Abstract
The emergence of Zika virus in the Americas in 2015 and its association with birth defects and other adverse health outcomes triggered an unprecedented public health response and a demand for testing. In 2016, when Florida exceeded state public health laboratory capacity for diagnostic testing, the state formed partnerships with federal and commercial laboratories. Eighty-two percent of the testing (n = 33 802 of 41 008 specimens) by the laboratory partners, including Florida’s Bureau of Public Health Laboratories (BPHL; n = 13 074), a commercial laboratory (n = 19 214), and the Centers for Disease Control and Prevention (CDC; n = 1514), occurred from July through November 2016, encompassing the peak period of local transmission. These partnerships allowed BPHL to maintain acceptable test turnaround times of 1 to 4 days for nucleic acid testing and 3 to 7 days for serologic testing. Lessons learned from this response to inform future outbreaks included the need for early planning to establish outside partnerships, adding specimen triage strategies to surge plans, and integrating state and CDC information systems.
Keywords
Zika virus, a mosquito-borne virus identified in 1947 in the Zika Forest of Uganda, spread to the Americas in 2015 and was later found to be linked to the particularly severe forms of several diseases, including Guillain-Barré syndrome and fetal microcephaly (in which an infant’s head is smaller than expected when compared with infants of the same sex and age). 1,2 The Florida Department of Health first identified travel-associated Zika virus disease in state residents in January 2016, and Florida was the first state to identify cases of locally transmitted Zika virus disease in the continental United States. In 2016, the Florida Department of Health reported more than 1400 cases, approximately 20% of which were locally acquired. Those cases reported symptom onsets from June through December 2016, with more than half of cases reported in August. Although locally acquired cases were reported from 4 counties in Florida, most exposures occurred in Miami-Dade County. 3,4
The Florida Department of Health’s Bureau of Public Health Laboratories (BPHL), a state agency comprising 3 laboratories (Jacksonville, Tampa, and Miami), is responsible for detecting biological and chemical threats and emerging infectious diseases. BPHL prepared for Zika virus response by adding Zika virus real-time reverse transcriptase polymerase chain reaction (rRT-PCR) and immunoglobulin M antibody enzyme-linked immunosorbent assay (ELISA) to its test menu in 2015 and early 2016. Although no locally acquired cases had been detected in Florida at that time, BPHL prepared for Zika virus testing in clinical specimens because of the arrival of Zika virus in South America; the presence of competent mosquito vectors in Florida, which increased the risk of local transmission; and the high volume of international travelers between Florida and the already Zika virus–affected areas in the Western hemisphere. BPHL anticipated an increase in testing volume from travelers and possibly from locally acquired cases; however, the need to screen asymptomatic pregnant women in response to the risk of fetal microcephaly was not predicted. Because of these risks, in February 2016, the governor of Florida declared a public health emergency, which released resources to local jurisdictions with imported cases of Zika virus disease so they could take precautionary actions to prevent local transmission of Zika virus. 5
As a lead agency for public health emergencies as designated in the Florida State statute, the Florida Department of Health established an Incident Management Team, which led the response through its Incident Command Center by using Incident Command System principles as outlined by the Federal Emergency Management Agency (FEMA). The Incident Command System is an organizational system that uses standardized processes and terminology to ensure command, control, and coordinated response operations. 6 The Incident Management Team organized the response and ensured coordination with all stakeholders to implement objectives outlined in the Florida Department of Health Zika Incident Response Playbook. 7 The Incident Management System manages incidents by objectives, and the Zika Incident Response Playbook provided the Incident Management Team with these baseline objectives. As the incident progressed, the objectives and operational tactics were adjusted during the planning cycle and documented in a daily incident action plan.
In August 2016, the governor directed the Florida Department of Health to provide free Zika virus infection risk assessment (based on exposure, travel history, and symptoms) and testing to any pregnant woman in the state because of locally acquired cases and increased risk of fetal microcephaly associated with maternal Zika virus infection.
This case study describes how a public health laboratory established private and public partnerships at the local, state, and federal levels to provide timely quality laboratory results during an infectious disease public health emergency. We describe the actions taken by BPHL to increase surge capacity through collaborative efforts with laboratory partners to test 41 008 Zika virus specimens in 2016. This case study, which to our knowledge is the first description of such a partnership in the literature, can assist public health laboratories in preparedness planning and response. Forming partnerships is 1 of the 11 core functions of state public health laboratories that can strengthen public health laboratory systems, 8 which can be difficult to establish and maintain, especially during a public health emergency response.
Methods
Laboratory Tests
BPHL performed Zika virus testing in all 3 of its laboratories: Jacksonville, Miami, and Tampa. The laboratories established laboratory-developed tests in preparation for test requests because US Food and Drug Administration (FDA)–approved assays were not initially available. Laboratory-developed tests are validated according to internal procedures. These tests are often created in response to unmet testing needs and are not regulated by the FDA. 9 The laboratories in Jacksonville and Tampa established capacity for rRT-PCR testing in July 2015 to detect Zika virus nucleic acid in serum specimens from acute infections, which have the potential to be transmissible via the bite of a mosquito during the viremic (virus circulating in the blood) phase. During 2016, BPHL expanded the acceptable specimens to include urine, saliva, amniotic fluid, placental tissue, and semen. The laboratories used a laboratory-developed test based on a Centers for Disease Control and Prevention (CDC) method that was established in response to a Zika virus outbreak in Micronesia in 2007. This method defined 2 genome position target sequences for which rRT-PCR primers and probes were designed. 10 As suggested by CDC, BPHL modified 1 of the primer sequences to detect the Zika virus Asian genotype that was circulating in the Western hemisphere.
In January 2016, the Jacksonville and Tampa laboratories added capacity to perform Zika immunoglobulin M antibody testing to detect recent but nonviremic infections in serum and cerebrospinal fluid by using a laboratory-developed test based on a CDC method. Viral-specific immunoglobulin M antibody detection indicates a recent infection because these antibodies are produced during the first few days after symptom onset. 11 The assay uses anti-immunoglobulin M antibody coated 96-well plates that capture immunoglobulin M antibody present in the specimen. A “sandwich” is then created with the addition of Zika virus noninfectious antigen that binds to Zika immunoglobulin M, followed by an enzyme-conjugated antiviral antibody that will cause a colorimetric change when a chromogenic substrate is added. The colorimetric change is detected with a spectrophotometer. This method was the basis for the FDA Emergency Use Authorization (EUA)–approved CDC immunoglobulin M antibody capture ELISA (MAC-ELISA) released in late February 2016. 11 The FDA allows for emergency use of laboratory methods that are not fully approved or cleared when a diagnostic test is needed to respond to public health emergencies.
In March 2016, BPHL adopted the EUA-approved ELISA, which was used in conjunction with dengue immunoglobulin M and dengue immunoglobulin G ELISA when required diagnostically. BPHL used a laboratory-developed dengue MAC-ELISA assay based on the same CDC method as for the Zika MAC-ELISA with dengue virus noninfectious antigen. BPHL used a laboratory-developed dengue immunoglobulin G ELISA based on a CDC method. Viral-specific immunoglobulin G antibody detection indicates a past infection because these antibodies are produced later than immunoglobulin M and are lifelong but cross-reactive with other flavivirus immunoglobulin G. 12 The assay is similar to the MAC-ELISA with sandwich of antigen, specimen antibody, and enzyme-conjugated anti-human immunoglobulin G antibody, but in a different sequence.
Also in March 2016, the FDA granted EUA approval of the Trioplex rRT-PCR assay, a multiplex assay with 1 genome position target for each of 3 viruses: dengue, chikungunya, and Zika, which makes it possible to distinguish Zika from clinically similar dengue and chikungunya infections. 11 The Miami laboratory added capacity to perform the Trioplex rRT-PCR test in July 2016, using this assay to test for Zika virus nucleic acid in serum and other specimen types (ie, cerebrospinal fluid, whole blood, urine, and amniotic fluid). These other specimen types were indicated as potential other sources of Zika virus detection because of persistence of viral nucleic acid in them longer than in serum, but the FDA–EUA approval required that they be tested alongside a patient-matched serum because they had not been validated as extensively as serum for arbovirus detection. 13 Also, BPHL-Miami performed the Trioplex assay on urine survey specimens, a process not described in this case study. When the Zika virus testing demand began to decrease and high-capacity nucleic acid extraction instruments approved for the EUA-approved rRT-PCR assay were obtained, the laboratories in Jacksonville and Tampa verified the Trioplex assay and transitioned to it in December 2016.
CDC performed testing by using the FDA EUA–approved CDC MAC-ELISA and Trioplex rRT-PCR assay. Two to 4 shipments of specimens were sent to CDC-Atlanta weekly from August 9 through October 5, 2016, and distributed among 3 divisions for testing (Division of Health Care Quality Promotion, Division of Bacterial Diseases, and Division of Viral Diseases). CDC and BPHL established a file transfer protocol to transmit a secure Excel file with data on specimens and patient demographic characteristics. The contracted commercial laboratory performed testing by using FDA EUA–approved PCR assays on serum and urine (alongside a patient-matched serum) and immunoglobulin M antibody assays on serum.
A positive Zika virus RNA detection by PCR assay was confirmatory. For a patient with only a positive, equivocal, or inconclusive Zika virus immunoglobulin M result, CDC confirmed the presence of neutralizing antibodies by a plaque reduction neutralization test following test request and specimen referral from BPHL or the commercial laboratory. This test is a complex serologic assay requiring as long as several weeks to obtain a result; details of it are not a part of this case study and as of this writing had not been published.
Test Reports
BPHL generated specimen reports from its laboratory information management system software, used by BPHL member laboratories to manage specimens and associated data, and sent reports to non–Florida Department of Health submitters by postal mail or fax. Clinic providers in the Florida Department of Health received electronic laboratory results through messaging to the county health department health management system. The state-level Florida Department of Health Bureau of Epidemiology and county health department epidemiologists received electronic laboratory results through messaging to Merlin, Florida’s reportable disease surveillance system. 14 BPHL sent these electronic laboratory results directly from the laboratory information management system via health level 7 messaging, which is compliant with the most widely used international standard for transfer of clinical and administrative data between clinical and public health entities. 15 CDC sent specimen reports through a secure automated email attachment to BPHL, and BPHL manually entered results into the BPHL laboratory information management system and reported them as described previously. Commercial laboratory results were sent to the submitting provider and reported to Merlin through electronic laboratory reporting.
Test Volume and Turnaround Time
We obtained data on test volume from the BPHL laboratory information management system for specimens tested by BPHL and CDC and from Merlin for specimens tested in the commercial laboratory for Zika virus PCR and immunoglobulin M tests. We calculated median test turnaround times weekly from day of specimen receipt at BPHL or the commercial laboratory to day of BPHL test result release or commercial laboratory electronic laboratory results message creation, respectively. When calculating test turnaround times for specimens sent to CDC-Atlanta, we included the day of specimen receipt at BPHL to account for the time needed to prepare specimens for shipping. We also projected weekly test turnaround times for BPHL if the specimens had not been sent to CDC using turnaround times before CDC shipment, the number of specimens received, and estimated weekly test capacity.
Laboratory Triage and Testing Algorithms
BPHL developed algorithms for triaging and testing the large volume of specimens submitted to it. BPHL established a testing partnership with a large for-profit commercial laboratory through the Florida Department of Health Incident Management Team within 2 days of the free testing for pregnant women announcement on August 3, 2016. County health department clinics sent specimens from low-risk patients (asymptomatic pregnant patients with no exposure) to the commercial laboratory and specimens from high-risk patients (pregnant or not pregnant symptomatic patients and pregnant or not pregnant asymptomatic patients with exposure) to BPHL. Exposure was defined as living in or visiting the Miami-Dade transmission area, traveling to a Zika-endemic country, having sexual contact with a person with confirmed Zika virus infection, or being born to a Zika-positive or symptomatic mother. Patients were considered symptomatic if they had symptoms such as fever, rash, joint pain, conjunctivitis, Guillain-Barré syndrome, or fetal ultrasound indicating possible congenital Zika syndrome. All specimens from patients who lived or worked in Miami-Dade County were considered high risk and were sent to BPHL-Miami for testing or to a triage unit because most local Zika virus exposures occurred in Miami-Dade County. BPHL-Miami established this Zika triage unit with staff members deployed by the Florida Department of Health Incident Management Team and CDC Emergency Operations Center, a unit within CDC that provides resources in public health emergencies.
BPHL-Miami received specimens from the Florida Department of Health that were collected in Miami-Dade County’s pregnancy Zika testing clinics (special clinics set up for that purpose by the county health department), obstetrics and gynecology providers, and multiple hospitals. BPHL-Miami tested specimens from the pregnancy clinics (which were considered high risk because they came from the Miami-Dade transmission area) by Trioplex and forwarded to BPHL-Tampa for MAC-ELISA, and all other specimens went to the triage unit. Prior to receipt of the specimens in the triage unit, Florida Department of Health epidemiologists indicated that testing was warranted for public health purposes by approving testing based on patient symptom, pregnancy, and possible exposure status. When samples arrived in the laboratory, the triage unit verified prior approval and triaged the specimens based on the low-risk or high-risk criteria for shipment to the commercial laboratory, BPHL-Tampa, or BPHL-Jacksonville. BPHL did not test specimens that did not meet epidemiologic requirements. Specimens from patients who were symptomatic and pregnant with or without exposure were sent to BPHL-Tampa. Specimens from nonpregnant patients who were asymptomatic with exposure were sent to BPHL-Jacksonville. Specimens from pregnant patients without symptoms or exposure were sent to a commercial laboratory. This triage allowed BPHL to retain specimens from patients meeting high-risk criteria and distribute the specimen load between BPHL-Tampa and BPHL-Jacksonville based on staffing and capacity. BPHL-Tampa forwarded a subset of specimens from pregnant or not pregnant symptomatic patients to CDC from August 9 through October 5, 2016. BPHL-Tampa made the decision to send specimens to CDC-Atlanta based on the rapid increase in specimen submissions after the free testing announcement and an anticipated backlog (Figure 1).
Triage algorithm for Zika virus specimens used by the Florida Department of Health Bureau of Public Health Laboratories (BPHL)–Miami triage unit, Florida, July–December 2016. Exposure was defined as living in or visiting the Miami-Dade transmission area, traveling to a Zika-endemic country, having sexual contact with a person with confirmed Zika virus infection, or being born to a Zika-positive or symptomatic mother. Patients were considered symptomatic if they had symptoms such as fever, rash, joint pain, conjunctivitis, Guillain-Barré syndrome, or fetal ultrasound indicating possible congenital Zika syndrome. “Has EPI stamp” means the specimen submission form had a stamp indicating testing approval by an epidemiologist (EPI). The triage unit verified the EPI stamp based on information on the specimen submission form and contacted an epidemiologist if the approval was questionable because of inaccurate forms, to confirm (or deny) the approval as appropriate. “EPI approved” means information on pregnancy status, symptoms, travel, and sexual history was required to approve or deny testing. Specimens from patients who were symptomatic and pregnant with or without exposure were sent to BPHL-Tampa. Specimens from nonpregnant patients who were asymptomatic with exposure were sent to BPHL-Jacksonville. Specimens from pregnant patients without symptoms or exposure were sent to a commercial laboratory. BPHL did not test specimens that did not meet epidemiologic requirements. BPHL-Tampa forwarded a subset of specimens from pregnant or not pregnant symptomatic patients to the Centers for Disease Control and Prevention (CDC) from August 9 through October 5, 2016.
BPHL developed testing algorithms based on guidance provided by CDC, although Florida made modifications based on local transmission and the decision to offer testing to all pregnant women in the state (Figures 2 and 3). These modifications included expanded serologic testing for suspected local cases (based on symptoms and no recent travel history outside of Florida), suspected Guillain-Barré cases (based on symptoms), and pregnant patients by Zika MAC-ELISA, dengue MAC-ELISA, and dengue immunoglobulin G ELISA testing regardless of Zika rRT-PCR result to obtain the full serologic profile for epidemiologic determination as an official case. Although immunoglobulin M levels vary, they may be detectable approximately 4 days and up to 12 weeks after symptom onset or exposure, and cross-reactivity between Zika and dengue viruses on serologic tests has been reported. 16 As another modification, BPHL performed rRT-PCR testing on all pregnant patients because the timing of a possible infection could not be established and the possibility existed of prolonged viremia during pregnancy, which was documented in other cases. 17 BPHL also performed rRT-PCR on suspected local cases because the time (or lack) of symptom onset or exposure may not have been clearly established and on suspected Guillain-Barré cases because little was known about the viremia in these patients.
Testing algorithm for Zika virus specimens collected from patients who had possible Zika virus symptom onset or exposure within 14 days before specimen collection, used by the Florida Department of Health Bureau of Public Health Laboratories, Florida, 2016. The algorithm focused on real-time reverse-transcriptase polymerase chain reaction (rRT-PCR) assays using the laboratory-developed and US Food and Drug Administration (FDA) emergency use authorization (EUA)–approved tests on specimens collected <14 days after symptom onset or exposure because of a greater likelihood of patient viremia (virus circulating in the blood) and detectable Zika virus nucleic acid in specimens. For specimens collected <14 days after symptom onset, symptoms included fever, rash, joint pain, conjunctivitis, Guillain-Barré syndrome (GBS), or fetal ultrasound indicating possible congenital Zika syndrome. Exposure was defined as living in or visiting the Miami-Dade transmission area, traveling to a Zika-endemic country, having sexual contact with a person with confirmed Zika virus infection, or being born to a Zika-positive or symptomatic mother. Although specimens collected <14 days after symptom onset are most likely to be rRT-PCR positive, Zika virus detectability may vary by specimen type, specimen handling practices, and timing of the immune response. For specimens from suspected local cases (based on symptoms and no recent history of travel outside of Florida), suspected GBS cases (based on symptoms), and pregnant patients, additional serologic testing including immunoglobulin M antibody capture enzyme-linked immunosorbent assay (IgM MAC-ELISA), dengue MAC-ELISA, and dengue immunoglobulin G (IgG) ELISA regardless of Zika rRT-PCR result was performed to obtain the full serologic profile for epidemiologic determination as an official case. Specimens testing positive by Zika MAC-ELISA but negative by rRT-PCR from suspected local cases, suspected GBS cases, and pregnant patients were sent to the Centers for Disease Control and Prevention (CDC) for plaque reduction neutralization test (PRNT).
Testing algorithm for Zika virus specimens collected from patients with possible Zika virus symptom onset or exposure ≥14 days before specimen collection, Florida Department of Health Bureau of Public Health Laboratories (BPHL), Florida, 2016. This algorithm focused on immunoglobulin M antibody capture enzyme-linked immunosorbent assay (IgM MAC-ELISA) testing on specimens collected ≥14 days after symptom onset or exposure due to a lower likelihood of patient viremia (virus circulating in the blood) but detectable Zika virus antibody, which develops later than detectable Zika virus nucleic acid in specimens. Symptoms included fever, rash, joint pain, conjunctivitis, Guillain-Barré syndrome (GBS), or fetal ultrasound indicating possible congenital Zika syndrome. Exposure was defined as living in or visiting the Miami-Dade transmission area, traveling to a Zika-endemic country, having sexual contact with a person with confirmed Zika virus infection, or being born to a Zika-positive or symptomatic mother. Specimens from suspected local cases (based on symptoms and no recent history of travel outside of Florida), suspected GBS cases (based on symptoms), and pregnant patients who tested positive, equivocal, or inconclusive by MAC-ELISA were also tested by real-time reverse-transcriptase polymerase chain reaction (rRT-PCR) using the laboratory-developed and US Food and Drug Administration (FDA) emergency use authorization (EUA)–approved tests for the following reasons: the timing of a possible infection could not be established and the possibility existed of prolonged viremia during pregnancy, the time (or lack) of symptom onset or exposure may not have been clearly established for the suspected local cases, little was known about the viremia in GBS patients. For these cases, BPHL also performed dengue MAC-ELISA and dengue immunoglobulin G (IgG) ELISA regardless of rRT-PCR result to obtain the full serologic profile for epidemiologic determination as an official case. Although IgM levels vary, they may be detectable approximately 4 days and up to 12 weeks after symptom onset, and cross-reactivity between Zika and dengue viruses on serologic tests has been reported. Specimens testing positive by Zika MAC-ELISA but negative or equivocal by rRT-PCR from suspected local cases, suspected GBS cases, and pregnant patients were sent to the Centers for Disease Control and Prevention (CDC) for plaque reduction neutralization test (PRNT). Abbreviation: CSF, cerebrospinal fluid.
BPHL followed 2 algorithms. The first algorithm, for specimens collected from patients with possible Zika virus symptom onset or exposure within 14 days prior to collection, focused on rRT-PCR testing using the laboratory-developed and FDA EUA–approved tests because of a greater likelihood of patient viremia and therefore detectable Zika virus nucleic acid in specimens. But Zika virus detectability may vary by specimen type, specimen handling practices, and timing of the immune response; therefore, for specimens from suspected local cases, suspected Guillain-Barré syndrome cases, and pregnant patients, BPHL performed the additional serologic testing described previously.
The second algorithm, for specimens collected from patients with possible Zika virus symptom onset or exposure ≥14 days prior to collection, focused on Zika MAC-ELISA testing by using the FDA EUA–approved test because of a lower likelihood of patient viremia at that time but a higher likelihood of the presence of detectable Zika virus antibody, which develops later in the course of infection, when Zika virus nucleic acid may no longer be detectable. Because Zika virus immunoglobulin M detection can be cross-reactive with dengue immunoglobulin M, 16 for specimens from suspected local cases, suspected Guillain-Barré syndrome cases, and pregnant patients, BPHL performed the additional dengue virus serologic testing described previously.
CDC-Atlanta performed testing following the Florida testing algorithms. The commercial laboratory performed FDA EUA–approved Zika virus PCR testing on serum and patient-matched urine specimens and FDA EUA–approved immunoglobulin M antibody testing on serum specimens unless indicated otherwise by the ordering submitter.
Outcomes
Test Volume
In 2016, all the laboratories combined tested 41 008 Zika specimens. In 2016, BPHL tested 17 749 Zika specimens, including 14 470 laboratory-developed PCR tests and FDA EUA–approved PCR tests and 8846 laboratory-developed immunoglobulin M antibody and FDA EUA–approved immunoglobulin M antibody tests. In 2016, the commercial laboratory tested 21 745 specimens, including 16 087 FDA EUA–approved PCR tests and 10 667 FDA EUA–approved immunoglobulin M antibody tests. In 2016, CDC-Atlanta tested 1514 specimens that BPHL sent from August 9 through October 5. During the peak specimen submission period, from July through November 2016, 82% (33 802/41 008) of all 2016 submissions were received. This period encompassed the peak of local transmission (July–October 2016), and the numbers of specimens tested by each entity during the period were 13 074 specimens (10 938 laboratory-developed PCR tests and FDA EUA–approved PCR tests and 6408 FDA EUA–approved immunoglobulin M antibody tests) at BPHL, 1514 specimens (1482 FDA EUA–approved PCR tests and 658 FDA EUA–approved immunoglobulin M antibody tests) at CDC-Atlanta, and 19 214 specimens (13 757 FDA EUA–approved PCR tests and 10 268 FDA EUA–approved immunoglobulin M antibody tests) at the commercial laboratory (Figure 4).
Zika virus test volume and weekly median turnaround time (TAT) for (a) polymerase chain reaction (PCR) and (b) immunoglobulin M (IgM) antibody tests, Florida Department of Health Bureau of Public Health Laboratories (BPHL), Centers for Disease Control and Prevention (CDC), and a large for-profit commercial laboratory, Florida, July through November 2016.
Turnaround Times
From July through November 2016, median weekly turnaround times (the medians of turnaround times for all tests performed during a particular week) for Zika virus PCR testing ranged from 1 to 4 days; the total range of individual turnaround times during that entire period (hereinafter, “total range”) was 1 to 64 days at BPHL, 6 to 8 days (total range = 3-29 days) at CDC, and 3 to 5 days (total range = 2-13 days) at the commercial laboratory. The median weekly turnaround times for Zika virus immunoglobulin M antibody testing was 3 to 7 days (total range = 1-73 days) at BPHL, 7 to 19 days (total range = 3-48 days) at CDC, and 3 to 13 days (total range = 2-19 days) at the commercial laboratory (Figure 4). Peak turnaround times tended to coincide with peak testing volumes, especially in the first few weeks after the free testing announcement. CDC turnaround times were longer because of the time involved in preparing and shipping the specimens to CDC, including transcription of data on patient demographic characteristics from paper requisition forms to computer systems, resolution of missing or questionable demographic information, labeling and aliquoting of specimens into leak-proof tubes suitable for air transport and not shipping every day, and delays in the resulting communication process, including results lost in email transmission or emails missed in the BPHL result extraction process. The median turnaround time for specimens tested by CDC from time of receipt at CDC to result was 2 days for Zika virus PCR and 5 days for Zika virus immunoglobulin M. The median weekly turnaround times at BPHL were only slightly longer than pre-Zika virus testing levels, which ranged from 1 to 3 days for PCR and 2 to 5 days for immunoglobulin M antibody assays despite the laboratories performing testing beyond anticipated peak capacity. The projected weekly test turnaround time for BPHL had the specimens not been sent to CDC from August through October 2016 would have been 4 to 10 days for PCR and 5 to 15 days for immunoglobulin M antibody.
Lessons Learned
Public health laboratories must respond to public health emergencies, but the extent and duration of that response cannot be predicted because each emergency comes with its own unique factors. After adding Zika virus testing to the BPHL test menu in late 2015, only 2 Zika specimens were tested in that year; in contrast, for the entire year (2015), BPHL performed 1174 PCR tests and 1331 immunoglobulin M antibody tests for all arboviruses (eg, West Nile, dengue, and chikungunya). In 2016, BPHL performed 24 775 PCR tests and 14 470 immunoglobulin M antibody tests for all arboviruses, including Zika virus. In 2016, BPHL responded to a 21-fold increase in PCR testing and an 8-fold increase in immunoglobulin M antibody testing compared with 2015 because of the Zika virus threat. Initially, BPHL had staffing, equipment, and supplies sufficient to test approximately 450 specimens per week based on capacity experience from the 2009 influenza pandemic. When this capacity became insufficient, additional staff members, equipment, and supplies were obtained that allowed testing of up to 1000 specimens per week. Testing volume continued to exceed this new laboratory surge capacity. Therefore, testing partnerships BPHL established with CDC laboratories and a commercial laboratory were important for maintaining quality testing with timely reporting of results. Specimen testing at CDC prevented a BPHL backlog that could have increased the median weekly turnaround times by up to 6 days for PCR and 8 days for immunoglobulin M antibody testing. Specimen testing at the commercial laboratory allowed BPHL to focus testing on suspected high-risk cases while maintaining acceptable turnaround times for testing of pregnant women in the state.
The decision to retain and test specimens from high-risk patients and refer specimens from low-risk patients to the commercial laboratory was effective for communicating and acting on high-priority case results. Because commercial laboratories are not routinely involved in testing epidemiologically important cases, rapid communication and the ability to accommodate additional testing are not common practices at these laboratories. The established lines of communication between BPHL and the Bureau of Epidemiology and an agreed-upon testing algorithm for additional testing facilitated prompt diagnostic test results, which was critically important for identifying locally acquired cases. The testing algorithm reduced unnecessary testing and focused additional testing, including dengue immunoglobulin M, dengue immunoglobulin G, and the plaque reduction neutralization test, on specimens from high-risk and pregnant patients.
These partnerships came with challenges that led to in-process changes and recommendations. The complex triage system for determining which specimens would be sent to the commercial laboratory required both substantial time and the development of robust preanalytical and postanalytical capabilities. BPHL often received specimens with incomplete demographic or submitter information that required research by the triage unit, resulting in delays in specimens being readied for shipment. The commercial laboratory had strict specimen submission requirements that resulted in rejection of specimens. To address these issues, BPHL and the Bureau of Epidemiology created guidance documents for specimen collection and submission. As the outbreak continued, the Incident Management Team established direct submission of specimens to the commercial laboratory for several large hospitals in Miami-Dade County.
The Florida Department of Health Incident Management Team responded rapidly to the increase in testing volume by establishing the testing partnership with the commercial laboratory, but a transition phase was needed to bring all the county health department clinics on board and start sending specimens to that laboratory. Therefore, BPHL initially performed testing of asymptomatic pregnant patients with low risk of exposure for many counties in the state, which contributed to the need to establish surge testing at CDC. The lesson learned from this surge in testing is to prepare submitters in advance for commercial laboratory partnership testing, with guidance documents for specimen collection and submission included in capacity plans so partnership testing can be implemented quickly.
Communication with CDC prior to exceeding in-house testing capacity was critically important to set up referral of specimens. However, if interaction with CDC had occurred earlier, more than 1 BPHL laboratory could have set up the referral process, possibly maintaining turnaround times within pre-Zika parameters. Delays in reporting a few CDC results through the BPHL laboratory information management system occurred because some reporting emails were overlooked. Integration of reporting systems in state health departments and the CDC laboratory information management system is a recommended priority.
This experience can be applied to future outbreaks of emerging or reemerging arboviruses or a resurgence of Zika virus. A successful public health laboratory response to a public health emergency depends on existing infrastructure (trained personnel, equipment with sufficient throughput, and adequate reagents and supplies) and the ability to expand infrastructure when needed. Coordination through the Incident Command System was key in acquiring needed resources during the Zika virus response because it affected many programs at the local and state levels. Local-level programs initially used local funds with the addition of funds from the state level when the public health emergency was declared. Later in the response, the release of federal funds sustained the response. Because of the magnitude of the response, a thorough cost breakdown and analysis was beyond the scope of this article. Generally, BPHL specimen testing was supported through both state and existing federal resources, and CDC specimen testing was supported through the CDC operating budget. The commercial laboratory testing was provided through state funding because this activity was not considered eligible for most federal sources.
In addition, as demonstrated, partnerships were important for maintaining high testing capability in Florida during local transmission of Zika virus in Miami-Dade County. Capacity plans for test volumes need to include outside partners to be able to handle unanticipated numbers of specimens. The Incident Management Team, through its Incident Command System, was essential in organizing the response to the increase in testing volume to ensure coordination and communication with public and private partners in many locations. A sustained public health laboratory infrastructure and effective partnerships will be important for the public health laboratory response to future outbreaks.
Footnotes
Acknowledgments
The authors acknowledge the dedicated personnel at the Florida Department of Health, Bureau of Public Health Laboratories (Miami, Tampa, and Jacksonville) and CDC (Atlanta, Georgia, and Fort Collins, Colorado) laboratories for Zika virus diagnostic testing.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by the Centers for Disease Control and Prevention (CDC) Epidemiology and Laboratory Capacity for Infectious Diseases grant. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of CDC.
