Point-of-Care Testing for the Emergency Department Patient: Quantity and Quality of the Available Evidence

An English-language PubMed search was performed using the following free text terms: (“EMERGENCY” AND “POINT OF CARE”) NOT ULTRASOUND (there was a large number of studies regarding the use of ED point-of-care ultrasound). In addition, a vein of studies associated with POCT (but not described as “point of care”) was identified in the area of rapid infectious disease (predominantly influenza) testing. After the initial free text search, the PubMed “similar articles” function was used to identify related studies that avoided the searched free text terms. References were then reviewed to select out those that studied changes in ED processes secondary to introduction of POCT. For a reference to be included it had to be designed such that parameters were compared between POCT and non-POCT tested ED populations. Inclusion of ED-LOS as a studied metric was not a requirement. Excluded references consisted of those that compared ED-POCT to central laboratory test characteristics or demonstrated new methodologies with hypothesized ED uses.

Selected references were then sorted by test/condition, with further distinction into studies that reported significant changes or not with special attention given to ED-LOS. Additional reported parameters (regardless of significance) included but were not limited to inpatient LOS, change in testing practice, change in treatment plan, disposition, or use of additional diagnostic services. Identified relevant systematic reviews of existing studies were also included in the literature review and were used as benchmarks against the findings in related individual studies.

A total of 1764 references were initially identified. After reference review (abstract and entire study), 74 studies were selected for further review. All demonstrated that POCT results were resulted and reported faster than similar results generated in the central laboratory. The references were then sorted by test/condition into 3 categories; viral-influenza (n = 21,^3–23  with 3 systematic reviews^24–26 ; Table 1), viral-respiratory not otherwise specified (n = 4,^27–29,34  with 4 systematic reviews^30–33 ; Table 2) and nonviral (n = 41)^{35–72,74–76}  tests/conditions, including a systematic review related to lactate use⁷³  (Table 3).

The “nonviral” group was further subdivided into the following test/condition groups: chemistry (n = 21)^45–65 ; cardiac (n = 10)^36–44 ; C-reactive protein (n = 5)^66–70 ; lactate (n = 2)^73–75 ; drugs of abuse (n = 1)⁷² ; renal (n = 1)⁷⁶ ; pregnancy (n = 1)⁷⁵ ; bacterial/strep (n = 1)³⁵ ; and D-dimer (n = 1).⁷¹  The “chemistry” group took into account POCT instruments designed to test the following analytes in combination or individually: sodium, potassium, chloride, bicarbonate, glucose, creatinine, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, bilirubin, amylase, albumin, total protein, human chorionic gonadotropin, pO2, pCO2, and pH. In some of these studies, cardiac markers (CK-MB, troponin I, and/or myoglobin) were part of the chemistry menu,^{45,46,53–55,57,60,62}  but it was clear from the discussions that they were focused more on general POCT instead of a cardiac evaluation per se.

Tables 1 through 3 summarize each study with a brief description of design and findings, annotated, as necessary, with clarifying comments, such as major reported significant and relevant nonsignificant findings as well as comments regarding the author's use of the term “significant” if the data did not appear to be statistically significant. These summaries are not granular in detail but are provided to give the reader a sense of each study. Readers are encouraged to review the articles themselves to determine if they answer a specific need.

In addition, these tables track in 2 separate columns the following: (1) ED-LOS finding (significant/nonsignificant) and if ED-LOS was not reported in the study (ie, unknown if studied or not); and (2) if no other parameters were studied (yes/no) or not, and if any of them were significant or not as well.

As noted in the viral-influenza group and seen in the systematic reviews, there was a trend toward decreased antibiotic use with an increased and earlier use of antivirals. With respect to ED-LOS, the systematic reviews noted no overall decrease in ED-LOS. However, in the viral influenza group, 7 studies^{3,4,11,14,19,20,22}  showed a significant decrease in ED-LOS, with 4 showing no significant change in ED-LOS.^6,12,18,21  The remaining 10 studies^{5,7–10,13,15–17,23}  did not report any findings related to ED-LOS.

In terms of the viral-respiratory not otherwise specified studies, 1 study²⁷  with ED-LOS findings had a significant change in ED-LOS but with increased ED-LOS, not decreased, and the other reference²⁹  showed no significant change. Two^28,34  additional studies did not report any ED-LOS–related findings. Their related systematic reviews^30–33  did not show any improvement in ED-LOS with POCT. Lastly, 1 study³⁴  combined the use of procalcitonin with viral POCT to show that when used in conjunction, antibiotic usage could be reduced.

In the nonvirus category (all groups), 14 studies^* showed a significant decrease in ED-LOS, and 12 showed no change.^† The remainder (n = 15)^‡ did not report any ED-LOS findings. When those with significant ED-LOS findings are broken down by test/condition category, we are left with the following: cardiac^37,43,44  (n = 3), chemistry^{47–49,52,54,56,59,61}  (n = 8), and 1 each in C-reactive protein,⁶⁸  drugs of abuse,⁷²  and lactate⁷⁴  groups. The lactate systematic review⁷³  reported improved hospital mortality and opposing findings related to hospital LOS but nothing about ED-LOS.

As indicated in Tables 1 through 3, a variety of additional significant findings were identified on a case-by-case basis, but it was not clear if they actually improved patient outcomes. Of note, several studies did show decreased admissions^{4,7,11,13–15,19}  and decreased hospital stays,^§ which could a priori reflect improved outcomes. However, inpatient LOS is not a typical reason for requesting ED-POCT.⁴²  Notably, several studies^45,59–62  reported process changes in ED patient care processes to take advantage of the POCT, with significantly decreased ED-LOS seen in 2 of them,^59,61  2 reporting no significant change,^45,60  and 1 not reporting findings related to ED-LOS.⁶²  The latter study looked at 11 different combinations of the following ED diagnostic processes (i-STAT (I-STAT refers to the muti-analyte handheld POC analyzer available from Abbott: complete blood count, radiology and electrocardiogram (ECG); the study found that all significantly improved treatment time except the combination of radiology and electrocardiogram. In this study, the author explicitly stated that ED-LOS was not evaluated because of the presence of confounding variables that could add time to a patient discharge. Of note, 17 studies reported either no significant findings for any parameter or had findings reported without evidence of statistical significance, using terms such as “better,” “positive impact,” improved, etc.^‖

Emergency department POCT influenza testing consistently demonstrates evidence supporting that use of such reduces antibiotic use and prompts and earlier and wider use of antiviral medication. This was seen in both this review and associated systematic reviews on the use of POCT influenza testing. Clearly, it is a significant tool for improving care in those presenting with upper respiratory disease to the ED in terms of diagnosis and treatment; however, it has not been shown to consistently improve ED-LOS. These findings were not seen with viral-respiratory not otherwise specified groups in both the papers studied and associated systematic reviews.

Non–viral-related POCT had similar findings; namely, significant changes on a study-by-study basis across a range of ED processes regardless of test/condition, but inconsistent significant improvement for the most important common denominator of ED patient care, LOS. In fact, the absence of ED-LOS findings in a number of ED-POCT studies raises the possibility that studies may have omitted ED-LOS findings if nonsignificant in favor of presenting only significant findings for other processes.

What is unclear from this review is why some studies showed significant ED-LOS decreases and why others did not. There was not a single study where laboratory test turnaround time did not improve, as expected, but it was unclear as to why some studies showed a decrease when others did not when looking at similar test/conditions (focusing on the “chemistry group”). Most likely these differences go back to the fact that processes differ across institutions and that patient care (ED or otherwise) is a multifactorial process,^42,51,55,65  whereas improvement of one factor does not translate into an overall decreased patient LOS.^42,52  Several studies did note in their discussions and findings that significant changes may have been associated with overall quality improvements^48,49,55,56  that had occurred prior to or as part of the study. In addition, several studies could not rule out the Hawthorne effect (observer bias) as the reason for their data.^42,60,64 

Emergency department POCT can have both significant and nonsignificant impacts on patient care depending on the parameter studied, especially ED-LOS, and depending on the study evaluated. In theory, a rapid laboratory test result would decrease ED-LOS, but this finding is not consistently seen across a wide range of tests and conditions. It stands to reason that to make that rapid test result decrease LOS, it should be coupled to a change in other ED-related patient care processes so that patients can be rapidly discharged to either home or admission.

For unclear reasons, the number of identified studies appeared to be disproportionately low relative to the ubiquity of ED POCT in use (demonstrated by where the identified studies originated). Perhaps this study will trigger additional studies attempting to prove that ED-POCT is worthwhile and can shorten ED-LOS.

Regardless of findings identified (significant or otherwise), the data in this review should help EDs and laboratories assess their needs against what has and has not been studied. This review should also help hospitals consider the need to address process improvements to be integrated with the use of POCT to improve the ED patient care experience with a decreased LOS.