Table 1. Summary of Studies on the Cost-Effectiveness of Screening for Breast Cancer after Age 65

Author, Year, Reference Time Horizon Type of Model Interval Mammography Effectiveness Sensitivity Specificity Costsa
Screen  Diagnosis  Treatment 		Utility Discount Rate Cost-Effectiveness Ratiob
Messecar, 200030 Lifetime Markov Biennial Based on
SEER stage distributionc		 95% 95% $118 $1,294 $40,475 .8 local
.26 mets		 5% 3.3 days saved for screening ages 75 -79 (vs ages 65-74) healthy  women; 1.5 days saved for women with dementia; cannot abstract CE ratio
Rosenquist and Lindfors, 199834 ages 40-79 Markov Annual ages 40-49, biennial ages 50-79 39% reduction in mortality with biennial for 50+; 13% for 40-49 yrs Not stated Not stated $72 $1,116 $7,991 (surgery only) None 3% $22,794-$27,248 average CE of screening for ages 50-79d
Lindfors and Rosenquist, 199525 ages 40-79 Markov Annual ages 40-49, biennial ages 50-79 Mortality reduction varies by age; 4%-23% for ages 40-49; 23%-32% for ages 60-79 Not stated Not stated $110 $1,116 $7,991 (surgery only) None 5% $50,131 for biennial at ages 65-79 (approx vs stopping at age 59)e
Brown, 199210 20 yrs starting at age 50 CANTROL Markov process Biennial Observed from RCTs ~30% reduction in mortality Not stated 98.6% $99 $2,520 Medicare costs:
$21,287 local;
$30,714 regional;
$30,714 distant; $63,455 terminal care 		None 5% $50,400 for ages 70-75 vs ages 65-70; $54,000 for ages 75-80 vs ages 70-75
Boer et al, 199831 Lifetime MISCAN Biennial; examines triennial Observed from RCTs ~30% reduction in mortality Varies by lesion size: 40% DCIS; 65% T1a; 80% T1b; 90% T1c; 95% >T2 Not stated $66 National Health Service costs $34,860 advanced stage None 6% $5,910 for ages 65-69 vs stopping at age 64
Boer et al, 199933 Lifetime MISCAN Biennial; examines annual and triennial Observed from RCTs ~30% reduction in mortality Same as Boer et al, 199831 Not stated $66 National Health Service costs $34,860 advanced stage Surgery .89-.93; tam .82; regional .63; mets .29 6% $48,433 for ages 65-94 vs ages 50-64
de Koning et al, 199122 1990-2017 MISCAN Biennial Observed from RCTs ~30% reduction in mortality Same as Boer et al, 199831 Not stated $66 National Health Service costs $34,860 advanced stage None 5% $13,280 for ages 71-75 vs ages 65-70
Eddy, 198926 10 yrs CANTROL Markov process Annual Unknown Not stated 98.6% $194 Medicare costs:
$21,287 local;
$30,714 regional;
$30,714 distant;
$63,455 terminal care 		Medicare costs:
$21,287 local;
$30,714 regional;
$30,714 distant; $63,455 terminal care 		None 5% $34,188-$86,614 for screening for ages 65-75
Kerlikowske et al, 199927 Lifetime Markov Biannual 27% reduction in mortality (22%-32%); assume benefits continue for 5 yrs after cessation of screening Not stated Not stated $108-$138 $451 Kaiser HMO costs: $31,258 DCIS; $45,220 None in base case; tested range in sensitivity analysis 3% $87,887 for ages 70-79 vs stopping at age 69
Mandelblatt et al, 199232 Cross section, 1 point in time Markov 1 point in annual program Based on
SEER stage distributionc 		75% 90% $146 N/A N/A None in base case; tested range in sensitivity analysis: local .9; regional .8; distant .5; short-term false-positive .10 None Varies by age and health group

CANTROL: a computer program to calculate outcomes (and costs); CE: cost-effectiveness; DCIS: IC ductal carcinoma in-situ; HMO: health maintenance organization; Markov Model: a type of simulation program with recurring defined health states used to portray disease process; Mets: distant metastatic disease (i.e., distant stage); MISCAN: microsimulation of cancer (a Monte Carlo simulation approach); RCT: randomized controlled trials; SEER: Surveillance, Epidemiology, and End Results; Tam: tamoxifen; T1a,b,c: tumor size

a Year 2002 dollars based on the consumer price index.
b Incremental costs per additional life-year, compared to screening until age 65, unless otherwise specified.
cEffectiveness based on stage is estimated by comparing stage distribution in the absence of screening to more favorable stage distribution in the presence of screening.
dRatio includes annual screening from ages 40-49, so overestimates results for women over age 50.  Note that results are average results over the age range and do not allow separation of data for extending screening after age 65.
eAnalysis includes annual screening from ages 40-64, then biennial from ages 65-79, compared to biennial screening for ages 50-59, so incremental ratio includes costs of starting earlier and extending screening to age 79.

Return to Document