Research ArticleOriginal Research

A Quantitative Study of the Decision Threshold for the Diagnosis of Infectious Mononucleosis

Xinyan Cai, Mark H. Ebell and Garth Russo
The Journal of the American Board of Family Medicine December 2022, jabfm.2022.210185R1; DOI: https://doi.org/10.3122/jabfm.2022.210185R1

Article Figures & Data

Figures

  • Figure 1.
    Figure 1.

    Illustration of the threshold model.

  • Figure 2.
    Figure 2.

    The distributions of probabilities for clinicians choosing to order a test for Infectious Mononucleosis (IM) and those choosing not to order a test (rule out).

  • Figure 3.
    Figure 3.

    Test (blue solid line) threshold based on the logistic regression model, obtained equaling to 0.5 of the probability of not ruling out Infectious Mononucleosis (IM) (test threshold) estimated according to model 2. Points (circles) represent the true probability of clinicians that decided to rule in IM and to order a diagnostic test for each scenario.

  • Figure A1.
    Figure A1.

    Distribution of the decision probabilities for subgroups: <=10 years’ practice versus >10 years’ practice.

  • Figure A2.
    Figure A2.

    Distribution of the decision probabilities for subgroups: works at student health center versus not works at student health center.

  • Figure A3.
    Figure A3.

    Distribution of the decision probabilities for subgroups: primary care physician versus Nonprimary care physician.

  • Figure A4.
    Figure A4.

    Distribution of the decision probabilities for subgroups: family physician versus nonfamily physician.

  • Figure A5.
    Figure A5.

    Test threshold by subgroups: Primary care physician versus nonprimary care physician.

  • Figure A6.
    Figure A6.

    Test threshold by subgroups: family physician versus nonfamily physician.

  • Figure A7.
    Figure A7.

    Test threshold by subgroups: ≤10 years in practice versus >10 years clinicians.

  • Figure A8.
    Figure A8.

    Test threshold by subgroups: working in a student health center versus not working in student health center.

Tables

  • Table 1.

    Demographic Characteristics of the Participating Clinicians

    Characteristic of clinicians (n = 122)n (%)
    Type of clinical setting 
     Family Medicine102 (83.6%)
     Internal Medicine11 (9%)
     Physician Assistant4 (3.3%)
     Nurse Assistant2 (1.6%)
     Other1 (0.8%)
     No response2 (1.6%)
    Time in practice, years 
     <=524 (19.7%)
     6 to 1010 (8.2%)
     11 to 2041 (33.6%)
     >2047 (38.5%)
    Practice site 
     Primary care101 (82.8%)
     Urgent care7 (5.7%)
     Emergency medicine5 (4.1%)
     Other7 (5.7%)
     No response2 (1.6%)
    Student health clinic setting
     Yes14 (11.5%)
     No108 (88.5%)
    Clinical decision: clinician ruled out IM without ordering a test
     Scenario 1 (1%)113 (92.6%)
     Scenario 2 (4%)103 (84.4%)
     Scenario 3 (7%)67 (54.9%)
     Scenario 4 (12%)28 (23%)
     Scenario 5 (18%)18 (14.8%)
     Scenario 6 (25%)1 (0.8%)
     Scenario 7 (30%)1 (0.8%)
  • Table 2.

    Estimation of Overall Test Threshold and By Subgroups

    Test ThresholdsProbability of IM (95% CI)p-Value
    All participants9.5 (8.2, 10.9) 
    Practice type 0.47
     Primary care9.8 (8.7, 10.9) 
     Non-primary care8.7 (7.1, 10.3) 
    Time in practice, years 0.02
     0 to 107.3 (5.7, 8.5) 
     >1010.5 (9.2, 11.8) 
    Specialty 0.62
     Family physician9.8 (8.5, 11.1) 
     Nonfamily physician9.0 (6.5, 11.5) 
    Student health center 0.59
     Yes9.7 (7.6, 11.3) 
     No10.7 (9.1, 11.8) 

    • Abbreviations: IM, infectious monoucleosis; CI, confidence interval.

    • Notes: The test threshold is estimated based on probability estimation from mix-effect logistic regression models.

  • Table A1.

    The Signs, Symptoms and Probability of (IM) for each clinical vignette

    Case 1Case 2Case 3Case 4Case 5Case 6Case 7
    Age30302418181818
    Days of symptoms106610101010
    Sore throatYesYesYesYesYesYesYes
    RashYesYesYesYesYesYesYes
    CoughYesYesYesYesYesYes
    Sore musclesYesYesYesYesYes
    Sore jointsYesYesYesYesYes
    Sleeping too muchYesYesYesYes
    NauseaYesYesYesYes
    FeverYesYesYes
    HeadacheYesYesYes
    Enlarged tonsils with exudateYesYes
    Posterior cervical adenopathyYes
    Probability of infectious mono1%4%7%12%18%25%30%
  • Table A2.

    Median and Interquartile Range for the Probability of Infectious Mononucleosis (IM) in the Vignettes Stratified by Clinical Decision and Physician Characteristics

    StratificationDecision Probability (%, Median (Interquartile Range))
    Practice TypeRule out (No test)Test
    Primary care4.0 (6.2)17.9 (12.8)
    Non-primary care4.0 (3.0)18.0 (12.6)
    Time in practice, years
     0 to 104.1 (6.3)18.1 (12.6)
     >104.1 (6.1)21.8 (13.6)
    Specialty
     Family physician4.0 (6.0)17.9 (12.8)
     Nonfamily physician3.9 (6.1)20.2 (13.8)
    Student health center
     Yes4.1 (6.1)18.1 (12.7)
     No4.1 (6.0)18.0 (11.7)
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section

Keywords