RICHARDSON ET AL.

17.

18.

19.

20.

21.

22.

23.

24.

Hsing AW, Guo W, Chen J, Li JY, et al. Correlates of liver cancer mortality in China. Inf J Epidmiol 1991; 20: 54-9. Trichopoulos D. The causes of primary heptocellular carci- noma in Greece. R o g Med Virol 1981; 27: 14-25. McMahon BJ, Lanier AP, Wainwright RB, et al. Hepatocellular carcinoma in Alaska Eskimos: epidemiology, clinical features, and early detection. In: Popper H, Schaffner F, editors. Progrem in l i w diseaw. Vol. IX. New York: Harcourt Brace Jovanovich, 1990: 643-55. Dunk AA, Spiliadis H, Sherlock, et al. Hepatocellular carci- noma and the hepatitis B virus: a study of British patients. QJ

Szmuness W. Hepatocellular carcinoma and the hepatitis B virus: evidence for a causal association. Prog Mcd Virol 1978;

Simmons GC, Yeong ML, Lee SP, et al. The association of hepatitis B viral infection and hepatocellular carcinoma in New Zealand. N Z Mcd J 1983; 9 6 669-7 1. MacNab GM, Urbabowicz JM, Geddes EW, et al. Hepatitis B surface antigen and antibody in Bantu patients with primary hepatocellular cancer. Br J Cancer 1976; 33: 544-8. Tong MJ, Sun SC, Schaeffer BT, et al. Hepatitis-associated antigen and hepatocellular carcinoma in Taiwan. Ann I n t m

Mcd 1987; 62: 109-16.

24: 40-69.

Mcd 1971; 75: 687-91.

25.

26.

27.

28.

29.

30.

31.

32.

Dusheiko GM. Hepatocellular carcinoma associated with chronic viral hepatitis: aetiology, diagnosis and treatment. Br

Zaman SN, Melia WM, Johnson RD, POrtmann BC, et al. Risk factors in development of hepatocellular carcinoma in cir- rhosis. Prospective study of 613 patients. Lancef 1985; 1:

Johnson PJ. Williams R. Cirrhosis and the aetiology of hepatocellular carcinoma. J Hgotol 1987; 4: 140-7. Maddrey WC. Chronic hepatitis. Dis Mon 1993; 39:

Shapiro CN. Epidemiology of hepatitis B. Pcdiafr Infccf Dis J 1993; 12: 433-7. Kew MC. Detection and treatment of small hepatocellular carcinomas. In: Hollinger FB, Lemon SM, Margolis HS, edi- tors. Viral hepafitis and liver disease. Rmxeding of the 1990 international symposium on viral hepafitis and liver disease: con- fempurary issucs andfufurc prmpecis. Baltimore: Williams and Wilkins, 1991: 535-40. Wright TL, Lau JYN. Clinical aspects of hepatitis B virus infection. Lantcl 1993; 342: 1340-4. Stevens CE, Toy PY, Tong ML, et al. Pennatal hepatitis B virus transmission in the United States: prevention by passive-active immunization. JAMA 1985; 253: 1740-5.

Mcd Bull 1990; 46: 492-51 1.

1357-60.

53-125.

Participation in breast cancer screening: randomised controlled trials of doctors' letters and of telephone reminders Ann Richardson, Sheila Williams, Mark Elwood and Margaret Bahr

Tim Medlicott

Department of Preventive and Social Medicine, University of Otago, Dunedin

Mornington Health Centre, Dunedin

Abstract: The study used a randomised controlled trial to find out whether supporting letters from general practitioners accompanying the invitations from a screening centre affected participation in a population-based breast cancer screening program for women aged 50 to 64. A further randomised controlled trial compared the effect o f postal reminders with telephone reminders for women who did not respond t o an initial invitation t o participate in the program. There were 482 women in the first trial and 641 in the second. Excluding women who were ineligible or could not be contacted, participation in screening was 71 per cent in the group which received letters from their general practitioners compared with 62 per cent in the group which did not receive letters (P = 0.059). In the group that received letters, 56 per cent were screened without a reminder compared with 43 per cent of the group that did not receive letters (P= 0.01). Fewer women who received letters from their general practitioners declined the invitation to be screened (P = 0.048). In the second trial, there was n o difference in participation between the group receiving telephone reminders and the group receiving postal reminders. As in breast cancer screening programs in other countries, general practitioner endorsement of invitations increased participation in breast cancer screening. Postal reminders were as effective as telephone reminders in encouraging women who did not respond to an initial invitation t o participate in screening. (AustJ Public Heulfh 1994; 18: 290-2)

omen living in Otago and Southland, New Zealand, and aged 50 to 64 are eligible to W participate in a pilot breast cancer

screening program which offers free two-yearly screening using two-view mammography. This pilot program will provide information about the accept- ability, effectiveness and efficiency of breast cancer screening in New Zealand.'.!

Correspondence to Professor J.M. Elwood, Director, Hugh Adam Cancer Epidemiology Unit, Otago Medical School, PO BOX 93, Dunedin, New Zealand. Fax 64 3 479 7298.

One of the most important determinants of the effectiveness and efficiency of a screening program is the participation of the target group. If participation is high, the benefits of screening are available t o a greater proportion of the target group, and the pro- gram will also have greater cost-effectiveness. Eli- gible women are identified using general practice age-sex registers, supplemented with information from the electoral roll. Registration on the electoral roll is compulsory in New Zealand, but the pro- portion of the population on general practitioner

290 AUSTRALIAN JOURNAL OF PUBLIC HEALTH 1994 v a . 18 NO. 3

BREAST CANCER SCREENING TRIALS

rolls is unknown. Each woman is sent an invitation with a suggested appointment time, an information leaflet, and a reply-paid card with which to accept or decline the invitation. The initial invitation is posted one month in advance of the proposed appointment.

Two randomised controlled trials were set up to find out, for New Zealand women: 1 . whether an invi- tation accompanied by an endorsement letter signed by the woman’s own general practitioner was any more effective than an unaccompanied invitation, and 2. whether a postal reminder or a telephone reminder was the more effective strategy for women who did not respond initially.

Methods Of women aged 50 to 64 whose names were on the age-sex register of a large urban health centre (prac- tice population 15 500), 482 were thought to be eli- gible for screening and were randomly allocated either to receive a letter from their general prac- titioners or not. The general practitioners at the health centre provided personally signed form letters on the practice letterhead. If there was no reply within two weeks of the invitation being sent, a postal reminder was sent from the screening centre (the reminders were not signed by the general prac- titioners). When all the appointment dates had passed, the two groups were compared to see whether they differed in the need for reminders and in the final participation rate. The study was designed to have 80 per cent power to detect an absolute dif- ference in participation of 15 per cent at 0.05 significance.

In a separate and independent study to investigate the effect of different reminders, the names of 641 women who had not replied within two weeks to an initial invitation were collected. Women with tele- phone numbers were randomly allocated to receive either a postal reminder or a telephone reminder. The telephone calls were made up to three times at different times of the day. The 146 women without telephone numbers formed a third group and were sent a postal reminder. This study was designed to have the same power as the first study.

Results The effect of letters from general practitioners Invitations were sent to 482 women, 248 with accom- panying letters from their general practitioners and 234 with letters from the program only. Randomisation produced similar groups with respect to age. Also, a similar percentage of each group was ineligible for screening (Table 1 ) .

A lower proportion of women who received letters from their general practitioners replied, declining the invitation to be screened (21 of 248,8.5 per cent, compared with 35 of 234, 15 per cent of those who had not received letters). This difference was statisti- cally significant ( P = 0.03, difference 6.5 per cent, 95 per cent confidence interval (CI) 0.8 to 12.2).

Excluding women who were ineligible or could not be contacted because they were no longer at that address left 203 women who had been sent general practitioner letters and 192 women who had not. Of those who were sent letters from general prac- titioners, 1 13 (56 per cent) participated as a result of

Table 1 : The effect of letters from geneml practitioners to women thought to be eligible for screening (N=482)

Letter sent No letter sent n = 248 n=234

Excluded Recent mammogram Mastectomy Out of the country No longer at oddress

Eligible

Not screened No reply Replied, declining invitation Did not attend appointment

Screened Without reminder After reminder Total screened

26 26 1 0 0 1

18 15

203 192

26 20 21 35 12 18

1 1 3 (56%) 82 (43%) 31 37

144 (71%) 119 162%)

the first invitation compared with 82 (43 per cent) of those who were not sent the letters (P = 0.0 1 , differ- ence 13 per cent, CI 3.2 to 22.7). Including those who required a further reminder, final participation among eligible women who had received letters from general practitioners was 144 of 203 (71 per cent) compared with 119 of 192 (62 per cent) among women who did not receive the letters (P = 0.06, dif- ference 9 per cent, CI -0.3 to 18.2).

Telephonv reminders compared with postal remin&rs In a separate study, of 641 women who had not replied to a postal invitation, 495 had identifiable telephone numbers and were randomly allocated to receive either a telephone reminder or a postal reminder.

Table 2 shows the results of telephone reminders compared with postal reminders. Of the 248 women randomised to receive telephone reminders, 1 18 (48 per cent) were screened. Thirteen women said that they would like later appointments but did not specify a time; none of these women has yet been screened and they have been included with the nonparticipants.

Of the 247 women who were sent postal reminders, 121 (49 per cent) were screened. There was no significant difference in participation between the study and control groups ( P = 0.8, dif- ference I .4 per cent, CI -10.2 to 7.4).

Of the 146 women who did not have telephone numbers and so could not be randomly allocated to either group, only 40 (27 per cent) were screened; 20 per cent of their invitations were returned because the address was incorrect.

Table 2: Cornprison of telephoned reminders and postal reminden for 495 women who did not reply to on initial

invitation

Postal reminder Telephone reminder n=247 n = 240

Not screened No reply 57 49 Declined 35 33

Did not attend appointment 17 33 Ineligible 17 15

Screened 121 (49%) 1 18 (48%)

AUSTRALIAN JOURNAL OF PUBLIC HEALTH 1994 v a . I8 NO. 3 29 1

RICHARDSON ET AL

Discussion In previous studies of participation in breast cancer screening, certain characteristics of eligible women (including younger age, higher socioeconomic group, and an awareness of prevention programs in general) have been associated with higher participation."-" Aspects of the screening service, such as the method of invitation, have also been shown to affect participation."'

In this study, as with programs in other countries, personal endorsement of invitations by general prac- titioners increased Only 43 per cent of the women in our study were screened as a result of an unaccompanied invitation without a reminder. This level of participation would be unsatisfactory for a population-based screening program, as the ben- efits of screening would be available to fewer than half of the eligible women, and the cost-effectiveness of the program would be lowered, as the major costs in screening are in salaries and equipment, while the cost of consumable items (for example, film and chemicals) is low. The cost per woman screened drops as the number of women participating increases.

The participation rate in this study is higher than that in a similar study carried out in Australia.'" The women in our study had consulted their general prac- titioners within the previous two years and their names were therefore on practice registers. Partici- pation may differ for women who do not regularly attend a general practitioner and are not on a general practice register.

Apart from the effect on participation rates there are other advantages to general practitioner involve- ment in breast cancer screening programs. Such advantages include eliminating inappropriate invi- tations (for instance, to women who are already being followed up after a diagnosis of breast cancer). Also, women can discuss the invitation and the screening program with their general practitioners, who will know when women in their practices are being invited and can provide support for women who require further assessment and investigation as part of the program. Although the system of invitations used in this study was time-consuming for the general practitioners involved, it contributed to updating the practice register and was seen by the general prac- titioners as a good exercise in quality assurance for the practice.

In the second randomised trial, a telephone reminder and a postal reminder were equally effec- tive for women who did not respond to an initial invi- tation. Neither method resulted in a large number of missed appointments. There were 13 women among those telephoned who said that they wished to be screened later but did not specify a time. None of these women have yet been screened. It is sometimes difficult to refuse a telephone reminder and it may be that these women did not wish to be screened. If this is so, telephone reminders may be undesirable as they

put undue pressure on some women who have already decided that they do not wish to participate in the screening program.

As the reminder methods are equally effective, whichever best suits the administrative arrangements of the screening program could be used. An advan- tage of postal reminders is that some were returned marked 'wrong address'. As there is then no possi- bility of the appointment being used, it can be allo- cated to another woman straight away. Telephone reminders may be more time-consuming than postal reminders (especially if several calls have to be made in order to contact some women) and they have been found to cost more than postal reminders.I' But it is easier to organise an appointment at relatively short notice by telephone. This provides the program with greater flexibility in scheduling appointments (for instance gaps in the screening schedule due to can- cellations can be filled at short notice).

Acknowledgments The authors wish to acknowledge the help of the staff of the Otago-Southland Breast Cancer Screening Program (Prof T.C.A. Doyle, Clinical Director, Mrs E. Bang, Coordinator, Dr B. Berkeley, Dr S. Chartres, Dr J. Nicol, Mr B. Phipps, Mrs B. Morgan, Mrs M. McPhee) and staff at the Mornington Health Centre. This study was carried out as part of the evaluation of the pilot program funded by a grant from the Cancer Society of New Zealand and the Health Research Council of New Zealand.

References 1.

2.

3.

4.

5.

6.

7.

8.

Skegg D, Paul C. Benson-Cooper D. Chetwynd J. et al. Mammographic screening for breast cancer: prospects for New Zealand. N Z Med J 1988; 101: 531-3. Elwood JM, Doyle TCA, Richardson AK. New Zealand's first population based breast cancer screening program. N Z MedJ

Fink R, Shapiro S. Lewison J. The reluctant participant in a breast cancer screening program. Public Health Rep 1968; 83:

Vernon SW. Laville EX, Jackson GL. Participation in breast screening programs: a review. Soc Sci Med 1990; 30: 1107-18. Kruse J. Phillips DM. Factors influencing women's decision to undergo mammography. Obsfet +wcoI 1987; 70: 744-8. Williams EMI. Vessey MP. Compliance with breast cancer screening achieved by the Aylesbury Vale mobile service (1984-1988). J Public Health Med 1990; 12: 51-5. Hobbs P. The behavioural aspects o f breast and cervical screening. Radiogruphy 1986; 52: 287-90. Invip. L, Turnbull D, McMurchie M. A randomised controlled

1991; 104: 258-60.

479-89.

., trial o f general practitioner-written invitations to encourage attendance at screening mammography. Community Health Stud 1990; 14: 357-64. Williams EMI, Vessey MP. Randomised trial o f two strategies offering women mobile screening for breast cancer. BMJ

Dorsch MM, Cheok F, Ingham HM. The effectiveness of invi- tations from general practitioners in recruiting women to mammographic screening. Med J Aurt 199 1 ; 155: 623-5. Hurley SF. Jolley DJ, Livingston PM, Reading D, et al. Effec- tiveness, costs, and cost-effectiveness o f recruitment strategies for a mammographic screening program to detect breast cancer. J Natl Cancer Inst 1992; 84: 855-63.

9.

1989; 299: 158-9. 0.

1.

292 AUSTRALIAN JOURNAL OF PUBLIC HEALTH 1994 VOL. 18 NO. 3