annual Meeting, Atlanta, Ga, May 6–9, 2001.

Author notes

Reprints: Theodore I. Benzer, MD, PhD, Department of Emergency Medicine, Massachusetts General Hospital, Clinics 115, 55 Fruit St, Boston, MA 02114 ([email protected])