Over the past decade, four Australian medical schools (Flinders University, the University of Queensland, the University of Sydney, and the University of Melbourne) have introduced four-year medical courses accessible only to graduates. Introducing these graduate medical programs (GMPs) required review and reform of the curricula. Five- or six-year undergraduate courses continue in other Australian and New Zealand medical schools, although there have been varying degrees of curriculum reform. Shorter courses may have less range and depth of content. Any such trend would run counter to efforts to increase teaching on cancer, which is a major and increasing health problem in both countries.1 Past surveys of medical students' oncology knowledge and skills2,3 have not indicated a high standard of training in this area.
We surveyed interns in early 2001 with a questionnaire based on one for an earlier study.2 Our aims were:
To compare the cancer knowledge and skills of interns in 2001 who graduated from GMP courses with those from non-GMP courses; and
To compare the cancer knowledge and skills of interns in 2001 with those who completed the survey in 1990.
The study population was recent medical graduates from all universities in Australia and New Zealand undergoing orientation as interns in teaching hospitals during the first two weeks of employment in January 2001. These doctors had just qualified but were not yet involved in hospital practice.
Sample size calculations and selection of hospitals were not straightforward because of likely cluster effects (the magnitude of which is difficult to estimate from published articles) and the lack of a comprehensive list of the numbers of interns accepted by each hospital from each medical school.
The earlier study obtained 389 responses.2 Our study was designed to have an 80% chance of detecting and declaring statistically significant, using a two-sided α = 0.05, absolute changes of about 10% in the proportions of interns reporting no exposure to palliative care (50% in 1990), to radiation oncology (42%) or medical oncology (18%), or "poor" instruction in palliative care (29%). This required at least 400 responses in 2001. We considered that, if the response of 84% obtained in 1990 could be duplicated, we would have to distribute about 500 questionnaires. There are about 1200 interns each year in Australia and New Zealand, so the simplest design involved a one-in-two sample. We stratified hospitals by State or country, ranked them by numbers of interns, divided the list into consecutive pairs, and then selected at random one member of each pair of hospitals.
Our questionnaire covered areas of knowledge and attitudes identified in the Australian Cancer Society (ACS) Ideal Oncology Curriculum as essential for graduating medical students.4 There were 90 questions divided into 34 topics. These included knowledge of prevention, screening and early detection of cancer, as well as the care and outcome of patients with cancer. We also asked about the interns' perceptions of their abilities to perform important tasks in clinical practice, such as preparing a patient for a hazardous procedure and discussing death with a dying patient. Interns rated the quality of their teaching about cancer, and reported the time in their medical courses that they thought was devoted to areas of clinical management, such as palliative care and radiotherapy.
We also asked interns about their career aspirations, whether they had a previous degree, and to identify the university where they studied medicine.
A copy of the questionnaire is available from the corresponding author.
All interns at each selected institution were asked to complete the survey. A representative of the ACS Oncology Education Committee contacted the intern orientation supervisors to arrange the time and venue for administration of the survey.
Relationships between variables were examined via contingency tables and χ2 tests. Contingency tables (for all response categories) were used to calculate χ2 values and the appropriate P values for all analyses involving categorical variables. We adjusted for the clustering effect of the sampling method using Stata.5 As confidence intervals are not presented for all relevant point estimates in the earlier report,2 we were not able to compare our results directly with those from the previous study, which were adjusted for sampling design. We relied on unadjusted comparisons of the differences in proportions, although these might lead to Type 1 errors (ie, detecting significant differences where none exist). Therefore, differences in proportions between the two studies that are only marginally significant should be interpreted with some caution.
Twenty-nine hospitals were sampled in Australia and New Zealand. There were 443 (62%) responses from the 719 interns who were surveyed. Questionnaires were completed and returned by 17% to 100% of interns at the selected hospitals. Response rate was low in Victoria because of lack of cooperation in administering the survey at some hospitals rather than the respondents' choice not to reply. Fifteen doctors were not trained in Australia or New Zealand and 27 did not nominate their training university; these respondents were excluded, leaving 401 surveys for analysis.
There were 118 respondents from interns trained in GMP courses and 283 from non-graduate intake courses. Interns trained in GMP courses were over-represented relative to the proportion of people graduating from such programs (33% of interns from GMP courses responded, compared with 22% of interns from non-GMP courses; P = 0.001).
Certain subjects have become more prominent in the past decade. These "new" subjects include communication skills, critical appraisal, and multidisciplinary care. Relative to interns from non-GMP courses, GMP interns perceived themselves as having significantly better communications skills, such as competence in advising patients to give up smoking, being able to break bad news, and to discuss death with a dying patient. However, they reported lower competence at preparing patients for hazardous procedures (Box 1).
More GMP than non-GMP interns would refer a patient with breast cancer to a multidisciplinary clinic for management (83% versus 70%; P = 0.03). There was no difference between interns in their knowledge of support services such as domiciliary palliative care or consumer groups.
All GMP interns had received training in critical appraisal, as opposed to 80% of non-GMP interns (P = 0.02). There was no difference, however, in their knowledge of the existence of evidence-based guidelines or their knowledge about the evidence on screening for cancer.
GMP interns reported a higher perception of their ability to perform a cervical smear compared with interns from non-GMP courses (P = 0.005). There was no difference in their perceptions of their ability to recognise a melanoma (Box 1).
There were no significant differences between the GMP and non-GMP intern groups in their reports of exposure to patients with lung, lymphoma, breast, oral, prostate or rectal cancer (Box 2). However, for both groups, exposure to patients with specific cancers had declined when compared with 1990 interns (Box 2).
There were no differences between the GMP and non-GMP groups in the proportion who had never examined a primary tumour in breast (GMP, 18%; non-GMP, 13%), mouth (GMP, 65%; non-GMP, 63%), prostate (GMP, 56%; non-GMP, 60%) or rectal cancer (GMP, 52%; non-GMP 51%) or melanoma (GMP, 28%; non-GMP, 25%).
The time spent in clinical attachments in medical oncology, palliative care, radiation oncology or surgery did not differ in the interns in 2001 between the two types of medical course. Attendance at clinics in palliative care, radiation oncology and surgical oncology was greater in 2001 than in 1990 (Box 3).
Significantly more GMP interns stated (incorrectly) that the age at which a woman was at greatest risk of cervical cancer was in her 30s, and significantly fewer GMP interns stated correctly that the age at which a person was at greatest risk of colorectal cancer was in the 60s (Box 4). There was no significant difference in estimating the age at greatest risk for breast cancer.
When asked which cancers had a five-year survival of more than 50%, fewer GMP students gave the correct answer for breast, lung and testis (Box 5). There was no difference for Hodgkin's disease, ovary, prostate or rectal cancer.
GMP interns reported more contact with patients in their first year (97% versus 28%; P < 0.001), role play (98% versus 74%; P < 0.001), clinical skills laboratories (97% versus 73%, P < 0.001), problem-based learning (100% versus 79%; P < 0.005) and formative assessment (100% versus 89%; P = 0.01).
GMP students rated their teaching more highly than interns from non-GMP courses for smoking cessation ("good" or "very good", 45% versus 29%; P = 0.04), the management of incurable cancer ("good" or "very good", 24% versus 17%; P = 0.15) and dying patients ("good" or "very good", 26% versus 19%; P = 0.07). There were no significant differences in the interns' rating of their teaching about the primary prevention of cancer, screening for cancer or the management of curable cancer according to the course intake requirements (Box 6).
Although there may be biases with the self-reporting of perceptions of competence and knowledge, systematic differences between GMP and non-GMP interns, or with the interns of 1990, are unlikely. It is therefore reasonable to make comparisons between the groups.
As expected, interns from GMP courses reported better communication skills, and greater awareness of multidisciplinary cancer management and critical appraisal than their counterparts from non-GMP courses. All are areas of new content that were not universally implemented in our previous survey of curricula.6
GMP interns did not report lower competence in clinical skills, but did exhibit less knowledge of areas such as the age of greatest risk of developing cancer and the outcomes of treatment.
Clinical exposure in terms of cancer patients seen and time spent in cancer-related clinical attachments were similar between the interns of 2001, but experience was less than ideal — fewer than half the interns surveyed had ever examined a patient with rectal or prostate cancer. Although the proportion that had attended oncology clinics was greater in 2001 than in 1990, more respondents in 2001 had never examined a patient with melanoma, mouth cancer or rectal cancer.
Others have reviewed the relationship between GMP and traditional teaching and students' knowledge and skills and reported equivalent outcomes.7-9 Where differences have been reported, they have been attributed to different course content or student selection.7 Student satisfaction may be higher in GMP courses.10 In Australia, it appears that GMP curricula have successfully introduced new course material and new methods of teaching, but have not always succeeded in producing doctors with better knowledge about cancer. The interns' perceptions of their skills in performing common clinical tasks have improved, but their exposure to patients with common cancer types remains poor.
From the public's perspective, two disturbing results are the ratings given to quality of instruction, and how little exposure interns in 2001 had had to patients with cancer compared with their colleagues 10 years previously. Consumers, who pay for the training of doctors as well as being obliged to use their services, will hope that these findings prompt more thorough teaching of oncology, such as that outlined in the ACS Ideal Oncology Curriculum.4
1: Proportion of respondents reporting levels of skills, by type of medical program
Skill |
Perceived competence |
GMP (n = 118) |
Non-GMP (n = 283) |
P* |
|||||||
Advising to give up smoking |
Very high |
8% |
3% |
0.02 |
|||||||
High |
33% |
27% |
|||||||||
Medium |
54% |
54% |
|||||||||
Low |
5% |
16% |
|||||||||
Nil |
0 |
0 |
|||||||||
Preparing a patient for a hazardous procedure |
Very high |
3% |
2% |
0.002 |
|||||||
High |
15% |
32% |
|||||||||
Medium |
44% |
47% |
|||||||||
Low |
35% |
18% |
|||||||||
Nil |
3% |
1% |
|||||||||
Discussing death with a dying patient |
Very high |
4% |
3% |
0.02 |
|||||||
High |
27% |
10% |
|||||||||
Medium |
30% |
40% |
|||||||||
Low |
36% |
42% |
|||||||||
Nil |
3% |
5% |
|||||||||
Breaking bad news |
Very high |
4% |
2% |
0.04 |
|||||||
High |
29% |
18% |
|||||||||
Medium |
42% |
54% |
|||||||||
Low |
22% |
26% |
|||||||||
Nil |
3% |
0 |
|||||||||
Recognising a melanoma |
Very high |
1% |
1% |
0.1 |
|||||||
High |
15% |
22% |
|||||||||
Medium |
48% |
53% |
|||||||||
Low |
35% |
24% |
|||||||||
Nil |
1% |
0 |
|||||||||
Performing a cervical smear |
Very high |
1% |
1% |
0.005 |
|||||||
High |
17% |
13% |
|||||||||
Medium |
62% |
47% |
|||||||||
Low |
19% |
38% |
|||||||||
Nil |
1% |
1% |
|||||||||
P value by χ2 test with four degrees of freedom. |
|||||||||||
2: Exposure to cancer patients with specific tumours
Tumour site |
"Never examined patient" |
||||||||||
1990 (n = 389) |
GMP (n = 118) |
Non-GMP (n = 283) |
P* |
||||||||
Breast |
na |
8% |
6% |
|
|||||||
Lung |
na |
2% |
5% |
|
|||||||
Lymphoma |
na |
21% |
20% |
|
|||||||
Melanoma |
14% |
20% |
19% |
0.03 |
|||||||
Mouth |
20% |
56% |
53% |
<0.0001 |
|||||||
Rectum |
17% |
31% |
32% |
<0.0001 |
|||||||
na = not asked in 1990. * Comparison of all 2001 interns (n = 401) with 1990 interns (n = 389), χ2 with 1 degree of freedom. |
|||||||||||
3: Attendance at oncology clinics
"Never attended" |
|||||||||||
Clinic |
1990 (n = 389) |
2001 (n = 401) |
P |
||||||||
Medical oncology |
18% |
20% |
0.71 |
||||||||
Palliative care |
50% |
18% |
< 0.001 |
||||||||
Radiation oncology |
42% |
22% |
< 0.001 |
||||||||
Surgical oncology |
11% |
3% |
< 0.001 |
||||||||
4: Knowledge of the age at which a patient is at greatest risk of developing specific cancers
Specific cancer |
Age at greatest risk
|
Proportion of responses |
P |
||||||||
GMP (n = 118) |
Non-GMP (n = 283) |
||||||||||
Breast |
30s |
3% |
1% |
0.5 |
|||||||
40s |
11% |
12% |
|||||||||
50s |
46% |
49% |
|||||||||
60s |
40% |
37% |
|||||||||
Cervix |
30s |
30% |
16% |
0.02 |
|||||||
40s |
38% |
41% |
|||||||||
50s |
21% |
29% |
|||||||||
60s |
11% |
14% |
|||||||||
Colorectal |
30s |
0 |
0 |
0.04 |
|||||||
40s |
7% |
2% |
|||||||||
50s |
33% |
32% |
|||||||||
60s |
60% |
66% |
|||||||||
P value from χ2 test with three degrees of freedom. In each case the highest age group is the correct answer. |
|||||||||||
5: Proportion of respondents who gave the correct answer when asked which of the following cancers had a five-year survival of greater than 50% with appropriate treatment
Specific cancer (correct answer) |
Proportion giving the correct answer |
P |
|||||||||
GMP (n = 118) |
Non-GMP (n = 283) |
||||||||||
Duke's C carcinoma of the colon (No) |
81% |
80% |
0.7 |
||||||||
Operable non-small-cell carcinoma of the lung (No) |
44% |
61% |
0.002 |
||||||||
A 2-cm node-negative breast cancer (Yes) |
89% |
96% |
0.01 |
||||||||
Hodgkin's disease above and below the diaphragm (Yes) |
26% |
32% |
0.2 |
||||||||
Stage I cancer of the prostate (Yes) |
98% |
98% |
0.7 |
||||||||
Cancer of the ovary with peritoneal metastases (No) |
96% |
97% |
0.6 |
||||||||
Cancer of the testis with pulmonary metastases (Yes) |
26% |
37% |
0.04 |
||||||||
P value from χ2 test with three degrees of freedom. In each case the highest age group is the correct answer. |
|||||||||||
6: Respondents' perceptions of the quality of training in specific subjects
Specific topic |
Rating of quality of training |
Proportion of responses |
P |
||||||||
GMP (n = 118) |
Non-GMP (n = 283) |
||||||||||
Cancer primary prevention |
Very good |
4% |
4% |
0.9 |
|||||||
Good |
33% |
30% |
|||||||||
Reasonable |
55% |
56% |
|||||||||
Poor |
7% |
9% |
|||||||||
Very poor |
1% |
1% |
|||||||||
Cancer screening |
Very good |
2% |
5% |
0.07 |
|||||||
Good |
46% |
34% |
|||||||||
Reasonable |
44% |
53% |
|||||||||
Poor |
8% |
8% |
|||||||||
Very poor |
0 |
0 |
|||||||||
Assisting patients to stop smoking |
Very good |
9% |
6% |
0.04 |
|||||||
Good |
36% |
23% |
|||||||||
Reasonable |
41% |
46% |
|||||||||
Poor |
14% |
24% |
|||||||||
Very poor |
0 |
1% |
|||||||||
Management of potentially curable cancer |
Very good |
1% |
4% |
0.3 |
|||||||
Good |
28% |
25% |
|||||||||
Reasonable |
47% |
54% |
|||||||||
Poor |
23% |
17% |
|||||||||
Very poor |
1% |
0 |
|||||||||
Management of incurable cancer |
Very good |
0 |
2% |
0.15 |
|||||||
Good |
24% |
15% |
|||||||||
Reasonable |
46% |
45% |
|||||||||
Poor |
29% |
34% |
|||||||||
Very poor |
1% |
4% |
|||||||||
Management of symptoms in patients who are dying of cancer |
Very good |
1% |
3% |
0.07 |
|||||||
Good |
25% |
16% |
|||||||||
Reasonable |
43% |
38% |
|||||||||
Poor |
28% |
38% |
|||||||||
Very poor |
3% |
5% |
|||||||||
P value from χ2 test with four degrees of freedom. |
|||||||||||
- Michael B Barton1
- Sharon E Miles2
- Martin H Tattersall3
- Phyllis N Butow4
- Sally Crossing5
- Konrad Jamrozik6
- Bin Jalaludin7
- Christopher H Atkinson8
- 1 Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Health Service, Sydney, NSW.
- 2 University of Sydney, Sydney, NSW.
- 3 Cancer Voices NSW, Sydney, NSW.
- 4 Imperial College at Charing Cross, London, United Kingdom.
- 5 Epidemiology Unit, Liverpool Hospital, Sydney, NSW.
- 6 Oncology Service, Christchurch Hospital, Christchurch, New Zealand.
We thank the Cancer Council Australia for helping with the costs of the survey and are grateful to the hospitals that facilitated participation of their interns. Funding was kindly provided by the Cancer Council Australia for printing and postage.
None identified.
- 1. Mathers C, Vos T, Stevenson C. The burden of disease and injury in Australia. Canberra: AIHW, 1999. (AIHW PHE 17.)
- 2. Smith WT, Tattersall MHN, Irwig LM, Langlands AO. Undergraduate education about cancer. Eur J Cancer 1991; 27: 1448-1453.
- 3. Tattersall MH, Langlands AO, Simpson JS, Forbes JF. Undergraduate education about cancer: a survey in Australian medical schools. Eur J Cancer Clin Oncol 1988; 24: 467-471.
- 4. Oncology Education Committee. Ideal oncology curriculum for medical schools. Sydney: Australian Cancer Society, 1999.
- 5. Stata [computer program]. Version 6. College Station, TX: Stata Corporation, 1999.
- 6. Barton MB, Simons RG. A survey of cancer curricula in Australian and New Zealand medical schools in 1997. Oncology Education Committee of the Australian Cancer Society. Med J Aust 1999; 170: 225-227.
- 7. Colliver JA. Effectiveness of problem-based learning curricula: research and theory. Acad Med 2000; 75: 259-266.
- 8. Albanese MA, Mitchell S. Problem-based learning: a review of literature on its outcomes and implementation issues. Acad Med 1993; 68: 52-81.
- 9. Vernon DT, Blake RL. Does problem-based learning work? A meta-analysis of evaluative research. Acad Med 1993; 68: 550-563.
- 10. Prideaux D. Researching the outcomes of educational interventions: a matter of design. BMJ 2002; 324: 126-127.
Abstract
Objective: To compare the cancer knowledge and skills of interns in 2001 who graduated from graduate medical program (GMP) courses with those from non-GMP courses, and to compare the cancer knowledge and skills of interns in 2001 with those who completed a similar survey in 1990.
Design: Questionnaire survey of recently graduated interns in a random sample of Australian and New Zealand hospitals. The questionnaire was designed to allow direct comparison with the 1990 survey, and was guided by the Australian Cancer Society's Ideal Oncology Curriculum for Medical Schools.
Results: 443 interns completed the survey (response rate, 62%; 42 were excluded, leaving 401 surveys for analysis: 118 from GMP courses and 283 from non-GMP courses). Interns from GMP courses felt more competent than those from non-GMP courses at discussing death (P = 0.02), breaking bad news (P = 0.04) and advising on smoking cessation (P = 0.02), but less competent at preparing a patient for a hazardous procedure (P = 0.02). More GMP interns would refer a breast cancer patient to a multidisciplinary clinic (83% versus 70%; P = 0.03). Knowledge about cancer risks and prognosis was significantly less in GMP interns, but GMP interns rated their clinical skills, such as taking a Pap smear, higher than non-GMP interns. The GMP and non-GMP groups did not differ in their exposure to cancer patients, but compared with 1990 interns recent graduates had less exposure to patients with cancer.
Conclusions: GMP curricula appear to have successfully introduced new course material and new methods of teaching, but have not always succeeded in producing doctors with better knowledge about cancer. Recent graduates have less exposure to cancer patients than those who trained 10 years ago.