Mammography screening in Italy: 2004 survey and 2005 preliminary data

Mammography screening in Italy: 2004 surveyand 2005 preliminary data

Daniela Giorgi,� Livia Giordano,� Leonardo Ventura,� Alfonso Frigerio,� Eugenio Paci,�

Marco Zappa�

�UO Epidemiologia, ASL 2 Lucca; �Unità di Epidemiologia dei tumori - CPO Piemonte, Torino; �UO di Epidemiologiaclinica e descrittiva - CSPO, Firenze - Istituto Scientifico della Regione Toscana; �Azienda Sanitaria OspedalieraOIRM S. Anna, Torino

Abstract

In 2004 regions were required to provide the ONS (Osservatorio Nazionale Screening, NationalCentre for Screening Monitoring) with data on screening activity. Previously, GISMa (GruppoItaliano Screening Mammografico) had for several years carried out a yearly survey to collect processindicators of mammography screening and compare them, using standard values agreed on at anational and international level. In 2006, an updated version of the operative report of processindicators was published by GISMa and ONS. Data for the survey came from several differentprogrammes that may have changed over time, and may have different settings of organisation and man-agement. The year 2004 can be considered almost exhaustive of the situation of Italian mammographyscreening; data were provided by 117 programmes from 18 different regions. In ten regions, the individ-ual programmes are part of a comprehensive regional project. In 2004, about 1,870,000 women aged50-69 years were invited to have a screening mammography, and over 1,050,000 were screened.Theoretical extension was 71.9%, while 52% of the target population received an invitation toscreening. An imbalance in coverage is still present when comparing Northern (more than 80%) andCentral Italy (almost 100%) to Southern Italy, that only has a 40% coverage by organised screening.The national crude attendance rate was 55.6%, slightly lower than the previous year. A decreasingtrend was evident from the North to the Centre to the South of Italy: 60.2%, 51.8%, and 39.1%, re-spectively. Referral rates of 8.8% at first screening and 4.2% at repeat screening were recorded. Directstandardised detection rate was 6.6 �1,000 at the first test and 5.0 at repeat test, while benign to ma-lignant ratio for first and repeat screening was 0.34 and 0.20, respectively.Indicators by 5-year age group confirm greater diagnostic problems at younger ages, with higher referralrates, higher frequency of surgical procedures with benign outcome (B/M ratio), and a substantiallylower detection rate as compared to older age groups.Preliminary data for the year 2005 shows a further increase in screening coverage: 76.4% of the Italiantarget population was enrolled in screening programmes, and 50.3% received an invitation to have amammography. Imbalance in coverage still exists between Northern and Central Italy and the South:figures are 92.4%, 98.6%, and 39.3%. The national crude attendance rate was 57.8%.

(Epidemiol Prev 2007; 31(2-3) Suppl 2: 7-20)Keywords: mammography screening, breast, survey, Italy

In 2004, as a consequence of the for-mal agreement between the Ministryof Health and the ONS, based on

which the regions must provide data tothe ONS, a great effort was made by the

ONS to collect screening data in all Ital-ian regions. For several years before that,GISMa had carried out a yearly survey to col-lect, as systematically and thoroughly as pos-sible, process indicators of mammography

anno 31 marzo-giugno 2007 supplemento 2 7

MAMMOGRAPHY SCREENING IN ITALY: 2004 SURVEY AND 2005 PRELIMINARY DATA

screening programmes implemented in Italy.Thanks to the work and collaboration of sev-eral agencies, the surveys gathered data frommost screening programmes, and it was possi-ble to produce regional and national statistics,and to compare different programmes usingstandardised indicators agreed on at a nationallevel. These comparisons represent a precioussource in helping to properly manage pro-grammes, and they provide informationwhich is helpful in making choices to improvescreening quality.In the last few years, the GISMa survey (towhich programmes have always participatedon a voluntary basis) has been consideredhighly representative of the Italian situation,despite its not covering the whole nation.Based on these experiences, a working groupwas set up within GISMa with the aim of re-vising and updating all parameters and stan-dard used for screening programme evalua-tion; the result of the group's work was thepublication of an updated operative report ofprocess indicators,� with the introduction of

new parameters and standards based both onthe results of the national surveys and the pub-lication of the updated European guidelines.�

The inclusion of cancer screening pro-grammes in the Basic Healthcare Parameters(LEA, Prime Ministerial Decree, 29 Novem-ber 2001) and the creation of the ONS hascertainly helped support the implementationand development of organised screeningprogrammes. The ONS, in particular, im-proved scope and systematic nature of the an-nual data collection: thanks to its support,surveys and the publication of the annual re-port on process indicators of the Italianscreening programmes were provided and amore up to date and comprehensive picture ofthe Italian screening activity at a national andregional level is now available.In 2004, data of the survey can be consideredalmost exhaustive of the Italian mammogra-phy screening situation, at least regarding ex-tension of screening programmes.The aim of this report, which is the fifth updateof similar previous publications by the ONS,��

is to provide summary data on the situation ofmammography screening in Italy. With this inmind, data referring to 2004 activity is re-ported stratified by region and 5-year agegroups. Preliminary data about extensionand attendance in 2005 is also reported. Itmust be always borne in mind that this is sum-marised data, coming from several differentprogrammes that may have experienced varia-tions over time, both for different levels of ex-perience and for dissimilar settings of organisa-tion and management. Recommended stan-dards are provided for every indicator at twolevels: minimal (acceptable) and optimal (desir-able). Therefore, when evaluating results thefollowing must be considered:

� not all programmes have the possibility todifferentiate between first and subsequent (re-peat) screening tests: for these programmesresults are assigned to the round that in-cludes the majority of the screened women;

8 anno 31 marzo-giugno 2007 supplemento 2

THE NATIONAL CENTRE FOR SCREENING MONITORING FIFTH REPORT

Region Number of programmes2004 survey

Valle d’Aosta* 1

Piemonte* 9

Liguria 5

Lombardia 14

Trentino 1

Alto Adige 1

Veneto* 17

Emilia-Romagna* 13

Toscana* 12

Umbria 4

Marche 10

Lazio* 12

Abruzzo* 1

Molise* 1

Campania* 12

Basilicata* 1

Calabria 1

Sicilia 2

Total 117

* active programmes within a regional project

Table 1: Distribution by region of active programmes thatprovided 2004 screening data to the GISMa survey.

� a few programmes are not yet able to pro-vide data stratified by 5-year age groups, sothe age-stratified results provided relate to asubset of programmes;

� not all programmes, particularly thosecovering large areas and with several territo-rial screening units, are able to provide com-plete data as to diagnosed cases: thus, somediagnostic results are slightly underesti-mated compared to the actual situation.

As for previous surveys, even in 2004 there wassome trouble in retrieving complete informa-tion on screen-detected cases; based on the in-formation provided by the programmes, data isstill missing for 288 cases (cases with mostlycomplete diagnostic information are around5,200); this number relates to subjects whohave been referred for surgical assessment andfor whom the final surgical outcome is yet un-known. The number of these cases is certainlyunderestimated as these cases are not reportedsystematically by all programmes.

Geographical extensionIn 2004, a further increase in the geographicalextension of mammography screening in Italyoccurred; overall, 117 programmes of 18 re-gions (table 1) contributed to the survey, al-though with different levels of data complete-ness. Compared to the 2003 survey, 29 addi-tional programmes are present, althoughthese programmes were not necessarily newlyimplemented; it is worth noting the emer-gence of regions, such as Marche and Cam-pania, with a large number of programmes,and that Calabria is present for the first time,although with a small programme. In ten re-gions, the single programmes are part of acomprehensive regional project.

In accordance with Italian and Europeanscreening guidelines, most programmes in-vite women in the 50-69 year age range;several programmes have a marginal activ-ity dedicated to women over 70 years of age,offering screening tests only to women at-



Region Theoretical coverageage 50-69 (%)

Women invitedduring 2004 age 50-69

Actual coverageage 50-69 (%)

Valle d’Aosta 100.0 7,299 95.5

Piemonte 100.0 188,101 64.8

Liguria 51.0 65,355 57.1

Lombardia 84.9 419,221 69.7

Trentino 100.0 32,028 110.0*

Alto Adige 100.0 22,082 84.5

Veneto 82.8 188,659 65.5

Emilia-Romagna 100.0 249,656 93.4

NORTH 83.6 1,172,401 68.7

Toscana 100.0 197,941 82.4

Umbria 100.0 51,150 92.6

Marche 73.5 54,687 58.0

Lazio 100.0 168,232 50.2

CENTRE 98.4 472,010 65.1

Abruzzo 44.7 24,807 32.3

Molise 100.0 25,222 133.7*

Campania 78.0 97,449 31.9

Basilicata 100.0 40,065 119.1*

Calabria 4.8 579 0.5

Sicilia 40.2 34,956 12.3

SOUTH 38.3 223,078 19.2

ITALY 71.9 1,867,489 52.0

* the number of invited women is larger than the target population for that year

Table 2: Theoretical and actual coverage of target population by organised screening programmes (age 50-69). Programmes active in 2004.

tending the last invitation. During the pastfew years, some programmes have startedincluding women aged 45-49 (one pro-gramme in Toscana and most of the pro-grammes in the Marche), partly as a conse-quence of a Ministerial Decree offering freebiennial mammography to women aged 45years or more.Table 2 (page 9) shows the number of invitedwomen in 2004 in the 50-69 age range, byregion. During this year, based on a popula-tion of almost 7,200,000 women (Istat dataas of 1/1/2004), target population of activescreening programmes in Italy was over 5million women.About 1,870,000 Italian women were in-vited to have a screening mammography,and over 1,050,000 responded to the invita-tion and were screened within organisedscreening programmes (tables 2 and 3).With extension (coverage) we generally indi-

cate the proportion of women involved in ascreening programme out of the total femalepopulation in the 50-69 age range resident inthe area. Table 2 shows two different data re-lated to extension: “theoretical” and “actual”coverage. Theoretical coverage refers to eligi-ble women residing in areas covered by anorganised screening programme; actual cov-erage refers to women who were sent an invi-tation to screening during 2004, based ondata provided to the GISMa survey.

Based on theoretical coverage, about 72%of Italian women in the 50-69 age range wascovered by a screening programme in 2004,showing a good increase in comparison withthe previous year (56.2% in 2003). A discrep-ancy between theoretical and actual coverage isstill present: the fraction of women actually re-ceiving an invitation to undergo screening in2004 (and who provided data to the survey) ac-count for 52% of the target population in the50-69 age range. This difference depends insome instances on organisation/managementdifficulties, due to which programmes havetrouble regularly inviting every year the entiretarget population; in other instances, the differ-ence is due to the failure on the part of theprogrammes to provide properly updated data.

Overall, around 1,950,000 women were in-vited to take a mammography in 2004; asmall part of them (about 82,000 women,i.e., 4.2% of the total), was invited by theprogrammes despite their being in the45-49 age range or over 69 years old.

The proportion of Italian women undergo-ing mammography in the absence of symp-toms is most likely higher than the reporteddata, as in several programmes spontaneousreferrals are accepted, which are not recordedin the database provided for the survey.

In spite of the large increase in the number ofprogrammes recorded even in the South ofItaly (table 2), where coverage increasedfrom 11% in 2003 to 38% in 2004, a largeimbalance in coverage is still present when



Region Testsperformed*

Levels ofcompleteness

Valle d’Aosta 5,044 1

Piemonte 116,443 1

Liguria 26,829 4

Lombardia 245,999 3

Trentino 18,964 1

Alto Adige 15,841 1

Veneto 119,672 1

Emilia-Romagna 168,429 1

NORTH 717,221 2

Toscana 128,489 1

Umbria 35,756 2

Marche 20,540 1

Lazio 66,792 4

CENTRE 251,577 2

Abruzzo 14.227 1

Molise 13,270 1

Campania 26.384 1

Basilicata 21,331 1

Calabria 189 1

Sicilia 7,652 4

SOUTH 83,053 1

ITALY 1,051,851 2

* first and repeat rounds

Table 3: Number of tests performed (first and repeat rounds)and levels of completeness of information about diagnostic process.Programmes active in 2004.

comparing Northern and Central Italy withthe South: here the percentage of targetpopulation not yet included in organisedscreening programmes is 60%, while inCentral Italy the target population is almosttotally covered, and in the North more than80% of the target population is included ina mammography screening programme.Considering the quality control of the pro-grammes, an important feature is suggestedby data shown in table 3, where the totalnumber of tests performed in 2004 and thedegree of completeness of the informationabout the diagnostic process is reported at aregional level. The degree of completeness isgiven based on a 4-level classification:� level 1: information available for >90%� level 2: information available for 60-90%� level 3: information available for 30-59%� level 4: information available for <30%.The overall difference between the twonumbers is about 238,000 tests, whichrepresents 22.5% of the total number ofwomen examined. Conditions are verydifferent at a regional level, and in severalareas differences are rather substantial (seeLiguria, Lombardia, Lazio, and so on); manyprogrammes were thus unable to providefurther information besides attendance, andthis is an important deficiency from thepoint of view of programme quality control.

AttendanceScreening programme attendance (compli-ance) is one of the main indicators for the eval-uation of the impact and efficiency of mam-mography screening.The crude attendance rate indicates the ratioof examined to invited women. Adjusted at-tendance (examined women/invited womenexcept women reporting a recent mammog-raphy), for those programmes that are able toprovide the information, is more representa-tive of real response to invitation, as it ex-cludes from the denominator those women re-porting a recent mammography outside the

programme. In these cases, mammography isnot repeated, and women are likely to complywith further invitation. Currently recom-mended standards are: �50% (acceptable)and �70% (desirable) for crude attendance;�60% and �75% for adjusted attendance,respectively.

Table 4 shows the results observed for crudeand adjusted attendance for single regions. Itis worth noting that not all programmescould provide data to calculate adjustedcompliance, thus the overall attendance rateis underestimated. With regards to crude at-tendance, only one region (Umbria) shows acompliance rate above the desirable standardof 70%, while 61% of the regions (11 out of18) are above the minimal standard of 50%;one third of the regions (6 out of 18) showsvalues below the minimal standards, andsometimes the difference is substantial.



Region Crudeattedance (%)

Adjustedattendance (%)

Valle d’Aosta 64.9 67.6

Piemonte 60.4 62.4

Liguria* 41.7 46.9

Lombardia 56.4 61.3

Trentino 64.6 71.7

Alto Adige 67.5 67.5

Veneto 63.4 72.4

E-Romagna 67.3 70.4

NORTH 60.2 64.8

Toscana 62.2 64.3

Umbria 71.4 72.3

Marche 31.2 35.6

Lazio 40.0 41.7

CENTRE 51.8 54.2

Abruzzo 56.0 58.1

Molise 52.6 52.6

Campania 29.4 29.7

Basilicata 52.0 52.0

Calabria 32.6 32.6

Sicilia 27.6 27.9

SOUTH 39.1 39.5

ITALY 55.6 59.0

*data refers to age group 50-64Values below minimal standard are shown in color.Values above optimal standard are shown in bold.

Table 4: Crude and adjusted attendance rate of invitedpopulation, age 50-69. Year 2004.

Contrasting with the increasing trend of thenational attendance rate recorded in the previ-ous years, a decrease, though not a substantialone, was observed in 2004 (55.6% versus58.3% in 2003). This decrease may be due tothe inclusion of programmes recently acti-vated in several regions; the initial years ofscreening programmes are often rather diffi-cult, although this is not always true, and itmay take some time before acceptable atten-dance rates are recorded. A decreasing trendis evident from Northern to Central to South-ern Italy, attendance rates being 60.2%,51.8%, and 39.1% respectively.Women in the 50-69 age range who werescreened in 2004 account for more than28% of the national target population.Considering adjusted attendance rate, no re-gion is above the optimal standard, whereas thenumber of regions below the minimal standardhas increased; it is important to consider that,as previously mentioned, this parameter is un-derestimated, since many programmes areunable to provide information about womenexcluded due to recent mammography.Attendance rates by 5-year age group areshown in table 5. Data confirms that youn-ger women have a higher compliance, a com-mon finding in other national programmes.�

The highest attendance is recorded amongwomen aged 55-64 years, relatively youngwomen who have been invited to take part inthe screening for several years; consequentlythey are more likely to participate, being alsoaware of the efficiency and the quality of thediagnostic procedures within an organised

screening programme. Caution should beused while interpreting national and regionalaggregated data, because they are the aver-ages of the results of individual programmes,which may vary substantially even within asingle region (figure 1). Crude attendancefor each programme ranges from 7.6% to88.2%, with 29% of the programmes belowthe minimal standard of 50%, and 21% ofthem above the optimal standard (70%).

Diagnostic indicatorsTables 6 to 8 show some of the main diag-nostic indicators, which are representative



Age Crudeattendance (%)


50-54 54.1 58.6

55-59 58.7 62.6

60-64 58.6 61.8

65-69 54.7 57.5

TOTAL 50-69 56.6 60.2

Results refer to a subset of programmes providing agestratified data.

Table 5: Crude and adjusted attendance by 5-year age groups.Year 2004.

Figure 1: Crude attendance rate in single programmes. Year 2004.

0 20 40%

60 80 100

AbruzzoAlto AdigeBasilicata

Campania

Emilia-Romagna

Lazio

Liguria

Lombardia

Marche

Molise

Piemonte

Sicilia

Toscana

Trentino

Umbria

Valle d’Aosta

Veneto

of the quality of screening performance.Indicators are reported separately for “firsttest” (women undergoing screening for thefirst time), and “repeat test” (women whopreviously underwent screening tests).

Referral rateIt indicates the proportion of screenedwomen referred for diagnostic assessments.This value needs to be reasonably low, in or-der to limit the negative psychological im-pact (anxiety), the invasive procedures (cy-tology, core, or surgical biopsies) which maybe required, as well as costs. Referral rate isthe main indicator of first-level screeningspecificity. Recommended standards are:<7% (acceptable) and <5% (desirable) atfirst screening; <5% (acceptable) and <3%(desirable) at repeat screening.

Detection rateIt indicates the proportion of detected can-cers every 1,000 screened women. It is themain indicator of the diagnostic sensitivityof the programme. Detection rate should beevaluated compared to expected incidencerate in the screened population, as breastcancer incidence shows both geographicaland age-related variations.

Benign/Malignant surgical biopsy ratioThis indicator is determined on women under-going recommended surgery. It indicates theratio of benign to malignant pathologicaloutcomes. It should be as low as possible, be-ing an optimal indicator of the diagnostic spec-ificity of the programme assessment phase.Recommended standards in Italy are: �1:1(acceptable) and �0.5:1 (desirable) at firstscreening, and �0.5:1 (acceptable) and �0.25:1(desirable) at repeat screening. The increasinguse of percutaneous core biopsy (e.g. vacuum-assisted biopsy) suggests caution in interpret-ing this indicator. The recent revision of the Eu-ropean guidelines seems to go in this direction;new suggested standards for this parameter arethe following: 0.5:1 (acceptable) and �0.25:1(desirable) both for first and repeat screening.

Detection rate of cancers �10 mm in size

It indicates the proportion of invasive cancers�10 mm detected every 1,000 screenedwomen. It summarises the capacity of theprogramme to detect “small” cancers, mostlikely “early” and associated with betterprognosis. It is an important indicator of thediagnostic sensitivity of the programme. Ex-cessively low values, especially lower than1‰ (where a low expected incidence cannotexplain them), could suggest the need to re-evaluate the quality of diagnostic proceduresused in the programme; on the other hand,values above 1.75-2‰ can be considered apositive indication of the diagnostic sensitiv-ity of the programme.

Proportion of in situ carcinomas

It indicates the ratio of carcinomas in situ every100 total detected cancers. Recommendedstandards are: 10% (acceptable) and 10-20%(desirable) at any screening round. The Italianstandards also give a maximum value, since toohigh a proportion of in situ carcinomas mightsuggest overdiagnosis or inadequate use of re-porting categories (overreporting) by patholo-gists. The new European standards do not in-clude a maximum value and suggested stan-dards are: >10% (acceptable) and >15% (de-sirable) at first and repeat screenings. In all ta-bles, values below minimal standard are in col-our, values above optimal standard are in bold.

Data refers to approximately 814,000 tests,accounting for more than 4,200 carcinomasdetected in 2004 at first screening (about1,900) or repeat screening (almost 2,800),and for 1,048 benign lesions. Screening centredatabases with a too limited number of testshave been excluded from the present analysis.

A critical aspect to take into considerationwhile analysing characteristics of detected car-cinomas is that information is not available foroverall identified cancers, and so in severalcases indicators are calculated in a subgroup ofcases; therefore data referring to detection rateof invasive cancers �10 mm and proportion



of in situ carcinomas should be interpreted withadditional caution. Though keeping theselimitations in mind, the indicators recordedfrom Italian programmes in 2004 appear over-

all rather good and comply at a satisfactorylevel with recommended national standards.

As already observed in previous surveys, anindicator that exceeds maximum standards



Region Total referralrate (%)

Total cancerdetection rate

(�1,000 screened)

B/M ratio Cancer �10 mmdetection rate

(�1,000 screened)

Carcinomain situ

(% cancers)

Valle d’Aosta 6.6 8.7 0.11 3.88 33.3

Piemonte 6.6 8.6 0.20 2.07 12.2

Liguria* 13.4 9.1 0.47 2.14 35.3

Lombardia 9.1 7.7 0.49 1.90 11.3

Trentino 6.7 7.6 0.00 3.81 0.0

Veneto 7.8 7.5 0.30 1.44 14.4

E-Romagna 8.6 6.9 0.40 1.31 15.9

NORTH 8.0 7.8 0.32 1.75 13.6

Toscana 8.2 5.1 0.24 1.66 11.1

Umbria 2.7 4.1 0.13 1.02 14.3

Lazio 10.9 5.9 0.37 1.27 8.7

CENTRE 8.5 5.2 0.27 1.55 10.9

Abruzzo 10.8 3.4 0.17 1.42 0.0

Molise 5.5 2.8 0.89 0.57 0.0

Campania 14.8 6.0 0.46 2.16 1.7

Basilicata 13.5 3.6 0.64 1.03 0.0

Sicilia 6.3 5.0 0.07 1.40 0.0

SOUTH 11.0 4.3 0.53 1.30 0.7

ITALY 8.8 6.6 0.34 1.63 11.9

*data refers to age group 50-64Values below minimal standards are shown in colour; values above optimal standards are in bold.

Table 6: Diagnostic indicators. First screening test.

Region Total referralrate(%)


(�1,000 screened)


(�1,000 screened)

Carcinomain situ

(% cancers)

Valle d’Aosta 3.9 8.0 0.09 3.00 6.3

Piemonte 3.6 5.4 0.21 1.85 15.8

Liguria* 3.3 6.5 0.40 2.60 20.0

Lombardia 4.5 4.3 0.34 1.20 10.3

Trentino 4.9 5.9 0.00 1.45 9.6

Veneto 4.1 5.0 0.19 1.56 11.7

E-Romagna 4.0 5.7 0.18 2.09 15.7

NORTH 4.1 5.3 0.21 1.75 13.8

Toscana 4.9 4.8 0.15 1.73 12.9

Umbria 2.5 4.6 0.37 0.81 9.5

Lazio 4.2 2.6 0.15 0.68 3.7

CENTRE 4.4 4.6 0.19 1.49 12.1

Abruzzo 6.9 2.5 0.00 0.84 7.1

Basilicata 5.0 2.6 0.26 0.57 17.4

Sicilia 8.6 5.2 0.00 1.29 12.5

SOUTH 5.9 2.7 0.15 0.71 13.4

ITALY 4.2 5.0 0.20 1.63 13.4

* data refers to age group 50-64Values below minimal standards are shown in colour; values above optimal standards are in bold.

Table 7: Diagnostic indicators. Repeat screening test.



at a national and often regional level is thereferral rate to diagnostic assessment, espe-cially for first tests (tables 6 and 7). Con-sidering first screening, 9 out of 15 regionsshow a referral rate above the acceptablestandard of 7%, and only one region has areferral rate below the optimal standard.The value of maximum standard is exceededvery often even when individual pro-grammes are considered (figures 2 and 3).Two thirds of the individual programmes(66.2% of the total) exceed the acceptablestandard for first screening tests of 7%, and39.2% of them (29 programmes) show a re-ferral rate above the value of 10%.As for repeat screenings, 19 programmes outof 65 (29.2% of the total) exceed the maxi-

mum standard, while about one fourth(27.7%) of the programmes record a referralrate below the optimal standard (<3%).Some values largely exceeding acceptablestandards may be explained, at least partially,by the limited experience of operators ofnewly implemented programmes, and be as-cribed to a learning curve effect: several pro-grammes were newly implemented in 2004.In other instances, where standards are ex-ceeded in programmes that have alreadybeen running for several years, further inves-tigation would be needed to better evaluateresults and identify reasons for failure.On the other hand, precise recording of re-sults, as well as comparison with other na-tional and foreign experiences and stan-

programmes0 20 40 60 80

30

25

20

15

10

5

0

%

12

10

8

6

4

2

0

programmes

%

0 2010 4030 50 60 70

Figure 2. Recall rate (%)in single programmes.First screening.Year 2004.One programme hasa recall rate over 50%(not shown).

Figure 3. Recall rate (%)in single programmes.Repeat screening.Year 2004.

dards, are essential in order to continuouslyimprove the screening procedures, and toachieve standardisation and quality controlof the different phases of the programme.

Data about referral rates suggests potentialcritical aspects for specificity in many pro-grammes; in the near future, more opportu-nities for discussing observed difficulties andsystematic interventions for quality assur-ance of the diagnostic procedures are re-quired, especially in programmes where sen-sitivity indicators (such as total detection rateand detection rate of invasive cancers �10mm), suggest the need for intervention.

Surely the regular collection of interval can-cers may improve the evaluation of the qual-ity of the programmes; this might be avail-able in the near future, thanks to the morefrequent and systematic use of formal record-ing of clinical information such as hospitaldischarge schedules (SDO). Table 8 showscrude and direct “truncated” standardised to-tal detection rates (standard: European popu-lation �1,000; age range: 50-69 years); use

of standardised rates allows for a better evalu-ation of the detection rate, which may vary byregion not only because of the different under-lying incidence, detection efficacy, and data re-cording completeness, but also due to differentage distribution of the examined population.

Tables 9 and 10 show diagnostic results by5-year age groups, and refer to a subgroupof programmes which provided data by agegroups. The 2004 results confirm previoussurveys, showing greater diagnostic problemsat younger ages. All indicators show worsevalues among women aged 50-54 years,namely a higher mammography positivityrate (higher referral rates) and a higher fre-quency of surgical procedures with benignoutcome (B/M ratio), together with a sub-stantially lower detection rate as comparedto older age groups, although the latter find-ing is expected due to a different age-specificincidence of breast cancer. From this point ofview, current inclusion of younger women(45-49 years) by several programmes shouldbe watchfully evaluated.



Region First screening Repeat screening

crudedetection rate

standardised**detection rate

crudedetection rate

standardised**detection rate

Valle d’Aosta 8.2 5.8 7.5 6.9

Piemonte 8.5 8.7 5.4 5.1

Liguria* 9.1 6.8 6.5 5.3

Lombardia 7.5 6.3 4.4 4.1

Trentino 7.7 7.6 5.9 6.0

Veneto 7.5 7.8 5.0 4.8

E-Romagna 6.9 9.7 5.7 5.1

NORTH 7.7 8.1 5.3 4.9

Toscana 5.5 5.8 4.7 4.6

Umbria 4.1 4.0 4.6 3.5

Lazio 5.9 5.9 2.6 2.7

CENTRE 5.5 5.7 4.5 4.3

Abruzzo 2.9 6.4 2.3 1.8

Molise 2.7 2.6 – –

Campania 5.9 7.3 – –

Basilicata 3.9 6.2 2.7 2.7

Sicilia 3.4 3.1 5.1 5.4

ITALY 6.6 6.9 5.0 4.7

* age 50-64 only ** European population

Table 8: Crude and standardised detection rate (European population) by regions (�100,000), age 50-69. First and repeat screenings.

Some preliminary data on the 2005GISMa survey of activity: coverageand attendanceAs a consequence of the formal agreementbetween the Ministry of Health and theONS, since 2004 regions must provide theONS with data about screening activity of allprogrammes implemented in their area.

Data reported in the following section may beconsidered almost exhaustive and provide atrue and complete picture of the implementa-tion and progression of organised mammog-raphy screening programmes in Italy, althoughfurther efforts will be required in the futureto improve data transmission and make it assystematic and complete as possible.

Preliminary data referring to geographical ex-tension and attendance rate for the year 2005are reported on a global and regional level.

Geographical extension

Table 11 shows data on theoretical and ac-tual extension, as previously defined.

During 2005, a further increase of screeningactivity was recorded and all Northern andCentral Italian regions were covered by screen-

ing programmes; although not all regions arecompletely covered). Theoretical extensiontherefore rose from 71.9% in 2004 to 76.4%in 2005, i.e., three quarters of the Italianwomen in the 50-69 age range are enrolled inorganised screening programmes.A major imbalance in coverage still exists be-tween Northern-Central and Southern Italy:while almost absolute coverage is observed inthe North and Centre of Italy, the situation isstill difficult in the South, where only 40% ofthe target population is involved in organisedscreening programmes.The target population of participating pro-grammes is above 5,500,000 women.According to the data provided, women whoreceived an invitation to screening in 2005 ac-count for 50.3% of the target population in the50-69 age range, with a light decrease as com-pared to the value of the previous year (52%).In 2005, approximately 1,820,000 women inthe 50-69 age range (table 11) received an in-vitation to take a screening mammography,and over 1,035,000 responded to the invita-tion and were examined within an organisedscreening programme. Women who were



Age Totalrecall rate

(%)


(�1,000 screened)


(�1,000 screened)

Carcinomain situ

(% cancers)

50-54 9.0 5.4 0.43 1.12 16.5

55-59 8.1 5.6 0.35 1.45 7.0

60-64 7.5 8.2 0.22 1.63 10.1

65-69 8.1 10.0 0.25 1.95 6.5

ITALY 50-69 8.4 6.7 0.33 1.41 11.3

Values below minimal standard are in colour; values above optimal standard are in bold.Results refer to a subset of programmes providing age stratified data.

Table 9: Diagnostic indicators by age groups. First screening.

Age Totalrecall rate

(%)


(�1,000 screened)


(�1,000 screened)

Carcinomain situ

(% cancers)

50-54 5.3 3.7 0.32 0.99 18.3

55-59 4.3 4.3 0.23 1.25 13.0

60-64 3.8 5.2 0.14 1.72 12.5

65-69 3.8 6.3 0.16 2.08 13.8

ITALY 50-69 4.2 5.0 0.19 1.55 13.7

Values below minimal standard are shown in bold.Results refer to a subset o programmes providing age stratified data.

Table 10: Diagnostic indicators by age groups. Repeat screening.

screened during 2005 accounted for 28.6%of the national target population.

Attendance

Table 12 shows preliminary results of crudeand adjusted attendance at a national andregional level.

A new inversion in the attendance rate wasevident in 2005, after the decreasing trendrecorded in 2004: a light increasing trend inattendance was observed in 2005, with anational mean value of 57.8%, compared to55.6% in 2004. Although a decreasing trendin participation is still evident from Northernto Southern Italy, it is noteworthy that in2005 the Central regions showed the highestincrease in attendance rates (values rose from51.8% in 2004 to 58.3% in 2005).

Several regions still have crude attendancerates below the minimal standards, as shownby the data reported. This shows the effortstill required to improve compliance: 8 out of

18 regions (44% of the total) are still unableto reach the acceptable standard, while onlytwo exceed the desirable standard (75%) forthis parameter.

ConclusionsIn March 2005, the Ministry of Health madean agreement with the ONS regardingoncological screenings, in order to define sev-eral aspects: a guideline for regional projects,ongoing monitoring of the programmesregarding their progress, consultation forcomputer-assisted management, training, qual-ity development and a plan for communica-tion. The major set goals of this agreement arethe following:� to extend coverage of screening programmesall over the Country, overcoming imbalancespresent between different areas of Italy;� to increase awareness and attendance ofscreening programmes among the group ofwomen who can benefit from them;



Region Theoretical coverageage 50-69 (%)

Women invited in 2005age 50-69

Actual coverageage 50-69 (%)

Valle d’Aosta 100.0 6,272 81.6

Piemonte 100.0 199,631 68.4

Liguria 47.4 32,497 28.4

Lombardia 95.5 421,131 69.7

Trentino 100.0 16,506 56.0

Alto Adige 96.9 23,283 88.2

Veneto 86.3 187,150 64.5

Friuli Venezia-Giulia 100.0 1,865 2.3*

Emilia-Romagna 100.0 237,452 88.5

NORTH 92.4 1,125,787 65.6

Toscana 100.0 199,328 82.7

Umbria 100.0 52,521 94.5

Marche 95.0 56,839 60.0

Lazio 98.4 149,472 44.2

CENTRE 98.6 458,160 62.8

Abruzzo 44.0 31,516 40.5

Molise 100.0 22,479 118.5

Campania 85.6 112,445 36.3

Basilicata 100.0 26,483 78.6

Sicilia 38.2 45,440 15.8

SOUTH 39.3 238,363 20.3

ITALY 76.4 1,822,310 50.3

* programme activated at the end of 2005

Table 11: Theoretiocal and actual coverage of target population by organised screening programmes. Programmes active in 2005.

� to guarantee the quality of organised screen-ing programmes.

The data reported for the years 2004 and2005 shows that a strong imbalance in mam-mography screening offer still persists in It-aly between the North-Centre and the Southof the country: in 2005, three quarters of thenational target population were enrolled inorganised screening programmes, but, whilealmost all the North and the Centre werecovered, more than 60% of women were notincluded in mammography screening in theSouth. The proportion of women who wereactually screened is just over one fourth(28%) of the national target population.

Periodic monitoring of data from screeningprogrammes is surely one of the mostimportant ways to guarantee the quality oforganised screening programmes, this beingone of the major issues pointed out by theMinistry of Health.

Lack of information on detected cases bymany programmes shows how much stillneeds to be done to evaluate the quality ofthe service provided to the population andto promote quality of the programmes.Surely the lack of invested financial resourcesand increasingly limited funds are one of themost important parameters affecting the ca-pacity of the programmes to maintain a highlevel of performance, from the point of view ofboth quality and quantity.The imbalance between theoretical and actualcoverage, shown by almost all programmes,stressed the problem of maintaining a con-stant invitation flow over time; this difficultysurely has an influence on the ability of pro-grammes in maintaining the two-year intervalbetween two consecutive screening rounds, asrecommended by the screening protocol.Information provided by several programmesconcerning time interval between two subse-quent screening tests goes in this direction: al-though the data of time interval are approxi-mate and collection of information is not ex-haustive, 22 programmes (out of 43 providingdata), corresponding to 51% of the total, sug-gest that the mean times between two screen-ing rounds are above 24 months, with a me-dian value of 27 months.

The decrease in the attendance rate re-corded in 2004, following an increasingtrend in recent years, may be partly due tothe inclusion of new programmes. In someinstances it may be ascribed to heterogene-ity of the invited population during the twoyears of the screening round: some pro-grammes have a tendency to concentratethe invitation of previous non-attendersover a short period of time. However, the re-corded attendance rate in 2005 againshowed an increasing trend.

Constant monitoring of attendance is stillnecessary, in any case, as it is one of the mainindicators of screening impact and efficacy inreducing breast cancer mortality.



Region Crudeattendance (%)


Val d’Aosta 71.1 71.9

Piemonte 56.7 59.0

Liguria 49.8 52.6

Lombardia 60.0 65.2

Trentino 82.3 84.7

Alto Adige 45.5 45.5

Veneto 64.8 72.6

Friuli V-G 47.2 47.2

E-Romagna 68.6 72.5

NORTH 61.8 66.3

Toscana 68.2 70.8

Umbria 60.3 62.5

Marche 60.7 61.6

Lazio 42.8 43.3

CENTRE 58.3 59.8

Abruzzo 48.4 50.8

Molise 40.0 40.1

Campania 31.4 31.4

Basilicata 51.5 51.5

Sicilia 37.9 39.3

SOUTH 37.9 38.5

ITALY 57.8 60.9

Values below minimal standard are in colour; valuesabove optimal standard are in bold.

Table 12: Crude and adjusted attendance rate among invitedwomen. Year 2005.

We must also consider that the attendance rateis biased towards underestimation, as GISMasurveys do not include women undergoingspontaneous screening, which in some pro-grammes may account for a substantialfraction of the target population; adjusting forthis spontaneous screening would allow a morecomplete understanding of mammographyscreening coverage in Italy.Inter-regional variability of some indicatorsand the recorded high referral rate would re-quire a more in-depth analysis; similarly, theincreasing introduction of new diagnostic tech-niques (such as microinvasive percutaneousbiopsy) might require a modified interpreta-tion of some indicators, or the introduction ofnew, more suitable indicators.A more critical evaluation should be used inthe analysis of sensitivity indicators, partic-ularly the identification of small invasive le-sions and the introduction, as soon as possi-ble, of the routine collection of interval can-cers, in order to achieve a higher efficacy ofthe screening programmes.Besides the presentation of the data stratifiedby region, a more detailed comparison at alocal level would most likely favour improve-ments of performance and better standardi-sation of procedures. From this point of view,the inclusion of screening activity in a com-prehensive regional project and the availabil-ity of regional reference centres may play asubstantial role in improving the quality ofindividual programmes.

Data for the ONS/GISMa survey for the years2004 and 2005 was provided by:

VALLE D'AOSTA: G. Furfaro, P. VittoriPIEMONTE: N. Segnan, L. GiordanoLOMBARDIA: E. Anghinoni, M. Bersani, L. Bisanti,R. Burani, N. Castelli, L. Cecconami, L. Fantini,L. Filippini, M.L. Gambino, S. Gatta, M. Gramegna,E. Invernizzi, S. Lopiccoli, G. Mangenes, G. Marazza,R. Paginoni, L. Pasquale, B. Pesenti, M.E. PirolaALTO ADIGE: A. Fanolla

TRENTINO: S. Piffer, A. BettaVENETO: M. Vettorazzi, C. Fedato, M. ZorziFRIULI-VENEZIA GIULIA: L. Zanier, A. FranzoLIGURIA: L. Bonelli, O. Brunetti, S. Erba,L. GaribottoEMILIA-ROMAGNA: C. Naldoni, P. Sassolide’ BianchiTOSCANA: E. Paci, F. Cipriani, S. De Masi,D. Giorgi, P. Mantellini, D. Marovelli, F. Mirri,M. Perco, B. Pocobelli, M. Rapanà, R. Rosati,L. Sbuelz, A. Scarfantoni, R. Turillazzi, L. VenturaUMBRIA: M.P. Lamberini, M. Petrella,S. Prandini, P. Tozzi, G. VintiMARCHE: C. Mancini, G. FuligniLAZIO: D. BaiocchiABRUZZO: G. Pizzicannella, C. Delli PizziMOLISE: F. CarrozzaCAMPANIA: R. Pizzuti, A. ChiancaBASILICATA: V. Barile, R. MagliettaCALABRIA: A. LeottaSICILIA: S. Ciriminna, D. La Verde, R. Musumeci,M.R. Russo, R. Scillieri, V. Spada.

References

1 Giordano L et al. and GISMa. Indicatori e stan-dard per la valutazione di processo dei program-mi di screening del cancro della mammella.Epidemiol Prev 2006; 2 (Suppl 1): 1-48.

2 Perry N et al. European guidelines for qualityassurance in breast cancer screening anddiagnosis, Fourth edition. European commission,Luxembourg 2006.

3 Frigerio A et al. Diffusione dello screening mam-mografico in Italia. In: Rosselli Del Turco M, Zap-pa M. Osservatorio Nazionale per la Prevenzionedei Tumori Femminili. Primo Rapporto. Roma2002: 14-21.

4 Giorgi D et al. Lo screening mammografico inItalia: dati GISMa 2001. In: Rosselli Del Turco M,Zappa M. Osservatorio Nazionale per la Preven-zione dei Tumori Femminili. Secondo Rapporto.Roma 2003: 14-23.

5 Giorgi D et al. Lo screening mammografico inItalia: dati GISMa 2002-2003. In: Rosselli DelTurco M, Zappa M. Osservatorio Nazionale per laPrevenzione dei Tumori Femminili. Terzo Rappor-to. Roma 2004: 16-25.

6 Giorgi D et al. Lo screening mammografico inItalia: survey 2003-2004. In: Rosselli Del TurcoM, Zappa M. Osservatorio Nazionale Scree-ning. Quarto Rapporto. Roma 2005: 16-27.

7 NHS Breast Screening Programme. Annual Re-view 2006: http://www.cancerscreen ing.nhs.uk



Mammography screening in Italy: 2004 survey and 2005 preliminary data

Documents

Transcript of Mammography screening in Italy: 2004 survey and 2005 preliminary data