In the last 10 years, multiplex polymerase chain reaction (PCR) technology has changed the basic diagnostic approach to respiratory infections.Reference Ramanan, Bryson and Binnicker1 Because multiplex PCR offers advantages over traditional culture-based techniques, including higher sensitivity and a shorter turnaround time, use of multiplex PCR testing for respiratory infections has become increasingly common at US medical centers. In some instances, targeted tests for viral pathogens, such as influenza, have been replaced by bundled tests for influenza, severe acute respiratory coronavirus virus 2 (SARS-CoV-2), and respiratory syncytial virus. In other instances, broad respiratory panels targeting 20 or more pathogens, potentially including a combination of viruses and bacteria, are used. During the early coronavirus disease 2019 (COVID-19) pandemic period, adoption of many of these respiratory panels accelerated as medical centers sought reliable access to rapid COVID-19 testing.Reference Klompas, Imrey and Yu2
However, despite widespread clinical use, the optimal implementation of respiratory panels to support diagnosis, patient care, and clinical outcomes remains undefined. Fewer than 1 in 5 academic institutions have implemented guidelines to support effective testing practices.Reference Kondo, Simon and Westblade3 In clinical trials, respiratory panel testing often fails to influence patient care, either because the treatment plan has been decided before testing or because antiviral therapy is not commonly prescribed for the virus detected.Reference Klompas, Imrey and Yu2,Reference Baghdadi, Yang and Lynen4–Reference May, Tatro and Poltavskiy7 The goals of this study were (1) to describe best practices in the implementation of respiratory panel testing based on expert perspectives and (2) to identify tools for diagnostic stewardship that might enhance the usefulness of testing.
Methods
The Society for Healthcare Epidemiology of America Research Network (SRN) is a consortium of hospitals conducting collaborative research projects in healthcare epidemiology and antibiotic stewardship. Qualtrics surveys were distributed by email to SRN member institutions between October 2021 and February 2022. Only 1 completed survey per SRN member institution was permitted.
The survey questions covered in this study are included in the Supplementary Material (online). Survey questions were multiple choice, multiple-selection check boxes, and free text. In the survey, respiratory panels were defined as “Rapid molecular diagnostics that perform PCR tests for >5 pathogenic targets simultaneously. Pathogenic targets may include resistance genes or genes related to toxin production. Examples of panels include Biofire, Verigene, ePlex.” This study was determined to be exempt from approval by the University of Maryland School of Medicine Institutional Review Board.
Data were exported from Qualtrics into Stata/IC version 14 software (StataCorp, College Station, TX) for descriptive analyses. Responses to multiple-choice questions were represented as proportions. Directed content analysis of free-text responses was conducted in NVivo 11 (QSR International, Burlington, MA) by 2 independent reviewers (J.D.B. and L.M.O.) using a predetermined codebook.Reference Hsieh and Shannon8 The pooled average percentage agreement between reviewers was 95.5% (κ = 0.58, calculated using Excel 16.0, Microsoft, Redmond, WA). Themes were summarized based on frequency within each question prompt. Where presented, quotations have been edited for punctuation and grammar.
Results
Surveys were sent to 82 study sites. Completed surveys were returned by 41 sites for a response rate of 50%. Sites were primarily located in the United States (n = 30, 73%), including 19 states and the District of Columbia. International sites were in Asia (n = 4, 9%), Europe (n = 2, 5%), North America (n = 2, 5%), and other (n = 3, 7%). Sites included 26 academic medical centers (63%), 7 community hospitals (17%), 2 federal hospitals (5%), and 6 other types of sites (15%). Also, 5 (12%) sites provided pediatric care, including 1 freestanding children’s hospital.
Respondents’ self-reported primary roles were hospital epidemiologist (n = 26, 63%), infectious disease consultant (n = 6, 15%), infection preventionist (n = 2, 5%), antibiotic stewardship physician (n = 3, 7%), infectious diseases pharmacist (n = 1, 2%), clinical microbiologist (n = 1, 2%), and dual clinical microbiologist–infection preventionist (n = 1, 2%). All respondents reported access to multiplex respiratory panel testing, including on-site access at 35 sites (85%).
Impact of respiratory panel testing
A joint display summarizing multiple-choice responses alongside excerpts of free text related to the clinical impact of respiratory panel testing is presented in Table 1. Respiratory panel testing was perceived as supporting or expediting accurate diagnosis at 35 sites (85%), supporting more efficient patient care at 33 sites (80%), and improving patient outcomes at 23 sites (56%). Among the 36 respondents who indicated that respiratory panel testing may either support accurate diagnosis or patient care, 23 (64%) believed that testing improves patient outcomes.
Table 1. Perceived Impact of Respiratory Panel Testing
Note. HE, hospital epidemiologist; IDP, infectious diseases pharmacist; IDC, infectious diseases consultant. Perceived impact was determined by response to multiple choice questions related to impact of multiplex PCR respiratory panel testing on accurate diagnosis, efficient patient care, and improved clinical outcomes. Related themes and representative quotations are drawn from qualitative analysis of free-text responses.
Free-text responses emphasized that the relationship between testing and clinical outcomes is less certain than the relationships between testing and diagnosis and testing and patient care:
-
Infectious disease consultant 1: “[Respiratory panels] help with isolation and diagnosis. Some respiratory pathogens are managed with supportive care and probably do not affect patient outcomes.”
-
Infectious disease consultant 4: “In general, they tend to help with more accurate diagnosis, however [it] is actually a mixed bag on patient outcomes … depending upon timing of ordering and who is actually in charge of care of the patient.”
-
Hospital epidemiologist 13: “I don’t have any data about their impact on patient outcomes, but I strongly suspect that having an accurate diagnosis improves outcomes.”
Perceived impact of respiratory panel testing based on respondent characteristics
When grouping respondents based on primary role, infectious diseases consultants were the most likely to perceive benefits from testing. All 6 infectious disease consultants reported that testing improves diagnosis and patient care, whereas 5 of 6 perceived testing as improving patient outcomes. On the other hand, antibiotic stewardship physicians were the least likely to perceive benefits from testing. Only 1 of 3 antibiotic stewardship physicians reported that testing improves diagnosis, 2 of 3 reported that testing improves patient care, and 1 of 3 reported that testing improves patient outcomes.
Compared to those with access to respiratory panel testing through outside sources (n = 6), respondents with access to onsite testing (n = 35) more frequently reported perceptions that testing supports diagnosis (89% vs 67%; P = .21 using the Fisher exact test), supports patient care (86% vs 50%; P = .034), and improves clinical outcomes (60% vs 33%; P = .15).
-
Infectious disease pharmacist #1: “Aside from influenza/COVID-19, we do not have access to an on-site respiratory multiplex PCR. When it’s sent, turnaround time is days and does not tend to impact [the] antimicrobial prescribed and is more used in immunocompromised hosts.”
Patterns of use in clinical areas
In the current state, respiratory panels were perceived as being used most frequently in intensive care (n = 38, 93% of respondents), inpatient acute care (n = 34, 83%), and the emergency department (n = 29, 71%). Respiratory panel testing was only perceived as being used frequently in the outpatient or clinic setting at 10 sites (24%). Respiratory panel testing was perceived as being useful in the same clinical areas: intensive care (n = 33, 80%), inpatient acute care (n=28, 68%), and emergency department (n = 25, 61%). Only 5 respondents (12%) perceived respiratory panels as being useful in outpatient care.
Overall, 25 respondents (61%) reported a mismatch between where respiratory panels are frequently used at their site and where testing is the most useful. Among those reporting mismatches, 16 (39% of the total sample) perceived respiratory panels as being used in more clinical areas than where they can be useful (eg, respiratory panels are most useful in intensive care only but are used most commonly in intensive care and the emergency department), and 4 (10%) perceived respiratory panels as being used in fewer areas than where they can be useful. Also, 5 respondents (12%) perceived respiratory panels as being used in the right number of clinical areas, albeit in the wrong areas.
Existing diagnostic stewardship
Among the 24 (58%) sites that implemented diagnostic stewardship to improve use of respiratory panel testing, the median number of interventions was 3 (IQR, 1–4). Interventions included education (n = 22, 54% of sites), structured order sets in the electronic health record with a “soft stop” that requires the clinician-user to click through to proceed with test ordering (n = 11, 27%), edited results reporting to facilitate interpretation (n = 8, 20%), best practice alerts (n = 6, 15%), communication of results through an intermediary such as the antimicrobial stewardship team or an infectious diseases consultant (n = 5, 12%), structured order sets with a “hard stop” that blocks test ordering without the option for clinician override (n = 2, 5%), and other (n = 3, 7%). Examples of “soft stops” included order sets that guide clinicians toward respiratory panel testing rather than COVID-19 testing and alerts to discourage repeat testing. Examples of hard stops included order sets that restrict testing to transplant patients or require either an infectious disease physician or pulmonologist to place the test order.
Across sites, the interventions most frequently reported as being effective were structured order sets built into the electronic health record to guide test ordering (n = 4, 10%), order restriction (n = 3, 7%), and structured communication of results to the primary team (n = 2, 5%).
-
Clinical microbiologist–infection preventionist 1: “Specific communication with the clinical team on interpretation of the test results has been very useful.”
Education was reported as useful at 4 sites (10%) but was often described in terms of its limitations (n = 3, 75% of those reporting benefit from education):
-
Hospital epidemiologist 15: “Education has reduced use, but this had led in some areas to a sense of total lack of value. Education missed when it IS valuable and focused on when it is NOT (all the time).”
Also, 1 site reported tracking variation in testing by provider and sharing these data with each department:
-
Hospital epidemiologist 10: “I look at the total number of orders per physician. Some order 25/year, and some order 250/year. I share the top 20 with the relevant departments. Just letting providers know we are counting, makes a huge difference.”
Target populations
Based on free-text responses, patient populations perceived as most likely to benefit from respiratory panel testing were those who are critically ill (9 mentions), immunocompromised (9 mentions), or otherwise requiring hospitalization (8 mentions). Conversely, respiratory panel testing was frequently perceived as having low utility among outpatients or emergency department discharges (8 mentions). Furthermore, 4 respondents suggested that the utility of respiratory panel testing depends on the pathogens under consideration, the differential diagnosis, or the need for further work-up:
-
Infection preventionist 1: “For patients with upper respiratory infections… It is not for patients with suspected bacterial pneumonia.”
-
Hospital epidemiologist 10: “If it doesn’t impact antibiotics or care, then it isn’t needed. Anyone in the hospital more than 5 days, doesn’t need a Biofire.”
Recommendations for optimal use
When considering how to improve testing practices, respondents emphasized the goals of testing rather than specific approaches (Table 2). Perspectives frequently focused on testing to facilitate antimicrobial stewardship (8 mentions), testing based on clinical syndrome (8 mentions), and testing to support infection control (5 mentions). Recommendations for specific interventions included developing protocols to guide use (2 mentions), restricting repeat testing (2 mentions), restricting testing to subspecialty use (1 mention), and grouping pathogens in the results report based on relevance (1 mention).
-
Antibiotic stewardship physician 1: “Rapid diagnostics help with [the] decision to use antibiotics or not.”
-
Hospital epidemiologist 20: “A well-defined pathway outlining clinical indications for testing and monitoring the impact [is needed] before introducing these rapid diagnostic panels.”
-
Hospital epidemiologist 8: “Non-ID physicians tend to order many more tests (serial tests for example). To have more rational use, infectious diseases authorization is needed.”
Table 2. Recommendations for Optimal Use
Note. HE, hospital epidemiologist; IP, infection preventionist; IDP, infectious diseases pharmacist; IDC, infectious diseases consultant.
Barriers to optimal use
The cost of testing (2 mentions), requirements to use respiratory panel testing related to COVID-19 (2 mentions), and access to test kits (1 mention) were cited as barriers:
-
Hospital epidemiologist 5: “The need to test for COVID and flu makes it hard to make optimal use of respiratory panels. We also rely on them to help assess appropriate isolation. Isolation orders are bundled with respiratory tests, and we track compliance with isolation status by test being sent. It is too complicated to audit isolation status by presence of respiratory symptoms alone.”
-
Hospital epidemiologist 21: “If they were cheaper, we’d use them more.”
Discussion
In this survey conducted via the SRN, multiplex PCR respiratory panels were perceived by experts in infectious diseases and hospital epidemiology as providing value in diagnosis and patient care. However, for one-third of respondents, value in these domains did not translate into a perceived positive impact on clinical outcomes. This gap between the recognized advantages of respiratory panel testing and potential lack of patient benefit represents an opportunity for diagnostic stewardship.
The questions on our survey sought to disentangle the impact of testing on diagnosis, patient care, and health outcomes. Based on the responses, we suspect the connections between these entities vary based on both the pathogen and the clinical syndrome. For instance, when a patient presents with upper respiratory infection due to influenza, diagnosis by respiratory panel testing can lead to targeted antiviral treatment that will shorten the duration of illness. On the other hand, in the setting of severe illness, a positive test indicating viral infection may inform use of isolation or other aspects of care without changing use of antibiotics, need for intubation, or admission to the intensive care unit. In the setting of mild illness, a positive test for rhinovirus may diagnose the etiology of infection without affecting care choices or the course of self-limited illness.
The COVID-19 pandemic forced many medical facilities to closely examine the role of multiplex respiratory panel testing in acute care. We suspect these experiences informed the responses to our survey, which were frequently framed around COVID-19. However, these experiences also likely enhanced the quality of responses, since the pandemic required many medical facilities to prioritize distribution of testing based on which patients were most likely to benefit from it. Notably, our respondents commonly expressed equivalence between SARS-CoV-2 and influenza in the sense that both are clinically significant viral pathogens for which antiviral therapy is available. Moving forward, we suspect that clinicians will likely continue to interpret test results as reflecting treatable versus nontreatable respiratory viruses.
About half of the sites responding to our survey had implemented multiple diagnostic stewardship interventions to enhance the usefulness of respiratory panel testing. Among these, the interventions most frequently considered effective targeted test ordering and results reporting. However, we did not elicit detailed descriptions of each intervention, and interventions designed to target the same steps in the diagnostic process may differ substantially in the extent to which they disrupt clinical care processes or require active maintenance. The ideal intervention for a given institution likely depends on the availability of staff to support and implement it, the accessibility and flexibility of clinical informatics, and the tolerance of frontline clinicians to changes to their workflow. For instance, requiring infectious diseases to provide preclearance for testing may be an efficient means of limiting overtesting at one institution and unacceptably disruptive at another.
Notably, educational interventions were implemented frequently but were perceived as limited. Multiple reasons were given regarding why an educational intervention for quality improvement may not have the desired impact: the underlying quality gap is not attributable to knowledge deficits, frequent repetition is required to maintain the effect, or other barriers challenge the application of gained knowledge.Reference Soong and Shojania9 In diagnostic stewardship, education may be considered the equivalent of a preintervention that enhances institutional readiness, rather than a true intervention on its own.Reference Hueth, Prinzi and Timbrook10 In particular, educating local champions and experts can facilitate dissemination of a new or innovative care process across an institution. In the absence of education, on the other hand, poor understanding of a new care process can act as a barrier to its uptake.
A theme expressed by multiple experts was preferred use of testing among immunocompromised or critically ill patients. This approach matches a conditional recommendation from American Thoracic Society guidelines based on the expectation that testing in these populations is more likely to inform treatment.Reference Evans, Jennerich and Azar11 However, antibiotic stewardship in the intensive care unit can be challenging, and clinicians may be more hesitant to de-escalate antibiotics among critically ill patients.Reference Wunderink, Srinivasan and Barie12 Scenarios in which a given patient’s risk factors and the detected pathogen align to create an opportunity for treatment, such as when parainfluenza-3 is diagnosed in a hematopoietic stem cell transplant recipient, are rare. We suspect that the frequency of instances in which respiratory panel testing influences patient care or outcomes among the highest risk patients is likely very low. Instead, the preference expressed by our respondents for testing in complicated patient populations may reflect the typical clinical experience of an infectious disease specialist, which tends to involve these types of patients. Real-world evidence is needed to demonstrate when and how respiratory panel testing can provide value among patients who are critically ill or immunocompromised.
In relation to use of isolation, no experts referenced the patient experience as a reason for testing. Patient satisfaction at reaching a diagnosis has been used as a justification for respiratory panel testing even when the results will not inform patient care.Reference Schreckenberger and McAdam13 Our results suggest that this argument does not reflect clinical reality.
This study had several limitations, including those inherent to a survey. Although our respondents were from multiple sites across the United States and abroad, their responses should be considered a collection of expert opinions, rather than representative of a larger population. Bias may have been introduced based on which institutions self-select to participate in the SRN and which individuals respond on behalf of those institutions. Further research, including input from frontline clinicians and other end-users of multiplex PCR respiratory panels, is needed to understand how diagnostic stewardship interventions can add value at the points in time when these tests are ordered, resulted, and interpreted. Finally, although our respondents included several pediatric hospitals, respiratory panel testing may serve purposes in the care of pediatric patients, such as providing guidance on when a patient can return to school, that are less applicable among adults. Thus, our results likely do not fully address the use of respiratory panels among pediatric patients.
In conclusion, many experts in infectious diseases and hospital epidemiology believe that multiplex PCR respiratory panel testing is a useful tool that can improve diagnosis, patient care, and patient outcomes. However, medical facilities frequently employ diagnostic stewardship interventions to enhance the usefulness of these tests.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2023.72
Acknowledgments
Financial support
This work was supported by the SHEA Research Foundation via the SHEA Epi Project Competition. J.D.B. received research funding from the Agency for Healthcare Research and Quality (grant no. 1K08HS028854-01). S.L.K. is supported by a VA Health Services Research and Development Service Research Career Scientist Award (no. RCS 11-222).
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
