Matching diagnostics development to clinical need: Target product profile development for a point of care test for community-acquired lower respiratory tract infection

Background Point of care tests (POCTs) are increasingly being promoted for guiding the primary medical care of community acquired lower respiratory tract infections (CA-LRTI). POCT development has seldom been guided by explicitly identified clinical need and requirements of the intended users. Approaches for identifying POCT priorities and developing target product profiles (TPPs) for POCTs in primary medical care are not well developed, and there is no published TPP for a CA-LRTI POCT aimed at developed countries. Methods We conducted workshops with expert stakeholders and a survey with primary care clinicians to produce a target product profile (TPP) to guide the development of a clinically relevant and technologically feasible POCT for CA-LRTI. Results Participants with clinical, academic, industrial, technological and basic scientific backgrounds contributed to four expert workshops, and 45 practicing primary care clinicians responded to an online survey and prioritised community-acquired pneumonia (CAP) as the CA-LRTI where a new POCT was most urgently needed. Consensus was reached on a TPP document that included information on the intended niche in the clinical pathway in primary medical care; diagnostic product specification (intended use statement and test concept), and minimum and ideal user specifications. Clinicians minimum requirements of a CA-LRTI POCT included the use of minimally invasive samples, a result in less than 30 minutes, no more than a single preparation step, minimum operational requirements, and detection of common respiratory pathogens and their resistance to commonly prescribed antibiotics. Conclusions This multidisciplinary, multistage partnership approach generated a clinically-driven TPP for guiding the development of a new POCT, and this approach as well as the TPP itself may be useful to others developing a new POCT.


Clinical Need:
A POCT test for Community Acquired Pneumonia (CAP) and Acute Cough in primary care: clinical considerations CAP and Acute Cough can be both Lower Respiratory Tract Infections (LRTI). LRTIs are one of the commonest cause of acute illness in adults and one of the leading reasons for seeking medical care.

Spectrum of disease POCT
The need for a new POCT in primary care

The less severely ill
For the majority of patients presenting with mild to moderate illness in primary care, a test that gives prognostic information and rules out the need for antibiotic treatment would be most helpful. However, there is already a POCT for a biomarker, namely C-Reactive protein (CRP), and there are no known other biomarkers that have better prognostic properties in this setting. A test to detect viral pathogens, which, if known by a treating clinician to be present, might rule out the need to prescribe antibiotics and might rule in the need for antiviral treatment, would be the next most useful test for enhancing the management of these presentations in primary care. During pandemics, viral identification becomes more important as targeted antiviral therapy may be recommended. For example, during the recent H1N1 pandemic, more people received oselatmivir than were infected with the virus in primary and secondary care. As with overuse of antibiotics, over-use of antiviral agents is associated with developing resistance. An etiological diagnosis can guide decisions about other aspects of management apart from pharmacological treatment such as the decision about whether or not patients need to be isolated, and can help patients and clinicians make plans in the light of improved prognostic information.

The more severely ill
Severe CAP

Mild
Bacterial pathogen (and resistance) identification

Viral Identification
For the more severe end of the symptom severity spectrum, a test to detect and identify bacterial pathogens and resistance factors would be the most useful POCT. In patients assessed in hospital with possible CAP, a rapid POCT could influence patient outcome by aiding prompt imitation of appropriate antibiotics, with a favourable impact on mortality, antibiotic resistance, and cost (e.g. shortening the length-of-stay, decreased antibiotic consumption). The sooner appropriate antibiotic treatment is started in patients with CAP, the better the outcomes for patients. At the present time, two broad-spectrum antibiotic agents are usually started empirically for patients with CAP. Clinical outcomes and impact on resistance may be improved by a POCT that guides more specific therapy from the start of treatment. POCTs could also guide steeping down treatment to narrower spectrum (but equally effective) agents after treatment has already been started.

Focus on CAP
Incidence and cost: CAP is most often a bacterial infection and is has a high mortality (hospitalised patient mortality: 8% ward, 28% ICU). CAP is the leading cause of severe sepsis and death from infection in the United States. 3 CAP accounts for 600, 000 hospitalisations in the United States, with an annual cost in excess of €8 billion.
In the UK, the annual incidence of CAP is 5-11 per 1,000 adult population. The illness results in about 83,000 hospital admissions each year and is the fifth leading cause of death in the UK. The thirty-day mortality for all patients admitted to hospital with CAP is 18.3%, and these patients constitute 5.9% of critical care admissions. 4 Patients admitted to critical care have more severe disease, and accordingly critical care and hospital mortality rates are high (35% and 50% respectively). 5 The healthcare costs incurred as a result of CAP are estimated to be around €530 million annually in the UK. The average cost for management in the community was estimated at €120 per episode, and this can increase up to €6,123 when the patient is admitted to hospital. Substantial cost savings might be made by strategies for early. Well targeted effective therapy in CAP and thereby reduce hospital admission and improve patient outcomes.

CAP: A POCT to detect and identify bacterial pathogens and resistance factors:
Which pathogens, how many, antibiotic resistances?
No two studies of the aetiology of CAP use the same procedures and produce consistent results. Observed differences in aetiology are due to healthcare setting, disease severity, population factors, and epidemics at the time of study (the prevalence of infections with some organisms is cyclical, and many vary at random from year to year), and depending on study methodology. A test that detected a range of common bacterial pathogens would be of great value to clinicians, and there may be a case for the POCT to detect atypical pathogens (such as Legionella) as concern about these organisms leads some clinicians to prescribe more broadspectrum antibiotics. However, following RAPP-ID workshop group discussions the decision was made to focus initially on the most common bacterial respiratory pathogen, namely There is evidence that resistance to macrolides may indeed result in failure of treatment with macrolide antibiotics. 8

Focus on Acute Cough
Cough often presents to clinicians working in primary and secondary care simply as a troublesome symptom . Acute cough is usually due to an infectious organism (bacteria or virus) though there are also non-infectious causes. The vast majority of cases of acute cough presenting to primary care are due to acute viral respiratory tract infection. While many episodes of acute cough are self-limiting and generate little diagnostic uncertainty and can easily be managed without recourse to diagnostic tests, antibiotics continue to be over prescribed for this condition. In the face of increasing diagnostic uncertainty, clinicians tend to prescribe empirical antibiotics. A major worry for them is to 'miss' a case of pneumonia.
Estimates vary, but around 5% of patients with acute cough have pneumonia. A test that rapidly differentiates these patients with pneumonia from those that are unlikely to benefit from antibiotic treatment would greatly benefit clinicians.

Incidence and cost
Acute cough is one of the most common symptoms for which patients seek medical attention, the most common new presentation in primary care, and the most frequent reason for visits to hospital-based outpatient clinics. Approximately 100/1000 of all people in developed countries will present to primary care with an acute cough, and 2/1000 will require hospital treatment. In the USA, acute cough accounted for 26 million office visits in 2004. 9 It is estimated that he cost of acute cough to the UK economy is estimated to be at least €1,175 million/year. This comprises €1.050 million to loss of productivity and €125 million cost to the healthcare system and the purchase of non-prescription medicines. 10 In the UK, annual prescribing costs for acute cough alone exceed €18 million. 11 reasonably be expected to wait in a busy general practice surgery during a time they are ill.

The need for a new test
Typically, a clinician may take a test sample, asks the patient to wait in the waiting room, then see another patient (possibly two patients), and then see the original patient again to give the test result and decide on management. Such a time frame would typically between 10 and 20 minutes but certainly within 30 minutes.

Acute cough: A POCT to detect and identify common respiratory pathogens:
Detection of common respiratory bacterial and viral pathogens from a naso-pharyngeal (NP) swab or a breath sample would be of value in a primary care setting for patients with less severe illness where the decision about prescribing antibiotics was uncertain. Detection viral and bacterial pathogens is likely to help in decisions about antibiotic prescribing, may help in reassuring patients, and may be of help in making specific treatment decisions (for example antivirals for treating influenza).
The rapid influenza component of the test would be valuable in everyday primary and secondary care (particularly during the influenza season), in pandemic situations, and in settings outside of the healthcare system, such as ports of entry.

Outline of clinical pathway showing possible use of a POCT in primary care
Diagram of POCT Concept generally should be collected as close as is possible to the start of symptoms and usually no more than 4-5 days later in adults. In very young children, influenza viruses can be shed for longer periods; therefore, in some instances, testing for a few days after this period may still be useful. The accuracy of the influenza tests depends on the prevalence of influenza, and the positive and negative predictive values vary considerably depending upon the prevalence of influenza in the community. False-positive (and true-negative) influenza test results are more likely to occur when disease prevalence is low, which is generally at the beginning and end of the influenza season. False-negative (and true-positive) influenza test results are more likely to occur when disease prevalence is high, which is typically at the height of the influenza season.

Rapid methods for the detection of other potentially pathogenic respiratory viruses
In general, rapid methods for the detection of respiratory viruses are based either on immunological, or on molecular techniques.

Notes on Clinical Samples
Intuitively, it seems that organism detection may be optimal from a breath test rather than nasopharyngeal (NP) swabs (NP). A breath test would also be optimal for ease of use and patient acceptability, as it is one of the least invasive way of sampling. It does not require blood, sputum, urine, or pharyngeal or nasopharyngeal swab (i.e. non invasive). Clinician and patient acceptability is a major concern for successful uptake of any POCT especially in those with milder illness in primary care, and nasopharyngeal washings were found to be more acceptable than nasopharyngeal swabs in 91% of participants in one study. 12 Additionally, over 30% of the clinicians consulted in the RAPP-ID survey stated that they would not use a test that required a nasopharyngeal swab to be taken, though this may be due to unfamiliarity with the technique. However, nasopharyngeal swab samples were found to be well tolerated in the FP6 funded Genomics to combat Resistance against Antibiotics in

Community-acquired LRTI in Europe (GRACE) Network of Excellence Randomised Controlled
Clinical Trial of amoxicillin versus placebo for acute cough (Goossens, personal communication) and so swab type and clinician/nurse training may be an important issue here.
In addition to causing CA-LRTIs, S. pneumoniae can also colonise healthy people and has been found in the nasopharynx of 5-20% of healthy adults, and in up to 50% of healthy children. Quantitation of pneumococci in nasopharyngeal samples is promising as method for distinguishing asymptomatic colonisation from active infection and may be predictive of pneumonia. Patients with CA-LRTI showed 28-fold higher average loads of S. pneumoniae than colonised individuals. 13 Results (using nasopharyngeal swab samples) obtained with a semi-quantitative culture method showed good correlation with real-time PCR, although this study was small. However, it is difficult to clearly define an quantification threshold that differentiates invasive pneumococcal disease from colonisation and so the practical use of this measurement should be further evaluated. 14 Despite this, it should be considered that the prescription of antibiotics to a person 'colonised' rather than infected with S. pneumoniae is perhaps of less importance than the number of patients treated with antibiotics for a viral infection. The threshold for quantitative cultures from nasopharyngeal aspirate for diagnosing probable pneumococcal pneumonia has been quoted as ≥10 6 cfu/ml 15 and ≥10 5 cfu/ml. 16 Exhaled breath condensate (EBC) has been used for the detection of bacteria including S.
pneumoniae from patients with acute exacerbation of chronic obstructive pulmonary disease (COPD). 17 However, PCR analysis found that EBC and sputum results did not correlate well in this study though this may be related to novel technology that is still in development. Specific Volatile Organic Compounds (VOCs) can be detected in the exhaled breath of patients with exacerbations of COPD, 18 and volatile metabolites specific for S. pneumoniae have been isolated when the organism is grown in culture. 19,20 However, no references were found for in vivo detection of VOCs from S. pneumoniae.
Therefore, data are needed for both types of sample to differentiate colonisation from lower airway infection.

Intended Use(s) of the test:
• A POCT to enhance the initial clinical management of suspected CAP in adult patients presenting in primary care. The POCT should achieve this by indicating with sufficient precision the presence or absence of the organisms that account for almost all cases of CAP, so as to inform decisions about whether an immediate antibiotic prescription is necessary or not.

•
The primary target population is in primary care.
• The POCT should be easy to use by a range of health care professionals in a variety of settings (e.g. doctors, nurses and any community health workers in any primary care setting doctors surgery, GP home visit, and nursing homes).
• The POCT should be feasible and cost effective in the primary care setting.
• The test is designed to be used with either nasopharyngeal (NP) swab or exhaled breath samples without need for prior culture.

Intended use
To aid the antibiotic and antiviral prescribing and treatment decisions for adults presenting with symptoms of CA-LRTI in any primary care setting including nursing homes, care homes, out-of-hours clinics, accident and emergency departments, GP practices etc.
To aid the antibiotic and antiviral prescribing and treatment decisions for adults presenting with symptoms of CA-LRTI in the primary care setting (clinician surgery).

Medical decision to be influenced
Whether or not to prescribe and antibiotic or antiviral agent for patients with CA-LRTI, and if so, which agent to select for maximum patient benefit and minimal impact on selection of antimicrobial resistance To guide clinical decision making, specifically about whether or not to prescribe antibiotic and antiviral agents, for the common causes of CA-LRTI.

Place of use
Any site where a patient may consult within the primary care setting (including nursing homes, care homes, out-ofhours clinics, accident and emergency departments, GP practices) and elsewhere.
Any site where a patient may consult within the primary care setting.

Patient criteria
All adult patients including antibiotic pre-treated and those with no prior antibiotic treatment for that episode of CA-LRTI.
All adults with no previous antimicrobial treatment for that episode of illness GP clinical diagnosis of a microbiologically proven influenza virus infection: NPV is 75% (for a combination of clinical signs. 22 ).

Type of analysis
-Qualitative and quantitative analysis (or semi-quantitative analysis of pathogens present as both colonisers and infective agents) of the above mentioned bacterial pathogens.
-Qualitative analysis of common respiratory viruses including influenza A and B.
-Qualitative detection of the listed bacterial pathogens (For S. pneumoniae the boundary for detection should be set at a level which is indicative of colonisation rather than infection.
-Qualitative detection of common respiratory viruses -Qualitative analysis of S. pneumoniae macrolide and penicillin resistance genes.
including influenza A and B.
-Readable for at least 1 hour.
(Note: The above may need to be modified in light of technical specifications) -Readable for at least 1 hour.

Test stability
Temperature etc at which test should remain stable for a defined length of time Stable at well beyond ambient room temperatures and any temperatures that may be encountered in the shipping process.
Stable at ambient room temperatures normally encountered and any temperatures that may be normally encountered in the shipping process.

Storage conditions Prior to utilisation
Storage at well beyond ambient room temperature and humidity.
Storage at ambient room temperature and humidity.

Need for additional equipment
No requirement for any additional equipment.

Maintenance of instrumentation
No maintenance. Annual maintenance at most.

Power requirement
Can work on battery and mains.
Can work on mains. Max power 300W Max start-up current from 12V, 15A

Cost of test to end user
Considerably less than €10 per test.
No more than €10 per test.
More allowed for clinical trials on case-by-case review but still less than €200 per test.

Cost of platform to end user
Free or leased as part of a per use contract.
(For reference, Cepheid GX-I lists at €10,000 but actually sells for around €8,500) Less than €2,000 for commercial sales but provided free for clinical trials for the duration of the trial.

Cost of manufacturing single use device
Less than €2 per test when established volume of much greater than 10,000 units per annum is achieved.
€4-8 per test with established volume of 10,000 units per annum.

Competitive landscape
The need ideally, is to provide: a much faster -less than 10 minutes turn around time (TAT), better PPV and NPV than current techniques and competitors, and at a much lower price.
The large established players currently focussed on the laboratory market must be considered as possible competitors and likely to be able to leverage on their large volumes in the central laboratory / hospital markets to lower overall test costs. Relatively simple lateral flow tests also create an expectation of fast, low cost solutions to serve this market.
The performance of such tests is unlikely to be challenged in the primary care setting as much as within the hospital or laboratory setting.
The performance of Cepheid's GeneXpert is the current benchmark for many POC molecular tests and an important target to beat.
Smiths Detection Diagnostics and Engima's ML are the other two emerging players to monitor for performance both clinically and commercially since they should be able to serve the market competitively. All three of these products are modular and scalable and so can also address the hospital based POCT opportunity.

Market Segmentation
Any primary care healthcare setting for respiratory pathogen identification and drug resistance determination. Primary use targeted at both routine GP use and to assist in making CA-LRTI drug trials more efficient and effective. Secondary use targeted at antimicrobial stewardship.
High resource community setting, where no laboratory skills are available, to test for the presence of specific respiratory pathogens. Use targeted at both routine GP use and to assist in making CA-LRTI drug trials more efficient and effective.
It is noted that according to the accounting consultancy PwC and Roche the global Mergers and acquisitions are currently considered to be at exceptional levels and reflect the need of pharmaceuticals to invest in diagnostics to ensure sustainable business models emerge for the future. It is expected, given current trends, that by 2020 the need for co-development of drug and diagnostics will drive all the major pharmaceutical companies to acquire or develop an in-house diagnostic capability.
The trends in business models can neatly be summarised as a move from blockbusters to niche-busters.
Early detection of disease and personalised medicine are the key areas of current activity and look set to be a significant part of new emerging healthcare delivery models. The desire of governments is move as much care and management of patients to the community as possible and keep costs as low as possible by ensuring early interventions are truly cost effective whilst delivering improvements in morbidity, mortality and patient quality of care.

Market Overview Useful References
ECRI Institute Health Technology Assessment Intelligence Report for molecular based POCT of LRTI provided to Medical Device Consultancy (MDC) as a custom report during January 2010. The copyright remains with ECRI Institute. The report includes information concerning GeneXpert, bioMerieux, HandyLab, Osmetech and others. The report is evidence based and includes research findings concerning: • • Rapid diagnostic test to detect influenza provides similar information to that indicated • above.
• Rapid swine flu tests under rapid development.
• Test for avian influenza provides a useful overview (as above) includes impacts, test prices and ECRI perspectives and predictions.
• Clinical Utility of Methicillin-resistant Staphylococcus Aureus Screening Tests provides a summary on clinical utility and has a useful review of clinical studies and extensive references.
• Internet-based infectious disease surveillance systems for predicting and/or tracking disease outbreaks and spread provides a very useful review of key issues that are highly relevant to RAPP-ID.
• Invader UGT1A1 Molecular Assay for Personalized Medicine Management document • provides some initial insights into areas that might suggest future use areas of application for RAPP-ID technology such as pharmacogenetics and metabolomics.
Molecular-based Diagnostic Testing Using PathFinderTG is not currently relevant to RAPP-ID but does provide further insights into personalized medicine that should eventually be of real interest.
• The PCR (Polymerase Chain Reaction) Test for Diagnosis of Lyme Disease is included for completeness and does give a benchmark on sensitivity specificity, PPV, NPV of various players -this is the sort of analysis that any major laboratory is likely to undertake for themselves.
• Hospital-Acquired Infection Diagnosic Market and Test Developments, Treatment and Trends. Kalorama Information, July 2011.
• World Health Organization (WHO) web site on infectious diseases, including drugresistant tuberculosis and other infectious diseases.

Competitor tests (based on current manufacturers marketing material):
Information is provided here about the best currently available tests in order to guide the development of tests that will be more useful. As part of the retrospective study, urine specimens from 35 blood culture positive pneumococcal pneumonia patients and 338 presumed S. pneumoniae negative patients (373 total patients) were collected at 3 different facilities and evaluated in the BinaxNOW® Test. BinaxNOW® Test performance was calculated using standard methods. Sensitivity was 86%, specificity was 94%, and overall accuracy was 93%. Ninety-five percent (95%) confidence intervals are listed below. Sensitivity = 86% (71% -94%) Specificity = 94% (91% -96%) Accuracy = 93% (90% -95%)

Clinical sensitivity and specificity (Prospective Study)
In a separate seven-center prospective study, the BinaxNOW® Test was used to evaluate urine specimens collected from 215 hospitalized and outpatients presenting with lower respiratory symptoms or sepsis and from patients otherwise suspected of pneumococcal pneumonia. Patients were considered positive for pneumococcal pneumonia if diagnosed by positive blood culture. The BinaxNOW® Test performed equivalently on both outpatients and hospitalized patients. Ninety-five percent (95%) confidence intervals are listed below.

4) A number of tests to detect respiratory viruses including influenza A and B
(Excel file with information on a number of these tests has been provided separately to RAPP-ID partners)

CLINICAL CONDITION CLINICAL DEFINITION Community acquired pneumonia (CAP)
An acute infection of the pulmonary parenchyma that is associated with at least some symptoms of acute infection accompanied by ausculatory findings consistent with pneumonia (such as localized breath sounds and localized rales) and/or the presence of an acute infiltrate on chest radiograph, in a patient not hospitalized or residing in a long--term care facility for 14 days prior to onset of symptoms.

Influenza
A person presenting with influenza like illness. Including both seasonal and pandemic influenza.

Acute exacerbation of asthma
The acute onset of an increase in asthma symptoms (e.g. dyspnoea, wheeze), a productive cough, and increase and/or change in colour (e.g. green or yellow) of sputum and/or other symptoms suggestive of infection.

Acute exacerbation of chronic obstructive pulmonary disease (aeCOPD)
A combination of, or all three of; worsening of dyspnea, increase in sputum purulence and volume, as well as one of the following clinical criteria: an upper respiratory infection in the past five days; fever without another apparent cause; increased wheezing; increased cough; or increase in respiratory rate or heart rate by 20% above the baseline.

Acute bronchitis
The acute or sub--acute onset of cough accompanied with evidence of sputum production in a patient with no history of chronic pulmonary disease and no evidence of pneumonia or sinusitis.  (5) 'initial treatment targeting --whether to prescribe an antibiotic or not' should have been included in options' Patient age group that a new CA--LRTI should benefit (2) 'ideal test should be available for all age groups' (4) 'in neonates, we are very careful and start antibiotics easily when infection appears' 'if I was considering near patient testing in a child, I would probably be admitting them..'. Category of staff that would use the test most often (8) 'We have no nurses, practice nurse or nurse practitioner' 'Depends on the organisation of the healthcare system'