Table 3. Differences in length of diagnostic interval in calendar days at the 50th and 90th percentilesa for patients with an index consultation in general practice. Patients with missing diagnostic interval were omitted (n = 53) (n = 416)
CrudeAdjustedbCrudeAdjustedb
n 50thDiff(CI 95%)50thDiff(CI 95%)90th Diff(CI 95%)90thDiff(CI 95%)
Index consultation assessment
 No suspicion12783ref92ref230ref151ref
 Suspected cancer19731 –52 (–66 to –38)32 –60 (–69 to –51)218 –12 (–22 to –1)1510(–11 to 10)
 Suspected other serious disease9236 –47 (–53 to –41)42 –50 (–66 to –34)199 –31 (–39 to –22)116 –35 (–39 to –31)
Cancer types
 Colorectal8254ref53ref238ref234ref
 Breast19345 –9 (–17 to –1)43–10(–36 to 15)209 –29 (–39 to –19)219–15(–30 to 0)
 Lung5452–2(–10 to 7)530(–13 to 14)185 –53 (–59 to –46)174 –60 (–78 to –43)
 Melanoma4649–5(–10 to 0)47–6(–24 to 12)347 109 (102 to 115)281 47 (32 to 62)
 Bladder1930 –24 (–30 to –17)31 –22 (–41 to –4)175 –63 (–68 to –57)137 –97 (–112 to –82)
 Endometrial1644 –10 (–15 to –5)44–9(–21 to 2)155 –83 (–88 to –77)153 –81 (–97 to –64)
 Ovarian668 14 (9 to 19)6916(–11 to 43)173 –65 (–70 to –60)2373(–18 to 23)
GP actionsc
 Watchful waiting49100 56 (44 to 67)74 37 (32 to 43)283 72 (60 to 84)226 25 (15 to 35)
 Await follow-upd 2978 26 (19 to 32)62 20 (6 to 33)185 –14 (–23 to –6)98 –20 (–25 to –14)
 Tests in general practicee 128537(–7 to 20)5210(–1 to 21)214–1(–12 to 11)169 25 (19 to 30)
 Confer with specialistf 4136 –13 (–17 to –9)37–8(–21 to 5)140 –77 (–89 to –65)134 –61 (–75 to –47)
 Referral25514 –24 (–34 to –14)14 –16 (–27 to –5)201–8(–20 to 4)126 –18 (–24 to –12)
Referral modality
 Cancer patient pathway12323ref19ref181ref204ref
 Specialist5464 41 (33 to 49)5233(–22 to 88)209 28 (21 to 36)243 39 (3 to 76)
 Imaging or other7854 31 (27 to 36)55 36 (13 to 58)279 98 (89 to 106)231 27 (20 to 34)
  • aPseudo-centiles were used for reporting of 50th and 90th percentile diagnostic interval length of references; for non-references the total diagnostic interval lengths are calculated from adding the point estimate to the reference value. bIndex consultation assessment adjusted for sex, age, follow-up status, comorbidity, and educational level. Cancer types adjusted for age and educational level. GP actions and referral modality were adjusted for age, sex, comorbidity level, follow-up status, and educational level. cSeparate analyses were conducted for each action with the reference group comprising all patients without that particular action (n = 416). dAmong patients in active follow-up (n = 235). eSuch as blood tests, urinalysis, spirometry. fNon-referral contact by phone or letter to specialist. Bold indicates statistical significance, P≤0.05. Diff = difference.