QuikRead go for C-reactive protein testing in primary care

Study

Details of intervention and comparator(s)

Outcomes

Strengths and limitations

Little et al. 2013

n=4,264

Cluster randomised factorial controlled trial

Multinational (246 practices from 6 countries)

Belgium, Spain, Wales, England, Poland, Netherlands

CRP testing using QuikRead 101 (n=1,062), enhanced-communication training (n=1,170), CRP testing in combination with enhanced-communication training (n=1,162) compared with standard care (n=870).

The antibiotic prescribing rate was lower with CRP training than without and with enhanced-communication training than without (33% compared with 48%, adjusted risk ratio 0.54, 95% confidence interval 0.42 to 0.69). The combined intervention was associated with the greatest reduction in prescribing rate.

There was a higher admission rate in the CRP group compared with the non-CRP group, but this was of borderline statistical significance when controlled for all potential confounders.

The symptom severities in the CRP group were similar to those in the non-CRP group. Rates of new or worsening symptoms did not differ significantly.

This was a large trial on the QuikRead 101 CRP system, powered to detect a 10% reduction in antibiotic prescribing.

Several practices were excluded before and after randomisation because of recruitment issues.

Cals et al. 2010

n=258

Randomised controlled trial

Multicentre (11 centres)

Netherlands

CRP-assisted prescribing strategy (QuikRead 101 analyser intervention, n=129) compared with standard care (non-CRP-assisted prescribing strategy, control, n=129).

Patients in the CRP-assisted group used fewer antibiotics than those in the control group after the index consultation and during 28‑day follow-up.

There were fewer delayed prescriptions in the CRP-assisted group. Of those with delayed prescriptions, fewer prescriptions were filled in the CRP-assisted group than in the control group.

Recovery was similar across both groups and patient satisfaction with care was higher in the CRP-assisted group.

The study was not powered to detect differences between the 2 groups.

Patients were individually randomised and allocation concealment was done, reducing selection bias.

The authors did a sensitivity analysis and accounted for variation at physician level.

Miravitlles et al. 2013

n=152

Prognostic study

Patients with COPD from placebo arm of randomised controlled trial

Spain

QuikRead 101 analyser with QuikRead CRP assay compared with Anthonisen criteria.

The only factors significantly associated with an increased risk of clinical failure (defined as incomplete resolution, persistence,

or worsening of symptoms that needed a new course of antibiotics, oral corticosteroids, or hospitalisation) without antibiotics were the increase in sputum purulence (included in the Anthonisen criteria) and a CRP concentration ≥ 40mg/litre.

The use of a point-of-care CRP test statistically significantly increased the predictive accuracy of clinical failure.

Only study to evaluate clinical outcomes associated with CRP ≥ 40mg/litre (cut-off determined on the basis of previous analysis).

Authors could not rule out some placebo effect because the criteria for success were based on clinical evaluation.

Minnaard et al. 2015

n=200

Retrospective diagnostic case-control study

Multinational (16 primary care research networks across 12 European countries)

Belgium, Finland, Germany, Hungary, Italy, Netherlands, Norway, Poland, Spain, Slovakia, Sweden, UK

5 point-of-care devices including QuikRead go and QuikRead 101, compared against a laboratory reference standard (Vitros 5.1 FS, Ortho Diagnostics; Dimension Vista Systems, Siemens).

Diagnostic accuracy outcomes were determined from 200 patient blood samples (100 with pneumonia, 100 without pneumonia).

A clinical algorithm was used to determine the incremental predictive power of each CRP test to predict pneumonia.

The sensitivity and specificity of QuikRead go in predicting pneumonia was consistent with other point-of-care systems and a laboratory reference test.

In all cases the discriminatory power of the tests to predict pneumonia were reduced when a higher threshold of CRP (>100mg/litre) was used. Sensitivity was 52% and 20% for low (20 mg/litre) and high (100 mg/litre) thresholds respectively for the detection of radiographic pneumonia, and specificity was 72% and 99%.

Retrospective case-control design subject to inherent bias (likely to overestimate effect).

Radiographs were used as the reference standard.

Samples were analysed retrospectively at a central laboratory.

A diagnostic model was used to evaluate the incremental predictive power of CRP when added to symptoms and signs. This may not completely represent clinical judgement in real life.

Brouwer and van Pelt, 2015

Analytical performance study

Netherlands

8 point-of-care devices including QuikRead go, compared with a comparative laboratory method (Synchron CRP).

All blood samples were from GPs' patients, aged over 18 years, with CRP concentrations ranging from 5 to 200 mg/litre, determined with the Synchron analyser.

The linear regression equation of QuikRead go revealed an underestimation of CRP values compared with the laboratory method.

The coefficient of variation of QuikRead go met the criteria of <10%.

The correlation of the CRP tests to the reference standard varied considerably.

Patient blood samples were used with CRP values determined from an appropriate reference standard.

Patient characteristics and diagnosis were not reported.

Minnaard et al. 2013

Analytical performance study

Netherlands

5 point-of-care devices including QuikRead 101 and QuikRead go, compared with a laboratory method (Olympus AU2700).

Residual stored material from routinely done laboratory blood tests was pooled to obtain lithium heparin plasma pools with CRP concentrations of approximately 10, 15, 20, 25, 50, 90, 100 and 110 mg/litre.

The within-day coefficient of variations for low and high CRP concentration samples were smaller for the QuikRead go compared with the QuikRead 101 and other point-of-care devices.

The between-day coefficient of variations for high CRP concentration samples was also lowest with the QuikRead go CRP test.

For high CRP values (>100mg/litre), agreement with the laboratory standard systematically decreased for all 5 point-of-care tests.

The analytical performance and agreement of the CRP tests to the laboratory method varied considerably, but was considered adequate.

Lithium heparin plasma samples were used instead of blood obtained by finger prick.

A separate plasma pool was created to evaluate the QuikRead 101 device.

Abbreviations: COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein.