14 J Med Screen 2000;7:14 18 Implementation of service screening with mammography in Sweden: from pilot study to nationwide programme Silas Olsson, Ingvar Andersson, Ingvar Karlberg, Nils Bjurstam, Ewa Frodis, Stefan Håkansson Swedish Institute for Health Services Development, Stockholm, Sweden S Olsson S Håkansson Malmö University Hospital, MAS, Malmö, Sweden I Andersson National Board of Health and Welfare, Stockholm, Sweden I Karlberg Institute of Selected Clinical Sciences, Department of Radiology, Gothenburg University, Sweden N Bjurstam Västerås Central Hospital, Västerås, Sweden E Frodis Correspondence to: Mr Silas Olsson, Telia Research, SE-123 86 Farsta, Sweden Email: silas.a.olsson@telia.se Accepted for publication 23 February 2000 Abstract Establishment of mammography screening in Sweden has progressed logically from pilot study through clinical trials to service screening. Screening with mammography for early detection of breast cancer has been provided by all Sweden s 26 county councils since 1997. It took 23 years from the initial pilot study through clinical trials to the establishment of mammography service screening throughout Sweden. In the screening rounds completed by 1995 96, and provided by all but one county council, 1 040 000 women participated, corresponding to 81% of those invited. The national average recall rate was 2.2%, and consequently 23 000 women were recalled for additional investigations. Eleven county councils invited women aged 40 74, six invited women aged 50 69, the remaining eight invited women between both these age intervals. Mammography outside screening programmes clinical mammography is available throughout Sweden. About 100 000 women a year were referred for clinical mammography and about 50% of these were either younger or older than those invited for screening. A negative relation between the use of clinical mammography and participation in the screening programmes was noticed. (J Med Screen 2000;7:14 18) Keywords: mammography; breast cancer; cost analysis Screening with mammography is one of the largest public health evorts to promote women s health in Sweden. The national service screening programme is an example of evidence based medicine, starting with a pilot study in Gävleborg county council in 1974 1 and continuing with several randomised clinical trials later in the 1970s. 2 4 Based on the results of the two counties (WE) study the National Board of Health and Welfare in Sweden published General guidelines on mammographic screening in 1986. 2 5 These basic recommendations for mammography screening included women aged between 40 and 74. As a result of the amendments of the national recommendations in 1987 and 1988, screening of women only from 50 to 69 years of age was allowed if county councils were short of trained stav. 6 7 The recommendations resulted in a rapid divusion of service screening among the county councils. In the following years, data from the WE study and from other clinical trials accumulated. 3 4 An overview (pooled data from five individual studies from Sweden) in 1993 showed a 24% reduction in breast cancer mortality after 12 16 years among those invited for screening compared with those not invited. 8 Additional data on women below age 50 at entry were published in 1997 and 1998 showing a statistically significant reduction in breast cancer mortality in this age group also. 9 11 On the basis of the new evidence, the National Board of Health and Welfare issued a revised version of the national recommendations towards the end of 1998. 12 The main difference from the previous version was a clearer recommendation to include women aged 40 49. The public part of Swedish health care was, at the time of this study, provided by 23 county councils (geographically defined areas governed by regional governments regarding for example, health care) and three municipalities, Gothenburg, Malmö, and Gotland that is, 26 healthcare areas. In 1999 some county councils merged to form larger regions. In this report, for practical purposes, we describe the healthcare system as having 26 county councils. The county councils operate independently within the legal frame of legislation set by the national government and are financed mainly by regional taxation, patient fees, and national governmental support. The decision to implement screening programmes or not is the responsibility of the county councils. The success of the screening programme depends on many factors, including participation rate and the total exposure to mammography that is, mammography screening and clinical mammography. Methods of survey Radiologists responsible for mammography screening, departmental heads and medical directors in radiology, and private providers of mammography in all county councils in Sweden were surveyed by postal questionnaires in 1989 and 1993. After reminder letters, the response rate from the two surveys was 100%. 13 15 During 1996 98 specific data of the surveys were updated by questionnaires, letters, and phone calls. 12
Mammography screening in Sweden 15 30 25 WE study published National recommendations County councils 20 15 10 5 0 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 Year Figure 1 Establishment of mammography screening by all the 26 county councils in Sweden. Pilot and trials during 1974 to 1986. The diagram indicates the publication dates of the WE report 2 and the General Guidelines on Mammographic Screening by the National Board of Health and Welfare in Sweden. 5 Results DIFFUSION OF MAMMOGRAPHY SCREENING PROGRAMMES In 1988, 15 county councils provided service screening programmes. By 1992, 22 of Sweden s 26 county councils provided screening programmes. The introduction was delayed in the remaining four county councils owing to lack of stav and facilities under construction. Three of these started service screening programmes during 1993 96. The last, Gotland, started screening in 1997. Consequently, from the pilot study in 1974, it took 23 years to establish mammography screening among all the county councils in Sweden. Figure 1 illustrates the establishment of mammography screening in Sweden. Most service screening programmes were introduced after publication of the results from the WE study 2 and the subsequent national recommendations. 5 By 1997 all county councils in Sweden provided mammography screening based on active invitation. The problem of stayng the mammography screening programmes can be illustrated by the major shortage in personnel reported by all but two of the county councils in 1989. In 1993 the situation had improved substantially, but still seven county councils continued to be short of qualified stav. Between 1996 and 1999, a lack of radiologists again became apparent, at least in part due to the retirement of the first generation of mammography screening radiologists. AGE GROUPS, PARTICIPATION, RECALL RATE, AND PATIENT FEES In Sweden there are 1.8 million women aged between 40 and 74 and 0.96 million women aged between 50 and 69. In the screening round, performed mainly between 1989 and 1992 in 22 county councils, 803 000 women participated, corresponding to a participation rate of 81% (63 89%). This can be compared with the screening round completed in 1995 96 in 25 county councils where 1 040 000 women participated, corresponding to 81% (66 91%) of those invited. In this round of screening, 11 county councils invited women aged 40 74, six invited women aged 50 69, and the remaining eight invited women between these two age intervals. Screening programmes which include women below the age of 50 had a higher participation rate. The national average recall rate for further investigations was 3.5% in 1988, 2.9% in 1992, and 2.2% in 1995 (0.8 5.1%). Given an average recall rate of 2.2%, this means that about 23 000 women were recalled for additional investigations, which in most cases implied a complete mammography examination only. Table 1 shows the basic data on mammography screening programmes in Sweden in 1995 96. Based on data from 23 county councils, nearly all screening programmes in 1993 used two x ray images for each breast in the first round of screening, and thereafter one or two images, depending mainly on the radiological density of the breast and the woman s age. Occasionally, programmes used the one or two image method throughout. Double reading of mammograms (two radiologists review each case), as recommended by the national guidelines, was used in 15 screening programmes, partially in five programmes, and in three programmes single reading was used. The fee paid by the woman (not reimbursed) for the screening in 1997 was about $13 (104 SEK, ranging between 0 and 170 SEK among the county councils). The remaining cost for the screening is paid by the county councils. There was no apparent relation between the fee and the rate of participation (R 2 (adj) = 0.06). WAITING TIMES One important factor determining the degree of anxiety produced by the screening programme is the waiting time for the results of the examination and recall for further examinations. 16 18 In 1993 13% of Sweden s screening centres normally sent a letter to women within three days describing the results of the examination, 35% sent the results within four to 10 days, and 17% sent the results within 11 to 14 days. Waiting times at the remaining screening centres (one third) were
16 Olsson, Andersson, Karlberg, et al Table 1 Basic data on mammography screening programmes in Sweden in 1995 96 The Swedish county councils and municipalities with authority for health care Start of mammography screening Age groups invited Screening interval (months) Number of women invited in the screening round 1995 96 Participation in screening round 1995 96 (%) Screening fee that womenhavetopay, 1995 96 (1997) (SEK) Stockholm (5 sites) 1989 90* 50 69 24 170 000 72 50 Uppsala 1988 40 74 18 24 53 400 81 120 Sörmland 1989 40 74 18 24 53 000 86 110 (120) Östergötland 1978 40 74 18 24 85 600 80 50 (100) Jönköping, East 1986 40 74 22 23 000 86 100 Jönköping, West 1987 40 74 24 39 000 86 100 Kronoberg 1990 50 69 18 18 500 75 170 Kalmar 1986 40 74 24 50 000 85 120 Blekinge 1988 45 69 18 24 21 000 87 0 Kristianstad, Krist. 1990 50 74 18 24 29 000 76 120 Kristianstad, Ängel. 1989 50 74 18 24 13 000 70 120 Malmöhus 1987 89 40 74 18 24 72 000 80 120 Halland 1989 50 74 21 31 000 84 60 Bohus 1986 50 74 24 37 000 90 100 Älvsborg, North 1993 40 74 24 21 000 83 100 Älvsborg, South 1988 40 74 24 55 000 84 100 Skaraborg 1989 50 69 24 31 500 87 125 Värmland 1993 50 69 18 24 33 000 83 100 Örebro 1987 40 74 18 24 47 000 83 50 Västmanland 1986 40 69 20 24 47 000 87 50 (100) Dalarna 1977 40 70 20 24 54 000 88 100 Gävleborg 1974 40 74 21 62 000 85 75 Västernorrland 1990 40 74 21 55 000 89 100 (150) Jämtland 1996 50 69 20 24 15 000** 91 150 Västerbotten 1995 50 69 24 27 000 89 100 Norrbotten 1989 40 74 18 24 55 000 82 150 Malmö, municipality 1976 46 69 18 24 18 000 66 120 Gothenburg, municipality 1983 50 74 18 60 000 70 125 Gotland, municipality Started 1997 (40 69) (18 20) (10 000) (89) (80) Total, year 1996 1 276 000 mean 81 mean 99 (104) (Total, year 1993) (996 168) (mean 81) (mean 91) * South hospital started 1981. Changed to 50 69 in 1997, to 45 69 in 1996. Depending on age or tissue density. **Screening programme started during 1996. insuyciently reported. If further investigation was required, women were notified by letter or telephone by 10% of the screening centres within three days, by one third of the centres within four to 10 days, and by 23% within 11 to 14 days. The remaining one third of providers either did not answer this question or provided insuycient information. CLINICAL MAMMOGRAPHY Clinical mammography denotes a mammography examination following referral, usually owing to symptoms or anxiety. Clinical mammography was introduced in Sweden during the 1960s, in Gothenburg, Stockholm, and Gävleborg county councils, and became established in the remaining county councils during the 1970s. Modern mammography using molybdenum technology was introduced in Sweden in 1968 in Gothenburg. In 1992, 61 hospitals in Sweden (out of about 90 in total) provided clinical mammography. In 1993 privately operated clinical mammography services were available at 12 locations in the metropolitan areas of Gothenburg, Malmö, Stockholm, and Uppsala. Between 1988 and 1992 the total volume of clinical mammography provided by the county councils and the private facilities declined from 133 000 to 122 000 women examined a year, corresponding to a reduction of 8%. During this time the volume of private clinical mammography declined from 44 656 to 38 066, a reduction of 15%, and public clinical Attendance rate in screening programmes (%) 95 90 85 Sweden (mean) 80 75 Gothenburg 70 Malmö Stockholm 65 0 10 20 30 40 50 60 Clinical mammography per 1000 women (all ages) Figure 2 Possible relation between participation in screening programmes (the screening rounds until 1992) and the scope of clinical mammography (1992). Each symbol in the diagram corresponds to one screening programme. 15
Mammography screening in Sweden 17 mammography declined from 88 253 to 83 699 women examined, a reduction of 5%. The tendency for women of all ages to seek clinical mammography has been studied. The results, based on a sample of approximately 40% of the clinical mammography volume in 1992, show that 30% of the women receiving clinical mammography from public health services are younger than the minimum screening age of women invited for screening in their respective county councils. The corresponding fraction in the private sector is 47%. In both the public and private sectors, 14% of women were older than the maximum age of women invited to screening programmes. This means that half of the women undergoing clinical mammography were either younger or older than the age group invited for screening in the county councils respectively. Participation in screening programmes completed by 1992 varied from 63% to 89%. This participation was compared with the distribution of clinical mammography to show how clinical mammography possibly influenced participation in the screening programmes. Figure 2 shows that participation in the screening programmes was lowest in the three metropolitan areas of Stockholm, Gothenburg, and Malmö, where clinical mammography was most frequent. The average use of clinical mammography in the 10 county councils that have provided screening programmes since 1988 or earlier (excluding Stockholm, Gothenburg, and Malmö) was 15 women per 1000 female inhabitants and per year (1992). The corresponding average for all county councils with screening programmes, excluding and including respectively the metropolitan areas, was 18 (17 and 20 for counties using age intervals of 40 74 years and 50 69/74 years respectively) and 28 women per 1000 female inhabitants and per year. The average clinical mammography volume in the three county councils with no screening programmes in 1994 was 39 women examined per 1000 female inhabitants and per year. Clinical mammography in Stockholm, Gothenburg, and Malmö averaged 50 women examined per 1000 female inhabitants and per year (1992). Discussion Service screening with mammography for early detection of breast cancer is an example of evidence based medicine. After a pilot study, clinical trials, and national recommendations it took only five years to implement service screening in most county councils (fig 1). The total time span from the pilot study to nationwide coverage of service screening was 23 years. However, the pros and cons of service screening with mammography have been debated by the professionals and politicians in Sweden, sometimes intensely. One such debate resulted in a complete stop of service screening in one county council for six months in 1996. Participation in Sweden s mammography service screening programmes was high, with an average of 81% and a range from 66% to 91% in the diverent programmes. The larger metropolitan areas were chiefly responsible for pulling the national average down. One of the reasons seems to be the relation found between the greater use of clinical mammography in the large metropolitan areas and lower participation in the screening programmes. Socioeconomic and other factors may also contribute to the lower participation in the metropolitan 19 20 areas. In the city of Malmö, which is the third largest city in Sweden, a correlation has been found between attendance and registered unemployment, income support, and foreign backgrounds (Matson S, et al, unpublished data). The county councils have adopted diverent age intervals when inviting women for screening owing to lack of stav or other resources and probably also because they valued the benefits of mammography screening diverently. The fraction of women recalled from the screening programmes for supplementary examinations decreased from 3.5% to 2.2% as national averages, from 1988 to 1995. The recall rate was higher in the first round of screening. This may be explained partly by the fact that more cases of cancer appeared in the first round than in subsequent screening rounds, and partly by the fact that the later rounds had the advantage of comparison films and also increasing experience among the radiologists. Thus the round of screening completed in 1995 96 resulted in recall of about 23 000 women. The anxiety experienced by some women in conjunction with additional investigations is receiving increasing attention. 16 18 Additional work up also represents a substantial cost. 12 21 Although this recall fraction is low by international levels, a reduction of the number of false positive cases is essential. Substantially improving the predictive value of initial screening will probably require new methods of investigation. Digital mammography and computer based image analysis systems as well as magnetic resonance tomography may contribute in this direction. Digital mammography together with telemammography, when appropriately developed, will open new possibilities for distant reading and review of the mammograms, quality assurance, and organisational development. Clinical mammography is available throughout Sweden. This activity declined somewhat when screening programmes were started. One reason why the decline has not been greater, despite the major expansion of mammography screening during the period, is probably that 50% of the women who demand and receive clinical mammograms are outside the ages invited for screening. This should be taken into consideration when discussing age groups to be invited to screening programmes. Another reason for the continued high use may be the easy access to clinical mammography in metropolitan areas. Conclusions The establishment of mammography screening in Sweden represents a logical series of events,
18 Olsson, Andersson, Karlberg, et al from pilot study through clinical trials to service screening. Although the entire process took 23 years, after the production of scientific evidence service screening was rapidly established in most of the county councils with a high degree of acceptance among invited women. We thank colleagues and others who responded to the survey and contributed information to this study, and also B Lundgren, Gävle, S Törnberg, Stockholm, and L Nyström, Umeå, for specific statistical data, and Erik Grönqvist, Stockholm, for statistical analysis. 1 Lundgren B, Jakobsson S. Single-view mammography. A simple and eycient approach to breast cancer screening. Cancer 1976;38:1124 9. 2 Tabár L, Gad A, Holmberg LH, et al. Reduction in mortality from breast cancer after mass screening with mammography. Lancet 1985;i:829 32. 3 Andersson I, Aspegren K, Janzon L, et al. Mammographic screening and mortality from breast cancer. The Malmö mammographic screening trial. BMJ 1988;297:943 8. 4 Frisell J, Eklund G, Hellström L, et al. Randomized study of mammography screening: preliminary report on mortality in the Stockholm trial. Breast Cancer Res Treat 1991;18:49 56. 5 Mammografiscreening: hälsokontroll för tidig upptäckt av bröstcancer. [General guidelines on mammographic screening: health check-up of early discovery of breast cancer.] Socialstyrelsen, Stockholm, Sweden: National Board of Health and Welfare, 1986. (Allmänna råd från Socialstyrelsen, Liber 1986:3.) 6 Om ändring i förslag till riktlinjer för mammografi-screening. [About changes in the suggested guide-lines for mammographic screening.] Socialstyrelsen, Stockholm, Sweden: National Board of Health and Welfare, 1987. (Meddelandeblad 15/87.) 7 Till samtliga landsting beträvande mammografiscreening.[to all county-councils concerning mammographic screening.] Socialstyrelsen, Stockholm, Sweden: National Board of Health and Welfare, 1988. (Letters to all county-councils.) 8 Nyström L, Rutqvist LE, Wall S, et al. Breast cancer screening with mammography: overview of Swedish randomised trials. Lancet 1993;341:973 8. 9 Andersson I, Jansson L. Reduced breast cancer mortality in women under age 50: updated results from the Malmö mammographic screening program. J Natl Cancer Inst 1997;22:63 7. 10 Bjurstam N, Björneld L, DuVy SW,et al. The Gothenburg Breast Screening Trial. First results on mortality, incidence, and mode of detection for women aged 39 49 years at randomization. Cancer 1997;80:2091 9. 11 Larsson LG, Andersson I, Bjurstam N, et al. Updated overview of the Swedish randomized trials on breast screening with mammography: age group 40 49 at randomization. J Natl Cancer Inst 1997;22:57 61. 12 Hälsoundersökning med mammografi. [Screening with mammography.] Socialstyrelsen, Stockholm, Sweden: National Board of Health and Welfare, 1998. 13 Olsson S, Lithander E. Mammografiverksamhetens utbredning och omfattning i Sverige: sammanställning av en enkätstudie. [The establishment of the mammographic work in Sweden: a specification of a study of inquiries.] Stockholm: Spri, 1989. 14 Lennholm B. Enkätstudie om mammografi i Sverige: regionalt ojämn utbyggnad - inga resurser för screening i åtta landsting. [A study of inquiries about mammography in Sweden: a regionally uneven extension - no resources for screening in eight county-councils.] Läkartidningen 1989; 86:3379 80, 84. 15 Olsson S, Andersson I, Bjurstam N, et al. 600 000 kvinnor per år undersöks med mammografi: var femte inbjuden avstår från screening. [600 000 women are each year examined using mammography: every fifth who is invited gives up screening.] Läkartidningen 1995;92:552 6. 16 Ellman R, Angeli N, Christians A, et al. Psychiatric morbidity associated with screening for breast cancer. Br J Cancer 1989;60:781 4. 17 Olsson P, Armelius K, Lenner P, et al. Mammografiscreening: en pilotstudie på de psykologiska evekterna. [Mammographic screening: a pilotstudy of the psychological evects.] Läkartidningen 1993;90:3213 14, 17. 18 Lidbrink E, Elfving J, Frisell J, et al. Neglected aspects of false positive findings of mammography in breast cancer screening: analysis from the Stockholm trial. BMJ 1996;312:273 6. 19 Hurley SF, Huggins RM, Jolley DJ, et al. Recruitment activities and sociodemographic factors that predict attendance at a mammographic screening program. Am J Public Health 1994;84:1655 8. 20 Kiefe CI, McKay S, Halevy A, et al. Is cost a barrier to screening mammography for low-income women receiving medicare benefits? Arch Intern Med 1994;154:1217 24. 21 Svensson H, Håkansson S. Hälsokontroll med mammografi: kostnader och nytta. [Health check-up with mammography: costs and benefits.] Stockholm: Spri, 1990. (Spri rapport 298.)