General approach
Diagnostic pathways for patients with suspected RTIs were developed based on clinical guidelines and data from interviews with GPs working in four European countries: Norway, the Netherlands, England and Italy. Practices that have adopted POC CRP testing in Netherlands and Norway were included as benchmarks of how testing could be implemented. Practices that do not currently use POC CRP testing based in England and Italy were included to evaluate the potential effect of implementing POC CRP testing as part of the RTI pathway [15, 16, 30].
Pathways were based on reported current standard of care (SoC) practice or hypothetical future practice. A micro-costing approach was taken in which the resources used and associated costs (staff time, consumables and diagnostic testing) were estimated for each pathway step to give a unit cost, weighted by the proportion of patients flowing through each step.
The final pathways were clinic-specific, one for each practice that uses POC CRP testing and two for each practice that does not. Additional details are presented in Supplementary Material S1-S13.
Clinician recruitment
Full-time GPs who have access (Netherlands/Norway) or do not have access (England/Italy) to POC CRP testing were recruited by a third-party market research agency based on an eligibility screening questionnaire (Supplementary Material S13). Limited inclusion criteria were applied to facilitate the recruitment of typical clinicians.
Data collection
Clinical guidelines [15, 16, 31] and published literature were used to inform draft SoC pathways for each country, outlining the current use of CRP testing and treatment [9, 15, 16, 31]. No RTI guidelines were found for Italy, so the SoC pathway was drafted using guidelines from England and the Netherlands [9, 16, 31]. These pathways were used as prompts alongside discussion guides within the semi-structured interviews.
CRP concentrations were split into three categories in the draft POC pathways: <20 mg/L, where antibiotics are not indicated, > 100 mg/L, antibiotics are indicated, and 20–100 mg/L, where delayed antibiotics are considered, in line with international guidance [13, 16]. In practices with implemented POC CRP testing, the results on CRP concentrations were readily available to inform clinical decision-making.
Pre-interview survey
Participants completed a pre-interview survey (Supplementary Material S11) to aid data collection regarding the RTI pathway in their current clinical practice. The survey responses were used to refine the discussion guides and pathways for each country and to create targeted questions for the interview.
Semi-structured interview
Interviews occurred between April and July 2022. They were scheduled for 90 minutes and conducted virtually using Microsoft Teams by two researchers per interview from a pool of five (EL, MW, KPS, AM, CHW). All interviews were conducted in English without use of interpreters. Interviews were transcribed to text, which was then reviewed and validated by the researchers. Participants received an honorarium to compensate them for their time.
During the interview, participants were taken through the proposed pathways and asked to comment on the accuracy of the pathway steps and to estimate the probability of patients moving through each pathway branch. Information was collected about staff costs, staff time and other resource use during patient consultation. In practices without access to POC CRP testing, clinicians were asked how the pathway might change with its introduction.
Post-interview validation
Following the interview, participants were either sent a second survey via email (n = 4; Netherlands/England) or invited to a 30-minute follow-up interview (n = 4; Norway/Italy). This enabled them to validate the updated pathway structure based on the interview, approve or amend their estimates of the patient proportions at each point in the pathway and provide any comments. Additional information that had not been covered in the initial interview was also requested, such as which antibiotics they prescribed and number of weekly telephone consultations. An example of a post-interview validation survey is presented in Supplementary Material S12. In some cases (n = 5), additional follow-up questions were emailed to the participant.
Outcomes
The outcomes calculated for each pathway were the mean number of antibiotic prescriptions per 100 patients, the percentage of patients referred to hospital for further assessment and management, and the average pathway cost per patient presenting with a suspected RTI (£, €).
To understand how much sooner CRP results could be available with the implementation of near-instant POC CRP testing, participants from England and Italy were asked to provide estimates for the time from the initial consultation to availability of CRP laboratory test.
They were additionally asked to estimate how their practice would change if they used POC testing.
The pathways reported by different interviewees were assessed and key similarities and differences were used to synthesise and validate the expected impact of the POC CRP test in different settings.
Cost inputs
Only direct costs to the healthcare provider were considered. Inputs for the micro-costing were derived from various sources (Table 1 and Supplementary Tables S2-S9), including clinician interviews, national reference prices, published literature, medical supply retailers and other online sources. All costs are presented in 2022 GBP (for England) or EUR (for Norway, the Netherlands and Italy). Costs were inflated using the Hospital & Community Health Services (HCSH) index [32] and the NHS Cost Inflation Index (NHSCII) [33] in England and national consumer price indices for Norway, Italy and the Netherlands [34] where necessary. Value-added tax (VAT) was included where applicable.
Clinicians were asked to provide estimates for local costs and resource use. Where available, these estimates were used and validated against published literature to identify responses inconsistent with previous evidence. If clinician estimates were not available, estimates were used from published literature. Assumptions were made as necessary to complete the pathways and cost calculations (Supplementary Tables S2-S9).
Salary costs, additional costs of employment and overheads were included for England [33]. For Italy, Norway and the Netherlands, only direct salary costs were included, as overhead estimates were unavailable.
Cost of hospital referrals (Base case and scenario analysis)
The cost of hospital referral was incorporated into the average pathway cost calculations based on the proportion of patients being referred in the pathway. No data were found from published literature or given by clinicians during the interview to inform what proportion of patients referred to the hospital attend for planned diagnostic testing or the emergency department (ED) and what proportion is admitted. As the base case, the cost of hospital referral was assumed conservatively to be the cost of admission for RTI patients for one day and/or night (Table 1). In reality, this may vary considerably. The proportion of hospital admissions after referral would be considerably longer than one day. Conversely, some patients may not be admitted and receive diagnostic testing as an outpatient visit. Due to this uncertainty around the costs of hospital admissions, two alternative scenarios were included: a) hospital costs set as zero, and b) hospital costs based on six nights per patient in hospital, the average stay in hospital for RTI patients identified from previous literature [35].
Table 1
Sources of data used in micro-costed pathways; for more detail, see Supplementary Tables S2-S9.
Country
|
Staff costs
|
Consumable costs
|
Drug costs
|
Additional healthcare costs
|
Staff time
|
Patient flow
|
Hospital referral cost
|
England
|
NHS Reference Costs 2019/20 [1]
|
Medical supplier [2, 3]
|
NHS Prescription Cost Analysis 2021/22 [4]
|
Hunter, 2015 [5]
|
Interview
|
Interview
|
National Schedule of NHS Costs 2020/21 [6]
|
Netherlands
|
Interview, salary websites [7–9]
|
Medical supplier [10]
|
Zorginstituut Nederland, 2022 [11]
|
No additional costs used
|
Interview
|
Interview
|
Kanters et al., 2017 [12]
|
Italy
|
Interview, salary websites [13]
|
Medical supplier [14]
|
Italian Medicines Agency, 2022 [15]
|
Interview
|
Interview
|
Interview
|
Tichy et al., 2020 [16]
|
Norway
|
Legeforeningen Normaltariffen 2022; salary websites [9, 17]
|
Medical supplier [18, 19]
|
Legemiddelvisning Legemiddelverket, 2022 [20]
|
Legeforeningen Normaltariffen 2022 [17]
|
Interview
|
Interview
|
Postma et al., 2015; Helsedirektoratet Innsatsstyrt finansiering (ISF) 2022 [21, 22]
|