Pathways to the diagnosis of epithelial ovarian cancer in Australia

Ovarian cancer, the sixth most common cause of cancer death among Australian women,1 has a poor prognosis, mainly because about 75% of cancers are diagnosed at an advanced stage when treatment is unlikely to be curative.2 The symptoms of ovarian cancer are often non-specific,3,4 and delays in diagnosis may, in part, explain why the disease has so often spread beyond the ovaries at diagnosis.5 Anecdotally, there is a perception that the journey from first presentation to diagnosis is often long and circuitous for women with ovarian cancer.

Few international and no Australian studies have attempted to formally document the diagnostic experience of women with ovarian cancer.5-7 While some have endeavoured to estimate diagnostic delays, estimates of time from first presentation to diagnosis are varied, with British and Swedish studies reporting that most women are diagnosed within 1 month,6,8,9 while others from the United States5 and Norway7 found that 40%–45% of women experienced delays of 3 or more months. As health system differences between countries probably explain some of this variation, it is important for Australian clinicians and policymakers to have access to information from the Australian setting.

Our aims were to describe the diagnostic pathways experienced by a large, representative group of Australian women with ovarian cancer and to document the time between their first presentation to a medical professional and clinical diagnosis.

The Australian Ovarian Cancer Study was an Australia-wide population-based study. Eligible participants were women aged 18–79 years with suspected invasive or borderline epithelial ovarian, fallopian tube or primary peritoneal cancer who were identified between January 2002 and June 2005 through gynaecological oncology units and state-based cancer registries.10 A total of 3550 women were identified; of these, 307 (8.6%) died before contact could be made, 194 (5.5%) could not be contacted, and the treating doctor refused contact with 133 (3.7%). A further 171 (4.8%) were excluded because they could not complete the questionnaires in English (70), could not give informed consent (35) or were too sick (66). The remaining 2745 women were invited to participate (most before surgery) and, of these, 2319 (84% of those approached) agreed to take part. After surgery, a further 634 women were excluded: 608 because their final diagnosis was not epithelial ovarian cancer, 25 because their cancer was diagnosed before the study period, and one because she was not an Australian resident.

The remaining 1685 participants completed a health questionnaire, and 1463 of these (87%) were also interviewed by research nurses by telephone (median time between histological diagnosis and interview was 3 months). This group formed the sample for our analysis. They were asked what symptoms prompted them to go to a doctor or, if they did not have symptoms, the reason for the consultation that led to their cancer diagnosis. Women were asked about each doctor consulted, dates of visits, tests done and treatment or advice given. The reported dates of first presentation and clinical diagnosis (ie, when the woman was first told she had ovarian cancer, not the date of histological diagnosis) were used to estimate the interval between presentation and diagnosis.

The numbers of symptoms reported, doctors seen and total doctor visits were all highly skewed. These variables were therefore summarised using means and medians, and groups were compared using the non-parametric log-rank test. We used χ2 tests to investigate associations between a variety of factors (age, family history of cancer, area of residence, education level, and types and duration of symptoms) and whether women were sent for investigations (yes/no) or referred to a gynaecologist, gynaecological oncologist or other specialist. Multiple logistic regression analysis was used to investigate factors associated with an interval of > 6 months between first presentation and diagnosis. All statistical tests were two-sided and a P value < 0.05 was considered statistically significant. All analyses were conducted using SAS statistical software (version 9.1; SAS Institute Inc, Cary, NC, USA).

This study was approved by the Human Research Ethics Committees at the Queensland Institute of Medical Research and all participating centres.

Characteristics of the 1463 women are shown in Box 1. Overall, 145 women (10%) reported that their cancer was diagnosed incidentally, most often at a routine check-up (50%) or when consulting for another condition (41%). These women were excluded from further analyses.

Our analyses suggest that the majority of symptomatic women with ovarian cancer presenting to a medical practitioner in Australia are investigated and diagnosed promptly. About 42% of women in our study were either given a diagnosis or referred to a gynaecologist, gynaecological oncologist or oncologist as a result of their first medical consultation, and this figure increased to 61% when repeat visits to the same doctor were included. Overall, 66% of women were diagnosed in a month or less and 80% within 3 months of their initial presentation. However, women reported seeing a wide variety of doctors, and just over 10% reported that the diagnostic process took more than 6 months. This delay was more common for women with lower incomes, those who lived in remote areas of Australia and those with abdominal or bowel symptoms. It is likely that some of the women reporting delays of several years had attributed unrelated symptoms to their ovarian cancer.

To our knowledge, this is the first study to describe in detail the diagnostic pathways experienced by women with ovarian cancer in Australia. Our results with respect to time to diagnosis are in close accord with another Australian study that found that 70% of women with ovarian cancer were diagnosed within 3 months of first experiencing symptoms.11

We found a strong association between residing in a remote area of Australia and being diagnosed more than 6 months after presentation, although the number of women involved was small. Women from remote areas may have poorer access to primary care services12 and make fewer visits to GPs and specialists than women living in urban areas.13 Access inequalities may also explain why women with lower incomes were more likely to have longer diagnostic delays than women with higher incomes, as evidence suggests that, in Australia, all else being equal, those with higher incomes are more likely to consult a specialist than those with a lower income.14

We also found that the number, type and duration of symptoms a woman presented with were associated with whether she was diagnosed more than 6 months after presentation. These findings largely concur with other studies that have investigated determinants of diagnostic delay in patients with ovarian cancer.5,7,15 Clinicians faced with a woman presenting with several non-gynaecological symptoms probably consider more common diagnoses associated with those symptoms in the first instance, thus potentially delaying diagnosis. However, we did not have information directly from medical practitioners and thus cannot provide insights into how doctors make clinical judgements when faced with the complex symptoms and signs associated with ovarian cancer.

Our study was large and included women from rural and metropolitan areas in every Australian state and territory. However, one limitation is that 14% of women ascertained as having suspected ovarian cancer were unable to participate due to illness or death, and a greater proportion of such women were identified through cancer registries rather than gynaecological oncology clinics. It is likely that a greater proportion of the women we were able to interview were ultimately treated by a gynaecological oncologist than is usual in the broader Australian population, and it is therefore possible that their diagnostic experiences differed from the women who did not participate. As our results suggested that women with higher-stage disease saw more doctors and had more doctor visits before diagnosis, it is possible that women who had died or were too sick to participate also had a more circuitous path to diagnosis. A further limitation is that our study was based on self-report, so women may have inaccurately recalled details of their diagnostic experience. However, it is unlikely that such inaccuracies would have systematically affected our findings. Finally, clinical practices may have changed since these data were collected between 2002 and 2005, although it seems unlikely that any such changes would result in greater diagnostic delays.

In conclusion, our study provides reassurance that, despite anecdotal evidence to the contrary, most women with ovarian cancer in Australia are diagnosed promptly once they present to a medical practitioner. For a small group of women (about 10%), diagnosis can take more than 6 months, and we have identified several factors associated with a longer delay. Further studies addressing these factors, especially lack of access to care, are warranted.

4 The referral pathways for 1318 women diagnosed with ovarian cancer in Australia

* Includes 13 women who reported seeing another clinician (n = 12) or other health care professional (n = 1) before consulting their GP. † Another 29 women (2.2%) did not present to their GP but to a gynaecologist (n = 16), an oncologist (n = 4) or another health professional (n = 9), and they are included from these points. ‡ Includes all doctors other than a gynaecologist, gynaecological oncologist or oncologist.