Public Health Agency of Canada
www.publichealth.gc.ca
Common menu bar links
Institutional links
Diseases & Conditions
Health & Safety
Research & Statistics
Agency Information
Search Box
E-mail this page
Report from the Evaluation Indicators Working Group: Guidelines for Monitoring Breast Screening Program Performance
[Previous] [Table of Contents] [Next]
Program Performance Measures
Participation Rate
| Definition: | Percentage of women who have a screening mammogram (calculated biennially) as a proportion of the eligible population. | ||||
| Context: | In order for a screening program to reduce
mortality in a population, that population must participate in the
program in sufficient numbers. A participation rate of 70% and over
was achieved in trials reporting substantial mortality
reductions. Note that program participation rate does not represent all breast cancer screening in Canada. In most provinces screening can also be delivered outside the structure of the program. |
||||
| Calculations: |
|
||||
| Details: | In the case of multiple screens, age at the
first screen is the criterion used to determine whether the woman
was in the target population. Target population (denominator) should be obtained from the most recent census results and/or forecasts of population available from Statistics Canada. |
||||
| Targets: | Canada ³70% of the eligible population (ages 50-69) Europe3 60% (ages 50-64) United Kingdom5 70% of women invited (ages 50-64) Australia7 ages 50-69) |
||||
| Status: | 34.3% of Canadian women (age 50-69, 1997-98)
received a program screen8. Note: From the results of the National Population Health Survey it is estimated that 53.6% of Canadian women received a (program or non-program) mammogram in the 1996/97 2-year period14. |
||||
| Evidence: | Based on basic principles of population
screening15,16. Extrapolation from the results of randomized controlled trials17,18. |
Retention Rate
| Definition: | The estimated percentage of women who are re-screened within 30 months of their previous screen. |
| Context: | Optimal benefits of screening are brought about by regular participation in the screening program (at least every 2 years). At present there is no indication that the benefits of screening are lost if re-screening occurs up to 6 months after the recommended interval (i.e., 30 month interval). |
| Calculations: | Actuarial Method for Survival Data st= 1 - (p0p1p2...pt) where pt = 1 - qt qt = et/ n*t n*t = nt . ½ ct st is the estimated cumulative probability of returning from baseline to the end of the study interval that begins at t; pt is the estimated probability of not returning during the study interval that begins at time t; qt is the estimated probability of women returning during the study interval that begins at time t; et is the number of women returning in the study interval that begins at time t; nt is the number of women present at the beginning of the study interval that begins at time t; ct is the number censored (because of death, breast cancer, or age limit-68 years) during the interval that begins at time t. |
| Targets: | Canada ³75% re-screened within 30 months (age 50-69) Australia7 75% screened in the previous round (age 50-69); of those rescreened, > 90% to be screened biennially within 27 months. |
| Status: | 79.0% of women (age 50-59, 1997-98) re-screened within 30 months of previous screen 65.0% of women (age 60-69, 1997-98) re-screened within 30 months of previous screen8. |
| Evidence: | Related to participation rate, sojourn time, screening interval studies19, and randomized controlled trials17,18. |
Abnormal Call Rate (%)
| Definition: | Percentage of women screened who are referred for further testing because of abnormalities found with a program screen. | |||||||||
| Context: | Abnormal call rate is a meaningful indicator when considered in the context of positive predictive value and cancer detection rate. Also, relative to the underlying breast cancer incidence rate, it is an indicator of the quality of mammographic image or interpretation. Abnormal call rate will generally be higher for initial screens (which detect prevalent cancers) than for re-screens. | |||||||||
| Calculations: |
|
|||||||||
| Targets: | Canada < 10% (initial screen); < 5% (re-screens) (age 50-69) Sweden2 9% (overall) Europe3 < 7% (initial screen); < 5% (subsequent screens) United Kingdom6 < 7% (initial screen); < 7% (subsequent screens) Australia7 < 10% (initial screen); < 5% (subsequent screens) |
|||||||||
| Status: | Abnormal recall rates by mode of
detection, 1997 and 1998 screen years8
|
|||||||||
| Evidence: | Measured in randomized controlled trial17. | |||||||||
Invasive Cancer Detection Rate
| Definition: | Number of women detected with invasive cancer during a screening episode per 1,000 women screened. | |||||||||
| Context: | Cancer detection rate is meaningful for program evaluation when considered in relation to the abnormal call rate, post-screen cancer detection rate, and the underlying breast cancer incidence rate. The cancer detection rate in an organized screening program should generally exceed the cancer incidence rate in the population prior to organized screening, because some cancers would remain asymptomatic in the absence of screening. Cancer detection rates will generally be higher for initial screens (which detect prevalent cancers) than for re-screens. | |||||||||
| Calculations: |
|
|||||||||
| Targets: | Canada > 5 per 1,000 on initial screen (age 50-69) > 3 per 1,000 on re-screens (age 50-69) United Kingdom6 > 3.6 per 1,000 first screen (age 50-64); > 4.0 per 1,000 subsequent screens (age 50-64) Australia7 > 5 per 1,000 first screen; > 2 per 1,000 subsequent screens (age 50-69) |
|||||||||
| Status: | Invasive cancer detection rates per
1,000 screens by mode of detection 1997 and 1998 screen
years8
|
|||||||||
| Evidence: | Based on randomized controlled trials17,18 and the experience of other breast cancer screening programs5,7. | |||||||||
In Situ Cancer Detection Rate
| Definition: | Number of women detected with ductal carcinoma in situ (DCIS) cancer (rather than invasive cancer) during a screening episode per 1,000 women screened. | ||||
| Context: | While there is no definitive link between in situ cancer and invasive cancer, in situ cancer detection may be interpreted as an indicator of screening quality when considered in relation to the cancer detection rate and underlying cancer incidence rate. | ||||
| Calculations: |
|
||||
| Targets: | Canada Surveillance and Monitoring Purposes Only United Kingdom5 0.4-0.9 per 1,000 (initial screen) 0.5-1.0 per 1,000 (subsequent screens) Australia7 10%-20% of cancers detected are in situ |
||||
| Status: | 0.9 per 1,000 in situ cancer (aged 50-69, 1997-98)8 | ||||
| Evidence: | It seems inappropriate to set targets for DCIS given the current paucity of evidence concerning the transition of DCIS to invasive cancer and the continually increasing sensitivity of screening techniques20. |
Diagnostic Interval
| Definition: | Total duration from abnormal screen to resolution of abnormal screen. | ||||
| Context: | An abnormal screen result can induce
morbidity, given the negative psychological impact it can have on a
client, even if follow-up is ultimately negative. Moreover,
excessive delay to diagnosis may worsen prognosis. Work-up should
therefore be completed expeditiously21. Note that some Canadian programs do not have integrated diagnostic capabilities, making measurement of diagnostic interval more difficult. |
||||
| Calculations: | (date of diagnosis) - (screen date) =
Diagnostic Interval
|
||||
| Targets: | Canada ³90% within 5 weeks if no open biopsy (age 50-69) ³90% within 7 weeks if open biopsy (age 50-69) Australia7 90% to have first assessment within 10 working days 70% to be provided with definitive diagnosis or recommendation for biopsy within 2 working days of first assessment. |
||||
| Status: | 74.5% within 5 weeks (no open biopsy, age
50-69, 1997-98) 44.9% within 7 weeks (open biopsy, age 50-69, 1997-98) |
||||
| Evidence: | Based on basic principles of screening15,16 and screening program evaluation research22. |
Positive Predictive Value
| Definition: | Proportion of abnormal cases with completed follow-up found to have breast cancer (invasive or in situ) after diagnostic work-up. | |||||||||
| Context: | Positive predictive value (PPV) is an indicator of the predictive validity of screening. The factors that influence cancer detection rate and abnormal call rate must also be taken into consideration when evaluating a program's PPV. PPV tends to improve with re-screening because the initial screen establishes a normal baseline. Consequently, PPV tends to be lower among initial screens relative to re-screens. | |||||||||
| Calculations: |
Note: Includes invasive and in situ cancers |
|||||||||
| Targets: | Canada ³5% (initial screen) (age 50-69) ³6% (re-screen) (age 50-69) |
|||||||||
| Status: | Positive predictive value by mode of
detection, 1997 and 1998 screen years8
|
|||||||||
| Evidence: | Based on methodology in screening program evaluation studies23. | |||||||||
Benign to Malignant Open Biopsy Ratio
| Definition: | Among open biopsies, the ratio of number of benign cases to the number of malignant cancer cases. | ||||
| Context: | Benign to malignant open biopsy ratios provide an indication of the quality of the presurgical assessment. Diagnostic specificity and sensitivity are reciprocal. Consequently there is a limit to the extent to which biopsy yield ratios can be improved. This indicator is most meaningful when considered in relation to the underlying breast cancer incidence rate and the post-screen detected cancer rate. | ||||
| Calculations: |
Note: Each open biopsy performed represents a case. It may be useful to present these figures with confidence intervals when small numbers of cases are observed. |
||||
| Targets: | Canada 2:1 open (initial & re-screen combined) (age 50-69) Sweden2 3:1 (first & re-screen combined) Europe3 2:1 (first screen), . 1:1 (re-screen) United Kingdom5 3:1 (first & re-screen combined) Australia7 2:1 (first screen), . 1:1 (re-screen) |
||||
| Status: | 1.6:1 benign to malignant open biopsy ratio (age 50-69, 1997-98) | ||||
| Evidence: | The targets are based on experience from research trials (e.g., Swedish Two County study)8. |
Invasive Cancer Tumour Size
| Definition: | Percentage of invasive cancers with tumour
size of 10 mm in greatest diameter as determined by the best
available evidence:
|
||||
| Context: | Invasive tumour size is the best known prognostic indicator. The purpose of mammographic screening is to detect pre-clinical cancers before symptoms are apparent. | ||||
| Calculations: |
Note: Each open biopsy performed represents a case. It may be useful to present these figures with confidence intervals when small numbers of cases are observed. |
||||
| Targets: | Canada 25% 10 mm (age 50-69) Sweden2 > 50% < 15 mm Europe3 20% 10 mm (initial screen) 25% 10 mm (subsequent screens) United Kingdom5 1.5 per 1,000 (< 15 mm; initial screen) 1.65 per 1,000 (< 15 mm; subsequent screens) Australia7 > 8 per 10,000 (10 mm) |
||||
| Status: | 37.6 % of tumours 10 mm (age 50-69, 1997-98)8 | ||||
| Evidence: | Stage-specific prospective studies and trials2,3,7. |
Positive Lymph Nodes in Cases of Invasive Cancer
| Definition: | Proportion of invasive cancers in which the cancer has invaded the lymph nodes. | ||||
| Context: | Positive lymph nodes are a prognostic indicator. The purpose of mammographic screening is to detect breast cancer as early as possible - before it spreads to the lymph nodes. | ||||
| Calculations: |
Note: Excludes cases in which lymph nodes are not assessed. |
||||
| Targets: | Canada < 30% node positive (age 50-69) Sweden2 > 70% node negative (age 50-64) |
||||
| Status: | 21.7% node positive in assessed cases of invasive cancer (age 50-69, 1997-98)8 | ||||
| Evidence: | Stage-specific prospective studies and trials2,3,7. |
Post-screen Detected Invasive Cancer Rate
| Definition: | Number of women with a diagnosis of invasive breast cancer after a negative screening episode per 10,000 person-years at risk, within 12 AND 24 months of the screen date. | ||||||||
| Context: | Post-screen detected invasive cancer rate is an indicator of the sensitivity of the screening program. This rate is affected by population incidence, age, rate of disease progression, and screening interval recommendation. A high rate may negatively affect the mortality reduction expected for a successful, organized screening program. The accuracy of this measure is dependent on the completeness of cancer registration. | ||||||||
| Calculations: |
Note: Calculations include all cases regardless of screening interval recommendation. |
||||||||
| Targets: | Canada < 6 per 10,000 person-years (within 12 months, age 50-69) < 12 per 10,000 person-years (within 24 months, age 50-69) United Kingdom5 12 per 10,000 (within 24 months) Australia24 < 6.5 per 10,000 (within 12 months) |
||||||||
| Status: | Cancers detected outside of program
after normal screening episode among program participants aged
50-69 at screening, 1994 and 1995 screen
years8
|
||||||||
| Evidence: | Studies of interval cancer25,26. |
[Previous] [Table of Contents] [Next]
