Table 4

Patient characteristics and call characteristics associated with clinically relevant undertriage and overtriage in high-risk calls

Undertriage*Overtriage†
N=32N=24
% (N)RR (95% CI)% (N)RR (95% CI)
Patient characteristics
 Age groups (in years)
  30–59 (n=272)6.3 (17)0.55 (0.29 to 1.07)7.4 (20)2.44 (0.85 to 7.01)
  >60 (n=133)11.3 (15)13.0 (4)1
 Sex
  Male (n=163)8.6 (14)16.8 (11)1
  Female (n=242)7.4 (18)0.87 (0.44 to 1.69)5.4 (13)0.80 (0.40 to 1.73)
Call characteristics
 Weekend
  Not weekend (n=208)6.7 (14)16.3 (13)1
  Weekend (n=197)9.1 (18)1.36 (0.69 to 2.65)5.6 (11)0.89 (0.41 to 1.95)
 Time of day
  Day or evening (n=314)6.4 (20)16.1 (19)1
  Nighttime (n=91)13.2 (12)2.1 (1.05 to 4.07)*5.5 (5)0.91 (0.35 to 2.36)
 Triage model
  GP-led triage (n=206)10.7 (22)12.4 (5)1
  Nurse-led triage with CDSS (n=199)5.0 (10)0.47 (0.23 to 0.97)‡10.0 (19)3.93 (1.50 to 10.33)‡
  • We excluded 30 calls as assessing accuracy of triage was not possible or not relevant.

  • *Undertriage: triage decision assessed as clinically relevant undertriage (rated ‘1’ or ‘2’).

  • †Overtriage: clinically relevant overtriage (rated ‘6’ or ‘7’).

  • ‡Significant difference in RR, using binomial regression analyses; p<0.05: Definitions: Weekend=Friday 16:00 to Monday 8:00 and holidays. Day=8:00 to 16:00, Evening=16:00 to midnight, Night=midnight to 8:00.

  • CDSS, computerised decision support system; GP, general practitioner.