Background: Despite increasingly widespread use of the Papanicolaou smear,
almost half of all women with invasive cervical cancer are diagnosed at a l
ate stage (regional or distant). Little is known about factors associated w
ith late-stage diagnosis of cervical cancer.
Objective: To examine the relationship of age, race, education level, incom
e level, smoking, marital status, health insurance type, comorbidity, and r
esidence in an urban or rural setting to late stage at diagnosis of cervica
l cancer.
Methods: Incident cases of invasive cervical cancer occurring in 1994 in Fl
orida were identified from the state tumor registry (N = 852). Cases were l
inked with state dis charge abstracts and the 1990 US census. Multiple logi
stic regression was used to determine the relationship between predictor va
riables (age, race or ethnicity, marital status, smoking status, education
level, income level, insurance type, comorbidity, and urban vs rural reside
nce) and the odds of late-stage diagnosis.
Results: Age, marital status, and insurance type were associated with late-
stage diagnosis. Each additional year of age was associated with a 3% incre
ased odds of late-stage diagnosis (odds ratio [OR], 1.03; 95% confidence in
terval [CI], 1.02-1.05; P<.001). Being unmarried was associated with a 63%
increased odds of late-stage diagnosis (OR, 1.63; 95% CI, 1.18-2.25; P=.003
). Being uninsured was associated with a 60% increased odds of late-stage d
iagnosis (OR, 1.60; 95% CI, 1.01-2.38, P=.02). Having commercial health mai
ntenance organization insurance was associated with a 46% decreased odds of
late-stage disease (OR, 0.54; 95% CI, 0.30-0.96; P=.04). Race, education l
evel, income level, smoking status, comorbidity, and urban residence were n
ot associated with stage at diagnosis.
Conclusions: Women with cervical cancer who are elderly, unmarried, and uni
nsured are more likely to be diagnosed at a late stage. These women should
be targeted for cervical cancer education and screening programs.