Research

Diagnostic accuracy of Fever-PAIN and Centor criteria for bacterial throat infection in adults with sore throat: a secondary analysis of a randomised controlled trial

Anna Seeley, Thomas Fanshawe, Merryn Voysey, Alastair Hay, Michael Moore and Gail Hayward
BJGP Open 2021; 5 (6): BJGPO.2021.0122. DOI: https://doi.org/10.3399/BJGPO.2021.0122

Article Figures & Data

Figures

  • Figure 1.
    Figure 1. Proportion with throat swabs positive for streptococcal † infection. Proportion of those with streptococcal infection within each clinical prediction score, with 95% confidence intervals represented by black bars. Lines on the graph added for thresholds of Centor criteria ≥3 and FeverPAIN ≥4 indicating high probability of infection, where NICE supports consideration of antibiotic prescription. Data below the graph indicate the number of participants at each score, and number who were throat-swab positive for group A, C, or G streptococci. All beta-Lancefield group streptococci to include group A, C, and G streptococci strains isolated on agar plates after 48 hours culture.
  • Figure 2.
    Figure 2. Receiver operating characteristic (ROC) curves for diagnostic accuracy in identifying streptococcal infections for FeverPAIN and Centor clinical prediction rules. All beta-Lancefield group streptococci to include group A, C, and G streptococci strains isolated on agar plates after 48 hours culture. ROC curve analysis showing sensitivity versus 1-specificity each threshold. The different scores are shown next to a point on the graph. Each colour represents the diagnostic accuracy of one clinical prediction rule for streptococcal infection in adults with sore throat.
  • Figure 3.
    Figure 3. Implications of diagnostic accuracy of FeverPAIN and Centor CPR in diagnosing streptococcal throat infections in 100 adults presenting with sore throat not deemed to require immediate antibiotics. The diagram illustrates how both CPR scores would operate if used on 100 people presenting with sore throat, based on a similar prevalence (16%) of streptococcal infection, as found in the present study. Different colours indicate possible outcomes: true positive (blue), false positive (yellow), false negative (red), and true negative (grey).
  • Figure 4.
    Figure 4. Delayed antibiotics prescribed in TOAST according to each clinical preduction rule score. The graph displays the percentage of participants in each clinical decision score who were prescribed delayed antibiotics at presentation. The numbers within each bar indicate the number of participants who received a delayed prescription at each score.

Tables

  • Table 1. Throat swab culture results
    All beta-Lancefield group streptococciaGroup A streptococciGroup C and G streptococci
    Negative swabPositive swabPNegative swabPositive swabPNegative swabPositive swabP
    n, %(n = 421)n, %(n = 81)n, %(n = 439)n, %(n = 63)n, %(n = 484)n, %(n = 18)
    Centor ≥3(n = 74)55 (13%)19 (24%)0.0159 (13%)15 (24%)0.0377 (16%)4 (22%)0.33
    FeverPAIN 2–3(n = 266)192 (46%)48 (60%)0.02205 (47%)35 (56%)0.16253 (52%)13 (72%)0.03
    FeverPAIN ≥4(n = 11)8 (2%)3 (3.7%)0.159 (2%)2 (3%)0.4210 (2%)1 (6%)0.09

    • ↵a All beta-Lancefield group streptococci to include group A, C, and G streptococci strains isolated on agar plates after 48 hours culture.

  • Table 2. Diagnostic accuracy at each threshold of clinical prediction rule
    Threshold n above / belowSensitivity(95% CI)Specificity(95% CI)ROC(95% CI)PPV(95% CI)NPV(95% CI)
    Centor ≥1424/7890.1 (81.5 to 95)16.6 (13.2 to 20.5)0.53 (0.5 to 0.57)17.2 (13.7 to 21.2)89.7 (80.8 to 95.5)
    Centor ≥2211/29861.7 (50.3 to 72.3)61.8 (56.9 to 66.4)0.62 (0.6 to 0.68)23.7 (18.1 to 30)89.3 (85.2 to 92.6)
    Centor ≥374/42823.5 (14.8 to 34.2)86.9 (83.3 to 90)0.55 (0.50 to 0.60)25.7 (16.2 to 37.2)85.5 (81.8 to 88.7)
    Centor ≥412/4903.7 (0.77 to 10.4)97.9 (96 to 99)0.51 (0.49 to 0.53)25 (5.49 to 57.2)84.1 (80.5 to 87.2)
    FeverPAIN ≥1434/6893.8 (86.2 to 98)15 (11.7 to 18.7)0.54 (0.51 to 0.58)17.5 (14.1 to 21.4)92.6 (83.7 to 97.6)
    FeverPAIN ≥2251/25163 (51.5 to 73.4)52.5 (47.6 to 57.4)0.58 (0.52 to 0.64)20.3 (15.5 to 25.8)88 (83.4 to 91.8)
    FeverPAIN ≥393/40922.2 (13.7 to 32.8)82.2 (78.2 to 85.7)0.52 (0.48 to 0.57)19.4 (11.9 to 28.9)84.6 (80.7 to 88)
    FeverPAIN ≥411/4913.7 (0.77 to 10.4)98.1 (96.3 to 99.2)0.51 (0.49 to 0.53)27.3 (6.02 to 61)84.1 (80.6 to 87.2)
    FeverPAIN ≥52/5001.29 (0.031 to 6.69)99.8 (98.7 to 100)0.51 (0.49 to 0.52)50 (1.26 to 98.7)84 (80.5 to 87.1)

    • ROC = area under receiver operating characteristics curve. PPV = positive predictive value. NPV = negative predictive value

Supplementary Data

  • BJGPO.2021.0122_Supp.pdf -

    Supplementary material is not copyedited or typeset, and is published as supplied by the author(s). The author(s) retain(s) responsibility for its accuracy.
Back to top

In this issue

BJGP Open
Vol. 5, Issue 6
January 2021
Download PDF
Download PowerPoint
Email Article
Citation Tools
Diagnostic accuracy of Fever-PAIN and Centor criteria for bacterial throat infection in adults with sore throat: a secondary analysis of a randomised controlled trial
Anna Seeley, Thomas Fanshawe, Merryn Voysey, Alastair Hay, Michael Moore, Gail Hayward
BJGP Open 2021; 5 (6): BJGPO.2021.0122. DOI: 10.3399/BJGPO.2021.0122
del.icio.us logo Facebook logo Mendeley logo Bluesky logo

Jump to section

Keywords

Intended for Healthcare Professionals
 
 

British Journal of General Practice