This manuscript has modeled several scenarios for the detection of Pandemic (H1N1) 2009, with the two most expensive scenarios being Scenario A and Scenario B. Pros and cons of different scenarios are shown in Table . Scenario A (RVP with/without CDC-M) consistently had the highest costs that were maintained throughout the whole pandemic period, although costs did drop as the % of specimens positive for influenza A increased. Not only were the costs high for Scenario A, the labor intensive nature of the assay and increased turn-around time for influenza A compared to other methods would make it difficult to deliver a timely answer to clinicians when compared to other scenarios. The costs for Scenario B (CDC-M first with H-typing and FluA-neg samples by RVP) was the second most expensive algorithm throughout the pandemic with costs dropping as the % of specimens positive for influenza A increased. However, the up-front use of the CDC-M assay would enable influenza diagnosis to be undertaken with less technologist hands-on time and a quicker turn-around time than RVP based methods. Test cost and increased technologist hands-on-time would occur during periods outside the influenza A peak, and diagnosis of other respiratory viruses would still be dependent on the use of the RVP.
Benefits of using each influenza A testing scenario during the pandemic
The two less costly algorithms were Scenario C and Scenario D. Scenario C (testing for influenza and respiratory virus by RVP only with H-typing) was the third lowest cost over all pandemic phases with costs crossing over with Scenario B at the peak. This algorithm would be focused on using the labor-intensive RVP assay and would require a longer turn-around-time to influenza A diagnosis than Scenario B. Furthermore, there have been some recent questions as to the use of RVP for the primary diagnosis of influenza A in patient specimens [10
]. The least expensive scenario across all phases of the pandemic was Scenario D (Influenza A by CDC-M only with H-typing). This scenario was also less labor intensive than scenarios focused on RVP and provided a quicker turn-around time to the diagnosis of influenza A than scenarios using RVP as the primary diagnostic tool. However, this scenario does not provide any data on other circulating respiratory viruses which had both diagnostic and surveillance value.
Outright costs alone should not be the primary driving force during the decision process regarding which algorithm is used. Other relevant factors in this decision-making process include needs for test performance data at the start of the pandemic, test turn-around-time, burden on laboratory staffing hours, availability of commercial kits, space available to carry out assays, and overall algorithm performance characteristics [10
]. Thus, laboratorians may decide to provide tests that are more costly to the laboratory but provide better patient care and may even save costs globally in patient care systems [11
]. These increased laboratory test cost per specimen may be recovered in increased efficiency in other areas such as clinical decision making, infection control and public health practice [12
The benefit of testing for other respiratory viruses needs to be determined by each laboratory following discussion with its client bases. These and other authors have often experienced that some laboratorians and clinicians believe that there is little benefit to testing viruses other than influenza due to the lack of widespread antivirals for other viral pathogens [14
]. However, testing for other viruses fulfills key surveillance roles and may have patient care and economic benefits in some settings [15
]. Steps such as cohorting patients based on viral etiology may be found by some institutions to ease the clinical management of patients [18
]. Identification of other viruses my also play a role in the discontinuation of antimicrobial therapy and decreased antibiotic use in some clinical settings [12
The identification of trends in viral was easy and readily accessible through the use of DIAL which is a partnership between Alberta's Provincial Laboratory for Public Health (ProvLab) and the Canadian Network for Public Health Intelligence (CNPHI). DIAL was founded by Drs. Jutta Preiksaitis, Bonita Lee and Shamir Mukhi in 2007 to address critical problems in extracting and managing laboratory-based data and was created so that ProvLab staff and other stakeholders would have an easier method for extracting, interpreting and analyzing laboratory data. Historically at ProvLab, the extraction of laboratory data from the Provlab information system (COHORT) was complex and could only be performed by computer programmers. Therefore, data was not easily accessible to laboratory staff, public health practitioners or other stakeholders. Moreover, the extracted data still required interpretation by laboratory experts to convert it into clinically meaningful final result. However, DIAL provides a solution to these problems by providing a simple web-based interface that enables users to access, summarize and analyze cleaned and interpreted real-time laboratory data [7
Pandemic preparedness involves not only the technical preparation of laboratories but also an understanding of both the cost implications of test utilization as well as the characteristics of each test algorithm. This manuscript indicates that a single methodology is not applicable to all conditions and that test characteristics may be as or more important than test-cost. Also it is notable that the cost of tests per specimen will vary depending on the prevalence of influenza A as well as other circulating viruses. Thus as the prevalence of influenza increases, an RVP only strategy (Scenario C) will increase in cost while a strategy that primarily using the CDC protocol with or without RVP (Scenario B) will decrease in cost. Clinician as well as patient needs may also have an impact on which algorithm is chosen as some situations may require quick turn-around-times (e.g. detection of influenza) while others may require an more comprehensive assessment of other respiratory viruses (e.g. cohorting patients with common respiratory infections) [19
This manuscript indicates that an ideal pandemic plan should allow for the laboratory to effectively shift between different algorithms as the pandemic progresses and depending on whether there is a need to identify other respiratory viral pathogens. The diagnostic and surveillance value in the identification of respiratory viruses supports the use of a combination of influenza testing and other RVP tests, or a multiplexed panel alone. The test volume, the proportion of specimens positive for influenza A and relative proportion of seasonal versus pH1N1 all affect the final cost, thus modeling for the optimal approach while fulfilling surveillance and diagnostic needs is complex. This movement to an RVP panel alone when influenza A prevalence is low would rely on effective near real-time surveillance systems that can provide decision makers the ability to analyze and review cleaned and interpreted laboratory data. In contrast, in a setting with a high prevalence of influenza, the laboratory leadership might decide, after consultation with the client base, whether only testing for influenza would be appropriate [22
]. Another cost-saving approach is to stop H-typing of all specimens at the peak of the pandemic when essentially over 90% of the positive specimens were pH1N1 which needs to be balanced with ongoing monitoring and surveillance initiatives. During the peak week of the second pH1N1 wave, 19-40% of the total cost was used for H-typing of influenza A positive specimens (data not shown). Therefore, laboratory planning and preparedness should include policies and procedures that ensure smooth algorithm transitions at all pre-analytical, analytical and post-analytical steps of the testing process.
It should also be noted that DIAL has applications outside of the pandemic and can be used for health care and public health planning during routine respiratory seasons. The near-real time capability of this system provides up-to-date information of circulating respirator virus and is of great benefit for the trending of respiratory virus overtime.
Pandemic planning should be process focused with well established standard operating procedures to ensure that staff are able to handle transitions effectively without extensive micromanagement [23
]. It is also important to have timely communications to the client base to indicate changes in algorithms during specific conditions and the impact of these changes on test ordering, clinical decision making and patient care [25
Such real-time decision making requires an interactive and simple to use data management system that allows decision makers to have access to the most up-to-date laboratory data. A system such as DIAL is ideal in this setting as it harvests real time information from laboratory information systems and allows for analysis of aggregate data [26
]. Access to this type of data may allow decision makers to potentially avoid decision making pitfalls such as; uncertainty, prejudice and optimism bias. However, the authors agree that biases will still exist even in the presence of DIAL, and decisions may still be made regardless of the data due to other factors impacting decision making such as group think, anchoring or choice-supportive bias [27