Skip to main content
Download PDF

Implementation of a rapid host-protein diagnostic test for distinguishing bacterial and viral infections in adults presenting to urgent care centers: a pragmatic cohort study

BMC Medicine volume 23, Article number: 63 (2025) Cite this article

Abstract

Background

Urgent care centers (UCCs) are a growing segment of healthcare with high rates of inappropriate antibiotic use. MeMed BV® (MMBV) is a blood test that differentiates bacterial from viral infections. Between April 2022 and March 2023, we introduced MMBV into routine care at ten UCCs. The primary objective was to assess MMBV’s impact on antibiotic use; the secondary objective was to assess whether MMBV aided in patient management.

Methods

A pragmatic prospective cohort study. Physicians who ordered MMBV reported electronically (in real-time) whether they intended to prescribe antibiotics before ordering the test and upon UCC discharge whether MMBV aided in patient management. Hospitalizations were recorded for 7 days post-UCC discharge.

Results

During implementation, 3920 MMBV tests were ordered for adults (age ≥ 18) by 144 physicians. The study cohort had 59% female patients and the median age was 42 years (IQR 31–58). For the primary objective, 3262 cases were included. MMBV indicated 629/3262 (19.3%) cases of potentially unwarranted antibiotics, of which physicians avoided prescriptions in 397/629 (63.1%). MMBV indicated 405/3262 (12.4%) cases of potentially missed bacterial infections. Physicians prescribed antibiotics to 283/405 (69.9%). MMBV adherence was associated with fewer hospitalizations (7.8% vs. 30.3%, p < 0.001). For the secondary objective, 2901 cases were included. Physicians reported MMBV aided patient management in 2494/2901 (86.0%) cases and contributed to avoiding emergency department referrals in 595/2901 (20.5%).

Conclusions

Implementing MMBV aided urgent care center physicians in their clinical decision-making and may have contributed to appropriate antibiotic use, better resource utilization, and patient management.

Peer Review reports

Background

Urgent care centers (UCCs) are a rapidly growing sector of the outpatient care setting [1, 2]. UCCs provide higher-level care than physician offices while alleviating some of the burden from emergency departments (EDs) [3,4,5].

Unwarranted use of antibiotics is common in outpatient settings, with rates as high as 46% in UCCs compared to 14–17% in other outpatient settings and 24.6% in EDs. Despite advances in diagnostic technologies, uncertainty in differentiating bacterial and viral infections persists, contributing to unwarranted antibiotic use [7,8,9].

MeMed BV® (MMBV) is a rapid blood test that differentiates bacterial and viral infections based on the computational integration of three immune proteins differentially expressed in response to acute infections: tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), interferon-gamma induced protein-10 (IP-10), and C-reactive protein (CRP) [9]. MMBV outputs a score ranging from 0 to 100, with higher scores indicative of bacterial infection (or bacterial/viral co-infection). MMBV’s diagnostic accuracy was validated in multiple observational studies on children and adults [6, 10,11,12,13,14]. In these studies, MMBV was evaluated against an adjudication-based reference standard for infection etiology. A panel of three or more experienced clinicians independently reviewed comprehensive case files—including clinical, laboratory, microbiological, radiological data and follow-up—to adjudicate cases as bacterial or viral. This methodology enables cases to be assigned an infection etiology when there is no “gold” standard microbiologically confirmed diagnosis and is acknowledged as a robust method for assessing the performance of diagnostic tests [15,16,17]. In a study of adults presenting to the ED with suspected lower respiratory infection, MMBV’s sensitivity for detecting bacterial infection or co-infections in adult patients was validated as 98.1% (95% confidence interval (CI) 95.4–100), with specificity 88.4% (95%CI: 83.7–93.1) and negative predictive value 98.8% (95%CI: 97.1–100) [6]. Similarly, other studies have reported sensitivity and/or specificity above 90% [10,11,12,13,14]. The area under the receiver operating characteristic curve (AUROC) for MMBV ranges from 0.91 to 0.98. Notably, this performance measure is calculated across all MMBV score thresholds (0–100) and includes cases with viral (< 35), equivocal (35–65), and bacterial (> 65) scores (11–15).

While such diagnostic accuracy studies establish that MMBV correctly differentiates bacterial and viral infections, they cannot provide insights into whether the test aids physicians in patient management in real-world settings. Recently, we implemented MMBV in a pilot study at three UCCs [18]. Here, we examined MMBV’s contribution to appropriate antibiotic use and whether it aids physicians in patient management during nationwide implementation in a network of ten UCCs.

Methods

Settings

The study was conducted at Maccabi Healthcare Services (MHS), Israel’s second largest Health Maintenance Organization, serving over 2.6 million citizens. MHS operates ten primary care outpatient UCCs nationwide, seeing patients of all ages. This network receives approximately 195,000 adult visits (excluding trauma, obstetrics, and gynecology) per year. The UCCs provide urgent medical care after-hours, i.e., between 19:00 and 23:00 on weekdays and 09:00 and 23:00 on weekends. They have on-site nursing and primary care physician services, lab facilities (including biochemistry and microbiology), and X-rays.

Study design

A pragmatic prospective cohort study designed to examine how routine use of MMBV affects antibiotic use and patient management at the MHS UCC network.

After initial evaluation, including patient history and physical examination, physicians could order MMBV tests for their patients alongside other tests according to their clinical judgment. Physicians who ordered MMBV tests were asked to complete an electronic report at two-time points during the UCC visit (Additional file 1: Fig. S1). First, upon ordering MMBV, physicians reported whether they intended to prescribe antibiotics for this patient. Then, before issuing the discharge note, physicians reported whether MMBV specifically aided in their decision-making process. Physician reports were integrated into the UCC visit management system and were completed in real-time.

Clinicians were not paid to implement the MMBV test, participate in the study, or recruit patients; their involvement was part of their routine clinical work.

The study was approved by the MHS institutional review board (IRB) number MHS-0138–21. Informed consent was waived by the IRB as identifying details of the participants were removed before the analysis.

Objectives and outcome measures

Primary objective

The primary objective was to assess the impact of MMBV on antibiotic use. The appropriateness of antibiotic use was assessed based on alignment between practice and the MMBV result. Given MMBV’s diagnostic accuracy, it can be considered that antibiotics were potentially unwarranted in patients with viral MMBV results (< 35) and were potentially warranted in patients with bacterial MMBV results (> 65).

We evaluated MMBV’s impact on appropriate use of antibiotics using the following outcomes:

  1. 1)

    The proportion of potentially unwarranted antibiotics avoided; defined as cases where physicians intended to prescribe antibiotics, and the patient had a viral MMBV result.

  2. 2)

    The proportion of potentially missed bacterial infections treated; defined as cases where physicians did not intend to prescribe antibiotics, and the patient had a bacterial MMBV result.

  3. 3)

    The proportion of MMBV adherence where physicians reported uncertainty; MMBV adherence was defined as prescribing antibiotics to patients with bacterial MMBV results and not prescribing antibiotics to patients with viral MMBV results.

We also examined whether adherence to MMBV results was associated with increased hospitalization within 7 days of UCC discharge. We hypothesized that the hospitalization rates when physicians adhered to the test would not be greater than when the test was overruled.

Secondary objective

The secondary objective was to assess whether MMBV specifically aided in patient management, according to the UCC physicians’ reports. Upon discharge, physicians reported whether MMBV (a) changed, (b) supported, or (c) did not affect their patient management decision. In addition, physicians reported whether MMBV (a) had no effect on ED referral decisions, (b) caused a referral, or (c) prevented referral.

Exploratory objective

An exploratory objective was to evaluate the added value of MMBV in relation to commonly used biomarkers for bacterial infection, CRP, and white blood cell count (WBC). At the UCCs, CRP and WBC can be ordered at the physician’s discretion and are measured independently of MMBV. Here, we examined three exploratory outcomes:

  1. (1)

    The number of cases where MMBV was equivocal, compared to the number of cases where CRP and WBC were equivocal; equivocal MMBV was defined as 35–65, in line with the manufacturer’s instructions for use [6, 14, 18], equivocal CRP was defined as 20–80 mg/L, and equivocal WBC was defined as 5–15 × 109/L [6, 19].

  2. (2)

    The number of cases where MMBV indicated a bacterial infection (MMBV > 65), but the other available biomarkers were not elevated; elevated CRP was defined as ≥ 80 mg/L and elevated WBC was defined as ≥ 15 × 109/L [19, 20].

  3. (3)

    Multivariable analyses calculating the contribution of MMBV, CRP, and WBC (and other parameters available at presentation) to the decision to avoid an antibiotic prescription. Odds ratios (OR) were used to compare the contribution of each parameter. Further details are given in the “Statistical analysis” section.

Follow-up

For each patient, antibiotic prescriptions, primary care visits, ED visits, and hospitalization records were retrieved for a 7-day period after UCC discharge. The hospitalization length, discharge diagnosis, antibiotic prescription, and microbiological test results were reviewed for patients hospitalized within 7 days.

MMBV implementation

As part of MMBV implementation into routine care, physicians were trained on how to order and interpret MMBV. MMBV scores fall within discrete interpretation bins, according to the manufacturer. MMBV score < 35 indicates viral (or other non-bacterial) infection; a 35 ≤ score ≤ 65 is an equivocal result, and a score > 65 indicates bacterial infection or bacterial-viral co-infection (Additional file 1: Fig. S2). Equivocal results are valid test results that indicate MMBV does not add further information to the physician’s infection etiology diagnosis.

Physician training also included MMBV’s intended use on patients with a suspected acute infectious disease and a clinical dilemma regarding etiology (bacterial/viral) within 7 days of symptom onset, and without certain conditions (e.g., immunosuppression, active oncological disease, hepatitis B, human immunodeficiency virus, hepatitis C carriers). Physicians were advised to avoid ordering MMBV for suspected gastrointestinal infections, tonsillitis, urinary tract infections, and skin infections.

Lab personnel conducted MMBV tests (MeMed BV®, MeMed) at the UCC. The tests were performed on a rapid immunoassay platform (MeMed Key®, MeMed). Results were available within 15 min of serum samples and delivered immediately to the treating physician via the electronic information system.

MMBV tests are single-use, multi-well cartridges. The test requires 100 µl of the patient’s serum to conduct three immunoassays measuring TRAIL, IP-10, and CRP. MMBV computationally integrates the biomarker measurements into a score ranging from 0 to 100 using an algorithm derived previously [9] and validated in multiple studies [6, 10, 12, 21, 22].

Statistical analysis

Descriptive statistics were used to summarize the demographic and clinical characteristics of the study cohort. Continuous variables were expressed as median and interquartile range (IQR), while categorical variables were expressed as frequencies and percentages. For demographic analysis, discharge diagnoses were grouped based on International Classification of Diseases, 9th Revision (ICD9) codes (Additional file 1: Table S1).

Comparison between groups for categorical variables (e.g., sex, antibiotic prescriptions) was conducted using chi-square tests and for continuous variables (e.g., age) using a Wilcoxon rank sum test. The 95%CI was calculated for each proportion using the Agresti-Coull21 method when comparing proportions and odds ratios.

Multivariable analysis relied on a generalized linear mixed model [23,24,25] to predict avoiding antibiotic prescription, where the predictors were age ≥ 65, sex, primary care visit (last 7 days), antibiotics prescribed (previous 7 days), intention to prescribe antibiotics (according to physician report), MMBV viral result (< 35), CRP < 20 mg/L, and WBC < 11 × 109/L. These parameters were selected to represent the measurable defining factors before antibiotic decision-making. The model included a random intercept for each physician to account for the random effect of individual prescription tendencies. ORs were calculated based on the exponent of the model calculated log-odds.

The data were analyzed using R version 4.4.0 and packages “gtsummary” [26, 27]. Patient and physician data was extracted for each UCC encounter where MMBV was ordered. The sample size was determined by the number of tests ordered during the 12-month evaluation period. Participant data was anonymized before analysis. For all statistical tests, a p value of < 0.01 was considered statistically significant.

Results

Study cohort

Between April 2022 and March 2023, 3920 MMBV tests were ordered for adults (age ≥ 18) across ten UCCs by 144 physicians (Additional file 1: Fig. S3, Tables S2, S3). Of these, 162 cases were excluded because MMBV results were missing from the records (Fig. 1). The study cohort included 2228 female patients (59%), had a median age of 42 years (IQR 31–58 years, Table 1), and included 650/3758 (17%) elderly patients (≥ 65 years). The most common discharge diagnoses were fever (1297/3758, 34%) and upper respiratory tract infections (686/3758, 18%). Patient characteristics stratified by discharge diagnosis are provided in Additional file 1: Table S4. In the 7-day follow-up, 252/3758 (6.7%) patients visited the ED, and 237/3758 (6.3%) were hospitalized. In addition, 2061/3758 (55%) patients visited a primary care physician within the 7-day follow-up period, and 532/3758 (14%) had antibiotics prescribed outside of the UCC.

Fig. 1
figure 1

Patient enrolment flow, MMBV results, and whether antibiotics were prescribed at the UCC

Full size image
Table 1 Patient characteristics of the study cohort for the entire cohort and for cases with bacterial, equivocal, or viral MMBV results
Full size table

In the study cohort, 2404/3758 (64.0%) of MMBV tests had a viral result (MMBV < 35), 858/3758 (22.8%) were bacterial (including bacterial-viral co-infection, MMBV > 65), and 496/3758 (13.2%) were equivocal (MMBV 35–65). Overall, antibiotics were prescribed in 1364/3758 (36.3%) of cases, with a higher prescription rate for bacterial cases (628/858, 73.2% prescription) than equivocal (241/496, 48.6%) and viral (495/2404, 20.6%). In a similar trend, ED referrals were 19.3% for bacterial cases, 11.1% for equivocal cases, and 7.2% for viral cases.

Primary objective: MMBV and appropriate use of antibiotics

Among patients with actionable MMBV results, physicians adhered to MMBV results regarding antibiotic prescriptions 2537/3262 (77.8%) of the time. Among elderly patients, adherence was 413/565 (73.1%, Additional file 1: Fig. S4). Before ordering MMBV tests, physicians reported intention to prescribe antibiotics in 928/3262 (28.4%) cases, uncertainty regarding prescribing for 528/3262 (16.2%) cases, and no intention to prescribe for 1806/3262 (55.4%) cases (Fig. 2).

Fig. 2
figure 2

Analysis of the primary objective. For patients with actionable MMBV results (i.e., excluding equivocal results), the physicians’ intention to prescribe antibiotics after the initial evaluation is presented, followed by the MMBV results and prescription practice. (A) Marks cases with potential for unwarranted antibiotics, B cases with diagnostic uncertainty, and C cases of potentially missed bacterial infections

Full size image

Among cases for whom physicians intended to prescribe antibiotics, 629/928 (67.8%) had viral MMBV results. Among those with viral MMBV results, 397/629 (63.1%) were not prescribed antibiotics, avoiding potentially unwarranted antibiotics (Fig. 2A). There were fewer hospitalizations when physicians adhered to MMBV (2.3% [95%CI: 1.1–4.3], compared to 7.3% [95%CI: 4.6–11.5], p = 0.002, Additional file 1: Table S5).

In cases where physicians reported diagnostic uncertainty (Fig. 2B), the rate of adherence to MMBV results was 417/528 (79.0%). For patients with bacterial MMBV results, 115/154 (74.7%) were prescribed antibiotics, and for patients with viral MMBV results, 302/374 (80.7%) were not prescribed antibiotics.

Among cases for whom physicians did not intend to prescribe antibiotics, 405/1806 (22.4%) had bacterial MMBV results, representing potentially missed bacterial infections (Fig. 2C). Among these patients, 283/405 (69.9%) were prescribed antibiotics. There were significantly fewer hospitalizations when physicians adhered to MMBV (i.e., prescribed antibiotics), compared to when MMBV was overruled (7.8% [95%CI: 5.1–11.5] vs. 30.3% [95%CI: 22.9–39.0], p < 0.001, Additional file 1: Table S6). Only 7/59 (11.9%) of the hospitalized patients had clinically relevant microbiological detection (Additional file 1: Tables S7–S8).

Secondary objective: MMBV and patient management decisions according to UCC physician reports

Physicians reported that MMBV results changed their patient management in 607/2901 (20.9%, 95%CI: 19.5–22.4) cases, supported the decision-making process in 1887/2901 (65.0%, 95%CI: 63.3–66.8), and did not affect patient management in 407/2901 (14.0%, 95%CI: 12.8–15.3). Physicians’ reports were consistent across different patient age groups and diagnoses (Fig. 3).

Fig. 3
figure 3

MMBV influence on patient management decisions, based on physician’s self-report at the end of UCC visit. Answers to the question, “How did the MeMed BV result affect your decision regarding this patient’s management?”. The top panel displays the breakdown based on the patient’s age. The bottom panel is stratified according to the discharge diagnosis. Only diagnoses with > 100 results are shown

Full size image

In cases where physicians reported that MMBV affected patient management, antibiotics were prescribed for 80.3% (512/638, 95%CI: 77.0–83.2) of bacterial cases and 16.2% (300/1856, 95%CI: 14.6–17.9) of viral ones (Fig. 4). On the other hand, where physicians reported that MMBV did not affect management, antibiotics were prescribed to 43.1% (44/102, 95%CI: 33.9–52.8) and 43.6% (133/305, 95%CI: 38.2–49.2) for MMBV bacterial and viral cases, respectively.

Fig. 4
figure 4

Antibiotic prescription in cases for whom physicians indicated MMBV affected (left)/not affected (right) patient management. Results were shown for bacterial and viral MMBV results. Asterisk denotes p < 0.001; n.s. denotes “not significant”

Full size image

Physicians reported that MMBV results had no effect on their ED referral decision in 2201/2901 (75.9%) cases, contributed to the decision to refer in 105/2901 (3.6%), and avoided referral in 595/2901 (20.5%). Of note, among the patients where the referral was avoided, in 7-day follow-up, 574/595 (96.5%) did not have an ED visit.

Exploratory objective: the added value of MMBV in relation to commonly used biomarkers for bacterial infection (CRP and WBC)

MMBV, CRP, and WBC were ordered for 2867 cases, and only these cases were used in the exploratory analysis.

The proportion of equivocal MMBV results (MMBV between 35 and 65) was 364/2867 (12.7%), equivocal CRP results (between 2 and 8 mg/L) was 943/2867 (32.9%), and equivocal WBC (between 5 and 15 × 109/L) was 2256/2867 (78.7%). There were 581/2867 (20.3%) cases where both CRP and WBC were equivocal, and MMBV provided actionable results, and 158/2867 (5.5%) cases where all three tests were equivocal (Additional file 1: Fig. S5).

Among cases with bacterial MMBV results, elevated WBC (≥ 15 × 109/L) and CRP (≥ 80 mg/L) were observed in 129/684 (18.9%) and 498/684 (72.8%) of cases, respectively.

In multivariable analysis, the following parameters had a significant contribution towards avoiding antibiotic prescription (i.e., OR > 1, Fig. 5): viral MMBV (OR 6.263), CRP < 20 mg/L (2.35), WBC < 11 × 109/L (OR 1.919), and antibiotics prescribed in the past 7 day (OR 1.82). Intention to prescribe antibiotics (based on physician reports) contributed against avoiding prescription (OR 0.379).

Fig. 5
figure 5

Model of drivers for avoiding antibiotic prescription. The results are presented as the odds ratio with the 95%CI

Full size image

Discussion

Implementing MMBV aided physicians in patient management decisions when we examined how their practice aligns with the test results (primary objective) and according to the physicians themselves (secondary objective). Physicians avoided 63% of potentially unwarranted antibiotic prescriptions, treated 70% of potentially missed bacterial infections as suggested by MMBV results, and attested that MMBV aided their patient management in 86% of all cases.

The observed reduction in unwarranted antibiotic use aligns with our pilot study [18]. Previously, we observed a 40.9% reduction in potentially unwarranted antibiotic use without an increase in the rate of hospitalizations. We also noted that adherence to MMBV results was not associated with increased hospitalization. In fact, prescribing antibiotics to cases defined as potentially missed bacterial infections was associated with fewer hospitalizations within 7-day follow-up. These findings support the safety of MMBV utilization at UCCs.

While it is difficult to determine whether antibiotics are unwarranted without microbiological and radiological testing (that was not part of the UCC visit), we used viral MMBV results (< 35) as proxies for cases that do not warrant antibiotics. This underlying assumption was based on data from previous studies indicating the test can rule out a bacterial infection with a negative predictive value > 98% [6, 14]. This was further corroborated by the fact that there was no increase in hospitalization within 7 days when antibiotics were avoided—suggesting the untreated patients did not significantly deteriorate.

Notably, a small proportion of patients (260/3262, 8.0%) for whom the physician’s original intentions aligned with MMBV, yet practice did not follow suit. These included cases with bacterial and viral MMBV results that may be due to underlying conditions, conflicting ancillary test results, or patient pressure (to either obtain or refrain from antibiotics) [28, 29]. Also, some patients were prescribed antibiotics before presenting to the UCC, and some were referred to the ED for treatment. Future qualitative studies are warranted to understand in greater depth the decision-making process.

Physicians indicated that MMBV supported their decision-making regarding ED referrals. Referral to ED is often prompted by the need for escalated care or further investigation using advanced tools, as most UCCs only offer limited on-site diagnostics [30,31,32]. As MMBV provides insights about the etiology of infection, we postulate that reduced diagnostic uncertainty is a possible explanation for the avoided referrals. Another explanation could be that patients with viral MMBV results (77% of avoided referrals, Additional file 1: Table S9) may have been considered by the physician as having a self-limiting and milder illness and be safe to discharge home.

We observed that in the small group of cases when physicians reported MMBV did not affect management, 44% of cases with viral MMBV results were prescribed antibiotics. These rates are similar to the rate of unwarranted prescriptions in the UCC setting in the USA (18–46% [3, 4, 7]). Prescribing antibiotics despite a viral MMBV result may be due to physicians’ lack of familiarity with the test, their own historic prescribing patterns, or possibly because they were concerned about patient satisfaction.

MMBV was integrated into routine care and ordered at the physician’s discretion alongside other tests, including CRP and WBC. Studies show the variable performance of CRP and WBC across different populations and settings [33,34,35,36,37]. Furthermore, MMBV was shown previously to outperform both biomarkers in diagnostic accuracy studies [6, 38]. We found that MMBV provides added value to these other commonly used biomarkers across all exploratory outcomes, showing how MMBV is informative while other tests are equivocal and that in multivariate analysis, it was the largest contributor to the decision to prescribe antibiotics. These results must be interpreted cautiously. It is not expected that MMBV will replace existing tools but rather complement them to improve management.

Limitations

The study was not designed as a randomized trial, limiting our ability to infer causality between MMBV results and practice. Also, no “true” infection etiology was identified, as such a process would require extensive testing that does not fit the pragmatic nature of the study (such as additional blood cultures, PCRs, or X-rays that were not ordered by the treating physician).

Only patients for whom physicians ordered MMBV were included, not representing all patients with suspected infectious diseases at UCCs. While this pragmatic approach focused on a realistic target population without disrupting routine care, it may have introduced selection bias towards early adopter physicians. Furthermore, the electronic report required to be completed at two points during the UCC visit is a form of intervention and may have impacted antibiotic decisions and adherence to test results. The absence of more detailed follow-up and microbiological investigation prevents ruling out potential complications in cases where antibiotics were avoided.

Conclusions

Implementing MMBV aided urgent care center physicians in their clinical decision-making and may have contributed to appropriate antibiotic use, better resource utilization, and improved patient management.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

UCCs:

Urgent care centers

MMBV:

MeMed BV®

ED:

Emergency department

TRAIL:

Tumor necrosis factor-related apoptosis-inducing ligand

IP-10:

Interferon-gamma induced protein-10

CRP:

C-reactive protein

CI:

Confidence interval

MHS:

Maccabi Healthcare Services

WBC:

White blood cell count

IQR:

Interquartile range

ICD9:

International Classification of Diseases, 9th Revision

References

  1. Ganguli I, Shi Z, Orav EJ, Rao A, Ray KN, Mehrotra A. Declining use of primary care among commercially insured adults in the United States, 2008–2016. Ann Intern Med. 2020;172(4):240–7.

    Article  PubMed  Google Scholar 

  2. Carlson LC, Raja AS, Dworkis DA, Lee J, Brown DFM, Samuels-Kalow M, et al. Impact of urgent care openings on emergency department visits to two academic medical centers within an integrated health care system. Ann Emerg Med. 2020;75(3):382–91.

    Article  PubMed  Google Scholar 

  3. Stenehjem E, Wallin A, Willis P, Kumar N, Seibert AM, Buckel WR, et al. Implementation of an antibiotic stewardship initiative in a large urgent care network. JAMA Netw Open. 2023;6(5):e2313011.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Stenehjem E, Wallin A, Fleming-Dutra KE, Buckel WR, Stanfield V, Brunisholz KD, et al. Antibiotic prescribing variability in a large urgent care network: a new target for outpatient stewardship. Clin Infect Dis. 2020;70(8):1781–7.

    Article  PubMed  Google Scholar 

  5. Poon SJ, Schuur JD, Mehrotra A. Trends in visits to acute care venues for treatment of low-acuity conditions in the United States from 2008 to 2015. JAMA Intern Med. 2018;178(10):1342–9.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Halabi S, Shiber S, Paz M, Gottlieb TM, Barash E, Navon R, et al. Host test based on TRAIL, IP-10 and CRP for differentiating bacterial and viral respiratory tract infections in adults: diagnostic accuracy study. Clin Microbiol Infect. 2023;S1198–743X(23)00257–4.

  7. Palms DL, Hicks LA, Bartoces M, Hersh AL, Zetts R, Hyun DY, et al. Comparison of antibiotic prescribing in retail clinics, urgent care centers, emergency departments, and traditional ambulatory care settings in the United States. JAMA Intern Med. 2018;178(9):1267–9.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Hersh AL, King LM, Shapiro DJ, Hicks LA, Fleming-Dutra KE. Unnecessary antibiotic prescribing in US ambulatory care settings, 2010–2015. Clin Infect Dis. 2021;72(1):133–7.

    Article  PubMed  Google Scholar 

  9. Oved K, Cohen A, Boico O, Navon R, Friedman T, Etshtein L, et al. A novel host-proteome signature for distinguishing between acute bacterial and viral infections. PLoS ONE. 2015;10(3):e0120012.

    Article  PubMed  PubMed Central  Google Scholar 

  10. van Houten CB, de Groot JAH, Klein A, Srugo I, Chistyakov I, de Waal W, et al. A host-protein based assay to differentiate between bacterial and viral infections in preschool children (OPPORTUNITY): a double-blind, multicentre, validation study. Lancet Infect Dis. 2017;17(4):431–40.

    Article  PubMed  Google Scholar 

  11. Srugo I, Klein A, Stein M, Golan-Shany O, Kerem N, Chistyakov I, et al. Validation of a novel assay to distinguish bacterial and viral infections. Pediatrics. 2017;140(4):e20163453.

    Article  PubMed  Google Scholar 

  12. Papan C, Argentiero A, Porwoll M, Hakim U, Farinelli E, Testa I, et al. A host signature based on TRAIL, IP-10, and CRP for reducing antibiotic overuse in children by differentiating bacterial from viral infections: a prospective, multicentre cohort study. Clin Microbiol Infect. 2022;28(5):723–30.

    Article  CAS  PubMed  Google Scholar 

  13. Mor M, Paz M, Amir L, Levy I, Scheuerman O, Livni G, et al. Bacterial vs viral etiology of fever: a prospective study of a host score for supporting etiologic accuracy of emergency department physicians. PLoS ONE. 2023;18(1):e0281018.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Bachur RG, Kaplan SL, Arias CA, Ballard N, Carroll KC, Cruz AT, et al. A rapid host–protein test for differentiating bacterial from viral infection: Apollo diagnostic accuracy study. J Am Coll Emerg Physicians Open. 2024;5(3):e13167.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Patel R, Tsalik EL, Evans S, Fowler VG, Doernberg SB, Antibacterial Resistance Leadership Group. Clinically adjudicated reference standards for evaluation of infectious diseases diagnostics. Clin Infect Dis. 2023;76:938–43.

  16. van Houten CB, Naaktgeboren CA, Ashkenazi-Hoffnung L, Ashkenazi S, Avis W, Chistyakov I, et al. Expert panel diagnosis demonstrated high reproducibility as reference standard in infectious diseases. J Clin Epidemiol. 2019;112:20–7.

    Article  PubMed  Google Scholar 

  17. Whitfield NN, Hogan CA, Chenoweth J, Hansen J, Hsu EB, Humphries R, et al. A standardized protocol using clinical adjudication to define true infection status in patients presenting to the emergency department with suspected infections and/or sepsis. Diagn Microbiol Infect Dis. 2024;110(1): 116382.

    Article  CAS  PubMed  Google Scholar 

  18. Kalmovich B, Rahamim-Cohen D, Shapiro Ben David S. Impact on patient management of a novel host response test for distinguishing bacterial and viral infections real world evidence from the urgent care setting . Biomedicines. 2023;11(5):1498.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Haran JP, Beaudoin FL, Suner S, Lu S. C-reactive protein as predictor of bacterial infection among patients with an influenza-like illness. Am J Emerg Med. 2013;31(1):137–44.

    Article  PubMed  Google Scholar 

  20. Escadafal C, Incardona S, Fernandez-Carballo BL, Dittrich S. The good and the bad: using C reactive protein to distinguish bacterial from non-bacterial infection among febrile patients in low-resource settings. BMJ Glob Health. 2020;5(5):e002396.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Klein A, Shapira M, Lipman-Arens S, Bamberger E, Srugo I, Chistyakov I, et al. Diagnostic accuracy of a real-time host-protein test for infection. Pediatrics. 2023;152(6):e2022060441.

    Article  PubMed  Google Scholar 

  22. Chokkalla AK, Tam E, Liang R, Cruz AT, Devaraj S. Validation of a multi-analyte immunoassay for distinguishing bacterial vs. viral infections in a pediatric cohort. Clin Chim Acta. 2023;546:117387.

    Article  CAS  PubMed  Google Scholar 

  23. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;7(67):1–48.

    Google Scholar 

  24. Luke SG. Evaluating significance in linear mixed-effects models in R. Behav Res Methods. 2017;49(4):1494–502.

    Article  PubMed  Google Scholar 

  25. Random effects in generalized linear models. In: Generalized linear models. Ltd: John Wiley & Sons; 2010. p. 319–407.

    Google Scholar 

  26. Sjoberg DD, Whiting K, Curry M, Lavery JA, Larmarange J. Reproducible summary tables with the gtsummary package. R J. 2021;13(1):570–80.

    Article  Google Scholar 

  27. Kassambara A. rstatix: pipe-friendly framework for basic statistical tests. 2023 [cited 2023 Nov 6]. Available from: https://cran.r-project.org/web/packages/rstatix/index.html.

  28. Scott JG, Cohen D, DiCicco-Bloom B, Orzano AJ, Jaen CR, Crabtree BF. Antibiotic use in acute respiratory infections and the ways patients pressure physicians for a prescription. J Fam Pract. 2001;50(10):853–8.

    CAS  PubMed  Google Scholar 

  29. Bagnulo A, Muñoz Sastre MT, Kpanake L, Sorum PC, Mullet E. Why patients want to take or refuse to take antibiotics: an inventory of motives. BMC Public Health. 2019;19(1):441.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Hart JH, Sakata T, Eve JR, Butler AM, Wallin A, Carman C, et al. Diagnosis and treatment of pneumonia in urgent care clinics: opportunities for improving care. Open Forum Infect Dis. 2024;11(3):ofae096.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Gorodetzer R, Alpert EA, Orr Z, Unger S, Zalut T. Lessons learned from an evaluation of referrals to the emergency department. Isr J Health Policy Res. 2020;9(1):18.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Siegfried I, Jacobs J, Olympia RP. Adult emergency department referrals from urgent care centers. Am J Emerg Med. 2019;37(10):1949–54.

    Article  PubMed  Google Scholar 

  33. Leticia Fernandez-Carballo B, Escadafal C, MacLean E, Kapasi AJ, Dittrich S. Distinguishing bacterial versus non-bacterial causes of febrile illness - a systematic review of host biomarkers. J Infect. 2021;82(4):1–10.

    Article  CAS  PubMed  Google Scholar 

  34. Roger PM, Keïta-Perse O, Mainardi JL. Diagnostic uncertainty in infectious diseases: advocacy for a nosological framework. Infect Dis Now. 2023;53(6):104751.

    Article  PubMed  Google Scholar 

  35. Gbinigie OA, Onakpoya IJ, Richards GC, Spencer EA, Koshiaris C, Bobrovitz N, et al. Biomarkers for diagnosing serious bacterial infections in older outpatients: a systematic review. BMC Geriatr. 2019;19(1):190.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Hill AT, Gold PM, Solh AAE, Metlay JP, Ireland B, Irwin RS, et al. Adult outpatients with acute cough due to suspected pneumonia or influenza: CHEST guideline and expert panel report. Chest. 2019;155(1):155–67.

    Article  PubMed  Google Scholar 

  37. Smedemark SA, Aabenhus R, Llor C, Fournaise A, Olsen O, Jørgensen KJ. Biomarkers as point-of-care tests to guide prescription of antibiotics in people with acute respiratory infections in primary care. Cochrane Database Syst Rev. 2022;10(10):CD010130.

    PubMed  Google Scholar 

  38. Ashkenazi-Hoffnung L, Oved K, Navon R, Friedman T, Boico O, Paz M, et al. A host-protein signature is superior to other biomarkers for differentiating between bacterial and viral disease in patients with respiratory infection and fever without source: a prospective observational study. Eur J Clin Microbiol Infect Dis. 2018;37(7):1361–71.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank our colleagues Tanya Gottlieb PhD, Amir Nakar PhD, Eran Reiner, MD, Roy Navon, and Boris Lebedenko for their input and technical support.

Funding

The study was conducted without external funding. Data collection, analysis, and manuscript preparation were completed using internal resources and as part of the authors’ routine professional responsibilities. None of the authors received additional payment or financial support related to the study and declare no competing interest.

Author information

Authors and Affiliations

  1. Maccabi Healthcare Services, 27 Hamered Street, Tel Aviv, 6812509, Israel

    Boaz Kalmovich, Daniella Rahamim-Cohen, Ilan Yehoshua, Sara Kivity, Noam Orvieto & Shirley Shapiro Ben David

  2. Faculty of Medicine, Tel Aviv University, Tel Aviv, 6997801, Israel

    Daniella Rahamim-Cohen, Ilan Yehoshua, Noam Orvieto & Shirley Shapiro Ben David

Authors
  1. Boaz Kalmovich
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Daniella Rahamim-Cohen
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Ilan Yehoshua
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Sara Kivity
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Noam Orvieto
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Shirley Shapiro Ben David
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Conceptualization, S.S.B.D and B.K; Data extraction: I.Y, S.K, D.R.C, methodology, S.S.B.D; formal analysis, S.S.B.D; writing—original draft preparation, S.S.B.D; writing—review and editing, B.K and D.R.C; All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Shirley Shapiro Ben David.

Ethics declarations

Ethics approval and consent to participate

The study was approved by the MHS institutional review board (IRB) number MHS-0138–21. Informed consent was waived by the IRB as identifying details of the participants were removed before the analysis.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests,

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

12916_2025_3903_MOESM1_ESM.docx

Additional file 1: All supplemental materials, Figures S1–S5 and Tables S1–S9. Fig. S1. Questionnaire presented to physicians in the visit management software. Fig. S2. MMBV test results and interpretation bins according to the manufacturer’s instructions for use. Table S1. Discharge diagnosis categories by ICD-9 codes. Fig. S3. Number of MMBV tests ordered by UCC and study week. Table S2. Participating physicians. Table S3. Tests ordered in addition to MMBV in the study cohort. Fig. S4. Study flow for elderly patients. Fig. S5. MMBV scores compared to CRP and WBC in cases where all three tests were ordered. Table S4. Patient characteristics stratified by discharge diagnosis. Table S5. Primary objective: potential unwarranted antibiotic use. Table S6. Primary objective: potentially missed bacterial infections. Table S7. Potentially missed bacterial infections: outcomes of patients that were hospitalized within 7 days of the UCC visit. Table S8. Potentially missed bacterial infections with clinically relevant microbiological detection at hospital. Table S9. Responses to the question “How did MeMed BV result affect ED referral,” stratified by MMBV result.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kalmovich, B., Rahamim-Cohen, D., Yehoshua, I. et al. Implementation of a rapid host-protein diagnostic test for distinguishing bacterial and viral infections in adults presenting to urgent care centers: a pragmatic cohort study. BMC Med 23, 63 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12916-025-03903-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12916-025-03903-8

Keywords

BMC Medicine

ISSN: 1741-7015

Contact us