Table 2.

Association between demographics and presentation–test intervals using negative binomial regressiona

Patients in cohortPresentation–test interval, days
n (%)Median (IQR)IRR (adjusted)b95% CI (P-value)
Total2102 (100)49 (5–172)
Constant35.7723.09 to 55.40 (<0.001)
Sex
  Male1148 (55)41 (4–162)Ref
  Female954 (45)63 (7–182)1.121.02 to 1.24 (0.016)
Age
  For every additional1.071.01 to 1.12 (0.013)
  10 years from age 27
Deprivation quintile
  1 (least deprived)367 (17)43 (4–162)Ref
  2370 (18)37 (3–147)0.860.73 to 1.02 (0.081)
  3421 (20)46 (6–173)1.010.86 to 1.18 (0.927)
  4438 (21)49 (6–184)1.030.88 to 1.21 (0.704)
  5 (most deprived)506 (24)73 (9–184)1.080.93 to 1.25 (0.326)
Smoking status
  Non-smoker149 (7)18 (1–85)RefRef
  Ex-smoker1168 (56)52 (6–173)1.581.26 to 1.99 (<0.001)
  Smoker785 (37)55 (6–181)1.631.29 to 2.05 (<0.001)
Number of features on first presentation
  11977 (94)48 (5–170)RefRef
  >1125 (6)91 (10–190)1.201.00 to 1.43 (0.052)
ln (alpha)0.820.77 to 0.88
  • a Likelihood ratio (LR) test of alpha = 0, χ2 = 240 000, P < 0.001.

  • b Variables included in the adjusted analysis include sex, age in years, level of deprivation, smoking status, and number of features at first presentation. IQR = interquartile range. IRR = incidence rate ratio. ln (alpha) = natural log of alpha. Alpha is the over-dispersion parameter (values >0 indicate the distribution’s variance is greater than its mean).