• Open Access

Prostate cancer screening for men aged 75 to 84 years in New South Wales

Authors


Correspondence to: Professor Dianne L. O'Connell, Cancer Council New South Wales, Cancer Research Division, PO Box 572, Kings Cross, New South Wales 1340; e-mail: dianneo@nswcc.org.au.

In 2012, an international randomised controlled trial demonstrated that 11 years after screening with the prostate-specific antigen (PSA) test for prostate cancer there was a population-wide mortality benefit from having had the test, but that this benefit may be outweighed by the harms associated with testing and treatment.1 In Australia, the Royal Australian College of General Practitioners does not recommend screening asymptomatic men older than 75 years, due to the long natural history of untreated localised prostate cancer, the impact of competing causes of death, the increase in false-positive results with age and adverse effects of radical treatment.2 However – despite these recommendations – in a recent Australian study, 47% of men aged 75–84 years reported having had a PSA test in the previous two years.3 We used Australian Medicare claims data to determine the screening status of men aged 75–84 years with no prior diagnosis of prostate cancer and examined demographic, lifestyle and health-related factors associated with prostate cancer screening.

We used a subsample of data from a larger population-based cross-sectional study of a random sample of men aged 50–84 years resident in New South Wales (NSW) selected from the Australian Electoral Roll. Participants completed a computer-assisted telephone interview undertaken by trained interviewers between January and December 2007. The participation rate for the overall study was 55%. The Human Research Ethics Committee of Cancer Council NSW approved this study.

This analysis included men aged 75–84 years, with no prior diagnosis of prostate cancer, who provided consent to obtain their Medicare claims data. Data relating to any PSA tests participants had between 2005 and 2008 were provided by Medicare Australia, and for this analysis we used Medicare item 66655 to identify PSA tests to “screen” for prostate cancer, and items 66656 and 66659 to identify PSA tests for monitoring prostatic disease or follow-up of a previous high PSA test result. Multivariable logistic regression was used to identify factors associated with having a screening PSA test in the previous two years, using a significance level of 0.05. Very few men (n=15) in this sample had a family history of prostate cancer, so the role of family history could not be examined. Data were analysed using SAS software version 9.2 (SAS Institute, Cary, NC, USA).

Of 315 men who were aged 75–84 years at baseline, 212 (67%) consented to our accessing their Medicare claims data. We obtained Medicare records for 207 men (98%). The only statistically significant difference between men who consented and those who did not was that those who had a university-level education were more likely to consent (81% versus 65%). Of the 207 men with Medicare records, 65% had a record of a PSA test in the two years prior to their interview, 35% for screening and 30% for monitoring. In the year prior to interview, 47% of men had a PSA test, 20% for screening and 27% for monitoring. Of the men who had a monitoring PSA test, 54% had another PSA test within 12 months of their interview. When comparing men who were screened to men who were not tested, we found that men who consumed 1–14 alcoholic drinks per week were more likely to have a PSA screening test than non-drinkers, and men whose highest education level was Certificate/Diploma/Trade/Apprenticeship were less likely to have a PSA screening test than men who had a university degree (Table 1). Having urinary symptoms and living in less remote areas were marginally non-significantly associated with PSA screening (Table 1). The median age of men who had a screening test or had not been tested was the same (79 years).

Table 1. Demographic, lifestyle and health-related factors associated with having a PSA ‘screening’ test in the previous two years for men aged 75 to 84 years in NSW.
 Men in sampleHad a screening PSA testaOdds ratiob95% confidence intervalb p-valueb
 N%n%   
  1. Notes:

  2. a. Weighted, totals may not match due to rounding

  3. b. For a “screening” PSA test compared to no PSA test, excluding men with a “monitoring” PSA test, adjusted for all variables shown in the table

  4. c. Age treated as continuous variable in logistic regression, odds ratio is for a one-year increase

  5. d. Other co-morbidities were emphysema, bowel cancer, heart conditions, stroke or lung cancer

  6. (ref): Reference category

Age c
 75–79
 80–84
 
127
80
 
62
38
 
44
28
 
35
35
0.950.82–1.090.44
Marital status
 Married/Living as married
 Other

157
50

76
24

53
19

34
39

1.00
0.71

(ref)
0.31–1.65
0.43
Education
 Higher School Certificate or less
 Certificate/Diploma/Trade/Apprenticeship
 University

75
103
29

36
50
14

29
32
11

39
31
37

0.59
0.27
1.00

0.15–2.27
0.07–0.98
(ref)
0.05
Country of birth
 Australia
 Other

164
43

79
21

54
18

33
41

1.00
1.16

(ref)
0.48–2.81
0.75
Accessibility/Remoteness of area
 Major cities
 Inner regional
 Outer regional/Remote

127
51
28

62
25
14

45
19
7

36
38
26

1.00
1.12
0.31

(ref)
0.46–2.76
0.10–0.91
0.07
Smoking
 Never smoker
 Ever smoker

81
126

39
61

24
48

30
38

1.00
1.28

(ref)
0.57–2.86
0.55
Alcohol consumption
 Non-drinker
 1–14 drinks per week
 15+ drinks per week

78
89
40

38
43
19

25
34
12

33
39
31

1.00
2.81
0.75

(ref)
1.14–6.96
0.27–2.11
0.03
Frequency seeing doctor
 At least every 2 years
 Less regularly

168
39

81
19

56
15

34
40

1.00
1.37

(ref)
0.56–3.38
0.49
Self-reported health
 Excellent/Very good/Good
 Fair/Poor

168
39

81
19

53
19

31
50

1.00
1.61

(ref)
0.62–4.18
0.33
Urinary symptoms
 No
 Yes

148
59

71
29

50
22

34
37

1.00
2.09

(ref)
0.97–4.50
0.06
Other co-morbidities d
 None
 At least one

93
114

45
55

29
43

31
38

1.00
0.96

(ref)
0.45–2.05
0.92

This analysis has some potential limitations. First, we used claims for one specific Medicare item to classify the test as a screening test. It is possible that an unknown number of tests may be misclassified as being for screening when they were performed for monitoring or follow-up, or vice versa. Further, Medicare Australia has a procedure called “coning” which may have affected the accuracy of the claims data.4 Coning occurs when more than three pathology items are requested and only the three most costly items are reimbursed and recorded by Medicare Australia. It has been estimated that 16–27% of screening PSA tests may be coned, in which case the true number of screening PSA tests could be higher.5

Despite guideline recommendations not to screen for prostate cancer in this age group, one-third of men aged 75–84 years were still undergoing “screening” PSA tests. We found associations between education and alcohol consumption and PSA screening, suggesting that there is no obvious driving factor in the practice of PSA screening for men aged over 75 years. Excessive PSA testing contributes to financial costs for the patient and government, in addition to the anxiety or other health issues that may accompany a high PSA test result and any subsequent procedures. Future studies should include general practitioners to examine the reasons for prostate cancer screening in older men.

Acknowledgements

We thank Dr David Smith, Dr Melina Gattellari, Dr Suzanne Chambers, Dr Carole Pinnock, Mr Terry Slevin and Professor Jeanette Ward for their substantial contributions to this study. We also acknowledge the contributions of the interviewers at the Hunter Valley Research Foundation who conducted the interviews, the Australian Electoral Commission who provided a random sample of men in New South Wales, and Medicare Australia and the Department of Veterans’ Affairs for making the Medicare claims data available for analysis.

This study was funded by National Health and Medical Research Council Project Grant #337601 and Cancer Council NSW.

An abstract was presented at the Australian Prostate Cancer Conference in August 2011, and the Clinical Oncological Society of Australia's Annual Scientific Meeting in November 2011.