Timely diagnosis and treatment of any cancer is a desire shared by patients, clinicians and politicians alike. This is based on strong belief that that early diagnosis and treatment of the cancer will positively impact on the patients overall outcome.1 Some studies to date have shown that diagnostic delays in cancers adversely affect outcome in studies about breast, lung, head and neck cancers.2 Endometrial cancer is the one of the most common gynaecological cancers in New Zealand, with approximately 315 new cases and 80 deaths being reported each year.4 Patients commonly present to medical practitioners with post-menopausal bleeding (PMB), and its relatively slow oncologic velocity and well defined referral pathway for PMB at Southern District Health Board makes it an ideal candidate to investigate delays. Within the literature, only a few studies to date have specifically looked at delays in diagnosis and treatment of gynaecological malignancies. A recent study by Vandborg et al5 showed median time from patient presentation at a GP clinic to initiation of treatment across all gynaecological cancers to be 14.4 weeks. Another Denmark-based national survey by Robinson et al showed median total timeframe (date of first cancer-related symptom to date of operation at gynaecological surgical centre) to be 12 weeks.6 Available New Zealand specific data includes a retrospective audit at Christchurch Womens Hospital in 2009 for all gynaecological cancers which showed median time from initial patient presentation to treatment to be 7.8 weeks, with time from referral to diagnosis at 3.8 weeks, diagnosis to treatment plan at 1.4 weeks, and treatment plan to surgery at 2.6 weeks.7 The Ministry of Health New Zealand (MoHNZ) is proposing target timeframes for cancers based on three indicators as shown on the figure one below.3 Hence the information regarding current performance of healthcare delivery in cancers is essential in putting the proposed target timeframes into perspective and to identify areas where improvements can be made. This audit aims to quantify the time taken in the diagnostic and treatment process for endometrial cancer at Dunedin Hospital, a tertiary hospital with the associated Medical School of the University of Otago, where a visiting Gynaecological Oncologist from Christchurch comes for clinics, combined gynaecology-oncology multidisciplinary meeting and surgery every 2 weeks. A second aim of this audit is to identify factors along this pathway contributing towards significant delays if present. Figure 1. Ministry of Health New Zealand indicators of best practice for diagnosis and treatment of all cancers Method The pathway of primary care referral, investigation and treatment for this audit was based on a one-directional pathway model as described by the MoHNZ.3 Patients discussed at a Gynaecology-Oncology Multi-Disciplinary Meeting (GOMDM) at Dunedin Hospital during 2008-2011 for primary endometrial cancer were included, if they were initially referred by their general practitioner or their private gynaecologist. Patients were excluded from the audit if they were referred from Southland hospital, or other than their general practitioner/private gynaecologist, as information was not readily available on these patients. The iSOFT clinical patient database, patient management system (iPM) and clinical notes where necessary were used to collect information required for this audit, which were; patient age; date when the initial referral was first received; type and duration of presenting complaint on initial referral; date of first specialist appointment (FSA); date of Pipelle biopsy and/or hysteroscopy and dilatation and curettage (D&C); date of staging CT and/or MRI; when the patients were first waitlisted for their treatment; FIGO (Federation of International Gynaecologists and Obstetricians) grade and stages of the cancer; date of multidisciplinary meeting when definitive management plan was decided; and finally the date when first medical/surgical treatment was initiated. Based on the dates collected, time duration between different points of the diagnosis and treatment pathway were calculated in days. The date when patients were waitlisted for surgery was regarded as the date of treatment decision. Descriptive statistics were used to summarize the data. The 10 longest outliers were further identified and information on their demographic details, medical comorbidities, and number of did-not-attended appointments (DNA), and waitlist urgency criteria for hysteroscopy were collected to identify reasons which may have contributed towards significant delay in this patient group. Results Eighty-seven cases of suspected endometrial cancers discussed at GOMDM during 2008-2011 were identified. Based on the exclusion criteria, 40 cases were excluded: 19 cases for being Southland hospital referrals; 7 cases due to final diagnosis not being endometrial carcinoma; 6 cases for being referrals other than from GPs or private gynaecologists; 5 cases for being re-referrals on the basis of previously diagnosed endometrial carcinoma; and finally 3 cases due to diagnosis of endometrial cancer being made outside the 2008-2011 period. Three further cases were not included in the analysis due to unavailable records. Table 1 describes demographic details of the included patients. After exclusions, 44 patients were included in the final audit. Mean age of the patients was 64 years (range 41-89), with the majority of the patients initially being referred by their general practitioners. 64% were referrals within the geographical boundaries of Dunedin city, and rest were referrals from outside Dunedin (which were defined as rural areas). Majority of the patients (86%) were post-menopausal women. The most common reason for referral to the specialist clinic was post-menopausal bleeding (75%), followed by incidental finding of abnormal cells on cervical smear (11%). In cases where patients were experiencing a specific complaint such as post-menopausal bleeding, the time from beginning of the complaint to presentation at a GP clinic varied between 1 to 24 months, with about one-third of the patients presenting within 6 months of symptom onset, and another third presenting after more than 6 months. Table 1. Demographic details of the patients Variables N % Age, mean (range) 64 (41-89) N/A Location Dunedin city Rural 28 16 64% 36% Original referrer General practitioner Private gynaecologist Unknown 38 4 2 86% 9% 5% Menopausal state Pre/Peri-menopausal Post-menopausal 6 38 14% 86% Reason for referral Post-menopausal bleeding Incidental finding on cervical smear/ultrasound scan Other 33 5 6 75% 11% 14% Duration of complaint Less than 6 months More than 6 months Unknown duration 16 14 14 36% 32% 32% Seen at first specialist assessment by Consultant Registrar 26 19 59% 41% In terms of diagnostic methods and cancer characteristics, (Table 2), 22 patients (51%) had pipelle biopsy, while 18 patients (42%) underwent hysteroscopy and D&C. Three patients (7%) who had pipelle biopsy done further underwent D&C for diagnosis due to inadequate samples obtained from the initial pipelle biopsy. Amongst 18 patients who underwent hysteroscopy and D&C, pipelle biopsy was attempted but unsuccessful in nine patients due to tight cervical os preventing the procedure. Pipelle biopsy was not attempted in the rest of the patients prior to D&C due to reasons which included patients having had successful or unsuccessful pipelle biopsy done by their private gynaecologist prior to referral, patients who have never been sexually active, or reasons otherwise unspecified from medical notes. Ultrasound scan of the pelvis during the diagnostic process was done in the majority of patients (82%). In total, 61% of these scans were done prior to FSA with 21% of them being done following FSA. Both CT and MRI of the pelvis were utilised with similar frequency, with 41% of patients having had MRI of the pelvis (which is routinely performed in Dunedin for G1 adenocarcinomas for preoperative assessment of depth of infiltration) and 36% of them receiving CT of the pelvis. In terms of cancer characteristics, the majority of the cancers were of low grade at grade 1 (66%) and stage IA (52 %), and of adenocarcinoma in origin (91%). Table 2. Diagnostic methods and cancer characteristics Variables N % Biopsy method Pipelle biopsy Hysteroscopy and D&C Both Unknown 22 18 3 1 51% 42% 7% 2% Radiological imaging CT MRI Both CT & MRI Unknown 16 18 7 3 36% 41% 16% 7% Ultrasound scan of pelvis Before FSA After FSA Unknown 27 9 8 61% 21% 18% FIGO grading G1 G2 G3 29 10 5 66% 23% 11% Revised 2010 FIGO Staging IA IB IIA IIIA IIIC IVB 23 10 4 1 5 1 52% 24% 10% 2% 12% 2% Histological diagnosis Adenocarcinoma Endometrioid Clear cell Serous Carcinosarcoma 40 32 5 3 4 91% 80% 13% 8% 9% Figure 2 shows the summary of the time taken between different components of the diagnostic pathway for endometrial cancer in patients included in this audit. Numbers represent the median time taken in days with ranges shown in brackets. Table 3 shows the definitions of each timeframes. Comparing the results of this audit to the MoHNZ recommendations, excess delay was present for time taken from referral to treatment initiation (93 days vs. 62 days), and a lesser delay was present from referral to FSA (21 days vs. 14 days). The waiting time for staging scans after confirmed diagnosis of cancer also appeared long with a median time of 22 days, contributing to the length of delay in FSA to treatment decision timeframe. The time taken between treatment decisions to treatment initiation was shown to be shorter in Dunedin Hospital compared to MoHNZ cancer proposals. (19 days vs. 31 days). Figure 2. Median times in days (range) between different components of the diagnosis and treatment pathway Table 3 shows that 75% of patients were treated within the MoHNZ best practice indicator 3 (time duration between waiting listed to first definitive treatment). In contrast, only 20% and 32% of patients were seen within the best practice timeframe for indicator 1 (time from referral to first treatment initiation) and indicator 2 (time from referral to FSA) respectively. Table 3. Proportion of patients meeting/not meeting MoHNZ best practice recommendations Indicator Recommendation met N (%) Recommendation not met N (%) Information unavailable N (%) Indicator 1 (62 days) Indicator 2 (14 days) Indicator 3 (31 days) 9 (20%) 14 (32%) 33 (75%) 29 (66%) 25 (57%) 6 (14%) 6 (14%) 5 (11%) 5 (11%) Figure 3. Comparison of times between pipelle biopsy vs. hysteroscopy and D&C groups When timeframes between patients who received pipelle biopsy for diagnosis as opposed to hysteroscopy and D&C were compared (Figure 3), timeframes were longer in the hysteroscopy and D&C group for REF-FSA (38.5 days vs. 19.5 days) and FSA-DIAG (53 days vs. 0 days). As a result of these initial differences, the Hysteroscopy and D&C group had considerably longer timeframes for FSA-TREAT and REF-TREAT when compared to the pipelle biopsy group. DIAG-SCAN, WAIT-TREAT, and SCAN-TREAT timeframes were similar for both groups. Ten patients with longest timeframes were further selected and their clinical history was looked into order to identify factors which contributed towards delays. A systemic aspect contributing to the delay within this patient group was the waiting time for hysteroscopy and D&C, where it took between 53 days to 136 days in the five longest outliers despite the hysteroscopy priority score (80-100) indicating the urgency for the procedure (80-100 out of 100 according to Clinical Priority Assessment Criteria score system). An important patient-related factor was multiple did-not-attends (DNAs) identified in two outliers, which likely mirrored patients denial of a potential cancer diagnosis. Time taken for patient to consider and agree to a treatment plan was also another significant factor with one patients case taking 103 days for treatment to be initiated after the staging scan. The most important clinical factors were insufficient or inconclusive histological Pipelle samples for diagnosis and documented difficult examination at FSA, which were present amongst five out of ten longest outliers. Other possible factors identified from these patients were rural referrals (2 out of the 10 longest outliers), multiple medical comorbidities requiring lengthy medical review prior to operation, and low clinical suspicion on FSA referral with 9 out of 10 longest outliers having been triaged as semi - urgent (to be seen within 6-12 weeks) or even lower priority. Discussion This audit aimed to quantify the duration of time taken for diagnosis and treatment of endometrial carcinoma at Dunedin Hospital New Zealand, and to identify areas where significant delays may be present. When compared to the findings from similar audit done at Christchurch Womens Hospital in 2009, times from referral to diagnosis were considerably longer in Dunedin Hospital (93 days vs. 26.6 days), but similar in treatment decision to treatment initiation (19 days vs. 18.2 days). It is important to note however that the Christchurch audit included all gynaecological cancers, not solely endometrial carcinomas; hence it is difficult to make a direct comparison of results. Symptom duration experienced by patients before presenting to their GP or private gynaecologist was seen to vary between 1 to 24 months. It is worth keeping mind that the symptom duration is not included in the calculation timeframes. Patient related delay therefore may contribute significantly towards overall delay in diagnosis and treatment of endometrial cancer. The delays seen compared to MoHNZ recommendations could be due to factors specific to Dunedin Hospital which is part of the Southern District Health Board (DHB). The Southern DHB has a land area of over 62,356 sq km making it geographically the largest DHB region in New Zealand. 7.1% of the national population live within the Southern DHB catchment area, with 59% of the catchment population residing outside Dunedin City.8 The nature of distribution of population and the large geographic catchment area means effective health resource distribution is challenging. Although referrals from Southland Hospital were excluded in this audit, patients referred from Oamaru, Dunstan, and other rural areas were included which may have contributed towards the delays seen in some patients. Attempts were made to compare time delays between referrals from urban Dunedin and rural areas; however results were not statistically significant, likely due to low power. As shown previously in Figure two, median times were much longer in the hysteroscopy and D&C group compared to the Pipelle biopsy group in REF-FSA time (38.5 days vs. 19.5 days), FSA-DIAG time (53 days vs. 0 days), FSA-TREAT time (91 days vs. 51.5 days), and REF-TREAT time (105.5 days vs. 76 days). Understandably, the longer duration in FSA-DIAG timeframe in the hysteroscopy and D&C group is likely to have been due to patients having to wait until the procedure could be done, while pipelle biopsies were able to be taken either at GP clinics prior to FSA or on the day of the FSA, which explains the median time of 0 days in FSA-DIAG timeframe for Pipelle biopsy group. These results suggest the waiting time for hysteroscopy and D&C in a patient for histological diagnosis is a prominent factor in timely diagnosis and treatment of endometrial cancer. Measures such as supporting the approach of GPs performing pipelle biopsies by increased training of general practitioners and streamlining of referral processes for hysteroscopy and D&C in suspected endometrial carcinoma may prove to be effective in reducing the delays seen. Inability to perform pipelle biopsy due to tight cervical os was also seen to be a common barrier, with 50% of the patients who had hysteroscopy and D&C requiring it for this reason. Difficulty in insertion of the pipelle is often encountered in endometrial biopsy with one study showing that even after excluding women with cervical stenosis, difficulty in passing pipelle through the cervix was experienced in 41.7% of the patients.9 Some have suggested the use of misoprostol for cervical ripening prior to pipelle biopsy as a means to improve the success rate. However there has been only one small randomized controlled trial to date,10 which showed little benefit of 400mcg oral misoprostol prior to biopsy, and even noted an increased incidence of pain and cramping during the procedure compared to controls. A larger randomised controlled trial on the topic is hence warranted. Hysteroscopy and directed biopsy, given its superior diagnostic potential compared to other endometrial biopsy methods, has been promoted to provide precise diagnosis of intrauterine pathologies even when provided in an office setting with a narrow scope.12 The option of an assessment of patients with post-menopausal bleeding at their first FSA appointment in an office hysteroscopy setting has recently become available in Dunedin hospital, and may help to shorten the timeframe to diagnosis as ideally a hysteroscopy and curettage can a be performed at the first FSA appointment. Shortages in access to theatre and imaging for patients to yet be diagnosed with cancer are additional aspects of delayed diagnosis and treatment of cancer patients. The new best practice standards of the Ministry of Health might improve the situation for cancer patients as they are likely to receive higher priority in accessing these resources, however it does not address the actual overall shortage in these areas. Factors identified from the 10 longest outliers have also been identified by other studies within the literature. A study by Robinson et al11 investigated the association between socioeconomic factors and diagnostic delays among all gynaecological cancers, and found that rural patients were at greater risk of experiencing delays compared to urban counterparts in time taken from GP referral to first specialist appointment (odds ratio [OR]=2.20). The presence of medical comorbidities was also associated with increased risk of experiencing secondary-care related delays (OR=1.66). Interestingly, younger patient age was associated with overall increased risk of experiencing delays, and the authors have suggested a few reasons for this; the main points being younger women feeling less urgency to contact their medical practitioner with initial onset of symptoms, less suspicion of cancer from referring doctors due to the patient demographics, and the fact that younger women are more likely to be working hence may find difficulty in scheduling a time for an appointment. While factors such as inconclusive histological samples and difficult gynaecological examinations may not be modifiable, factors such as multiple DNAs may be due to the patients poor understanding of the significance of symptoms such as post-menopausal bleeding. No studies to date have specifically explored New Zealand publics perception of risk in post-menopausal bleeding. Hence, such research may be helpful in determining whether a stronger emphasis on patient education is required in order to reduce impact of patient-related factors such as multiple DNAs in diagnostic delay of endometrial carcinoma. Internationally, specific cancer timeframe recommendations such as NICE guidelines in United Kingdom have demonstrated improvement in faster cancer diagnosis since its implementation13 and some centres are now adopting multi-centred accreditation process14 to standardise care for cancer patients. The Faster Cancer Treatment programme of the Ministry of Health NZ aims to improve services by standardising care pathways and timeliness of services for cancer patients throughout New Zealand. The four main focus areas of this programme are: faster cancer treatment indicators, patient pathway coordination, tumour specific standards and multidisciplinary meetings. The planned collection of consistent data throughout the country will provide up-to-date information about the appropriateness of our current target timeframes as well as aid in its development, and hopefully improve in the context of work done in parallel within the other three main foci of work. This audit is a first assessment of time taken for referral, diagnosis and treatment of endometrial cancer in Dunedin Hospital. Although seemingly small in size, this single-centre based audit involved comparable caseload of endometrial carcinomas to the diagnostic delay study carried out by Vandborg et al5 as well as the aforementioned 2009 Christchurch Womens Hospital Audit. Given its retrospective design however, this audit is affected by bias from missing data and it is possible that some patients with endometrial cancer were lost to follow up, therefore affecting the final result. The results are specific to Dunedin Hospital, therefore is not readily generalizable to other regions in New Zealand, however it is likely that the problem of delay in diagnosis and treatment does not involve Dunedin Hospital alone. This audit specifically looked at patients diagnosed with endometrial carcinoma; hence patients who may have been investigated for PMB but were subsequently found not to have endometrial cancer were not included. Therefore, further research looking into timeframes for patients presenting with post-menopausal bleeding may provide more robust information regarding our current PMB diagnostic pathway performance. Conclusion The results of this audit show that the best practice times as proposed by MoHNZ for timely cancer diagnosis and treatment have not been met for patients with endometrial cancer, specifically with regards to time from initial referral to treatment, and to a certain extent from initial referral to first specialist appointment. Once the diagnosis was made, treatment was usually within the appropriate timeframe. The main factor contributing to the delays which could be identified was waiting times for hysteroscopy and D&C for histological diagnosis. Better education of patients about post-menopausal bleeding, training of general practitioners in pipelle biopsy, expedited hysteroscopy and D&C in an office gynaecology setting, and reduced waiting times for imaging and surgery could be possible ways to address the delays.

Timely diagnosis and treatment of any cancer is a desire shared by patients, clinicians and politicians alike. This is based on strong belief that that early diagnosis and treatment of the cancer will positively impact on the patients overall outcome.1 Some studies to date have shown that diagnostic delays in cancers adversely affect outcome in studies about breast, lung, head and neck cancers.2 Endometrial cancer is the one of the most common gynaecological cancers in New Zealand, with approximately 315 new cases and 80 deaths being reported each year.4 Patients commonly present to medical practitioners with post-menopausal bleeding (PMB), and its relatively slow oncologic velocity and well defined referral pathway for PMB at Southern District Health Board makes it an ideal candidate to investigate delays. Within the literature, only a few studies to date have specifically looked at delays in diagnosis and treatment of gynaecological malignancies. A recent study by Vandborg et al5 showed median time from patient presentation at a GP clinic to initiation of treatment across all gynaecological cancers to be 14.4 weeks. Another Denmark-based national survey by Robinson et al showed median total timeframe (date of first cancer-related symptom to date of operation at gynaecological surgical centre) to be 12 weeks.6 Available New Zealand specific data includes a retrospective audit at Christchurch Womens Hospital in 2009 for all gynaecological cancers which showed median time from initial patient presentation to treatment to be 7.8 weeks, with time from referral to diagnosis at 3.8 weeks, diagnosis to treatment plan at 1.4 weeks, and treatment plan to surgery at 2.6 weeks.7 The Ministry of Health New Zealand (MoHNZ) is proposing target timeframes for cancers based on three indicators as shown on the figure one below.3 Hence the information regarding current performance of healthcare delivery in cancers is essential in putting the proposed target timeframes into perspective and to identify areas where improvements can be made. This audit aims to quantify the time taken in the diagnostic and treatment process for endometrial cancer at Dunedin Hospital, a tertiary hospital with the associated Medical School of the University of Otago, where a visiting Gynaecological Oncologist from Christchurch comes for clinics, combined gynaecology-oncology multidisciplinary meeting and surgery every 2 weeks. A second aim of this audit is to identify factors along this pathway contributing towards significant delays if present. Figure 1. Ministry of Health New Zealand indicators of best practice for diagnosis and treatment of all cancers Method The pathway of primary care referral, investigation and treatment for this audit was based on a one-directional pathway model as described by the MoHNZ.3 Patients discussed at a Gynaecology-Oncology Multi-Disciplinary Meeting (GOMDM) at Dunedin Hospital during 2008-2011 for primary endometrial cancer were included, if they were initially referred by their general practitioner or their private gynaecologist. Patients were excluded from the audit if they were referred from Southland hospital, or other than their general practitioner/private gynaecologist, as information was not readily available on these patients. The iSOFT clinical patient database, patient management system (iPM) and clinical notes where necessary were used to collect information required for this audit, which were; patient age; date when the initial referral was first received; type and duration of presenting complaint on initial referral; date of first specialist appointment (FSA); date of Pipelle biopsy and/or hysteroscopy and dilatation and curettage (D&C); date of staging CT and/or MRI; when the patients were first waitlisted for their treatment; FIGO (Federation of International Gynaecologists and Obstetricians) grade and stages of the cancer; date of multidisciplinary meeting when definitive management plan was decided; and finally the date when first medical/surgical treatment was initiated. Based on the dates collected, time duration between different points of the diagnosis and treatment pathway were calculated in days. The date when patients were waitlisted for surgery was regarded as the date of treatment decision. Descriptive statistics were used to summarize the data. The 10 longest outliers were further identified and information on their demographic details, medical comorbidities, and number of did-not-attended appointments (DNA), and waitlist urgency criteria for hysteroscopy were collected to identify reasons which may have contributed towards significant delay in this patient group. Results Eighty-seven cases of suspected endometrial cancers discussed at GOMDM during 2008-2011 were identified. Based on the exclusion criteria, 40 cases were excluded: 19 cases for being Southland hospital referrals; 7 cases due to final diagnosis not being endometrial carcinoma; 6 cases for being referrals other than from GPs or private gynaecologists; 5 cases for being re-referrals on the basis of previously diagnosed endometrial carcinoma; and finally 3 cases due to diagnosis of endometrial cancer being made outside the 2008-2011 period. Three further cases were not included in the analysis due to unavailable records. Table 1 describes demographic details of the included patients. After exclusions, 44 patients were included in the final audit. Mean age of the patients was 64 years (range 41-89), with the majority of the patients initially being referred by their general practitioners. 64% were referrals within the geographical boundaries of Dunedin city, and rest were referrals from outside Dunedin (which were defined as rural areas). Majority of the patients (86%) were post-menopausal women. The most common reason for referral to the specialist clinic was post-menopausal bleeding (75%), followed by incidental finding of abnormal cells on cervical smear (11%). In cases where patients were experiencing a specific complaint such as post-menopausal bleeding, the time from beginning of the complaint to presentation at a GP clinic varied between 1 to 24 months, with about one-third of the patients presenting within 6 months of symptom onset, and another third presenting after more than 6 months. Table 1. Demographic details of the patients Variables N % Age, mean (range) 64 (41-89) N/A Location Dunedin city Rural 28 16 64% 36% Original referrer General practitioner Private gynaecologist Unknown 38 4 2 86% 9% 5% Menopausal state Pre/Peri-menopausal Post-menopausal 6 38 14% 86% Reason for referral Post-menopausal bleeding Incidental finding on cervical smear/ultrasound scan Other 33 5 6 75% 11% 14% Duration of complaint Less than 6 months More than 6 months Unknown duration 16 14 14 36% 32% 32% Seen at first specialist assessment by Consultant Registrar 26 19 59% 41% In terms of diagnostic methods and cancer characteristics, (Table 2), 22 patients (51%) had pipelle biopsy, while 18 patients (42%) underwent hysteroscopy and D&C. Three patients (7%) who had pipelle biopsy done further underwent D&C for diagnosis due to inadequate samples obtained from the initial pipelle biopsy. Amongst 18 patients who underwent hysteroscopy and D&C, pipelle biopsy was attempted but unsuccessful in nine patients due to tight cervical os preventing the procedure. Pipelle biopsy was not attempted in the rest of the patients prior to D&C due to reasons which included patients having had successful or unsuccessful pipelle biopsy done by their private gynaecologist prior to referral, patients who have never been sexually active, or reasons otherwise unspecified from medical notes. Ultrasound scan of the pelvis during the diagnostic process was done in the majority of patients (82%). In total, 61% of these scans were done prior to FSA with 21% of them being done following FSA. Both CT and MRI of the pelvis were utilised with similar frequency, with 41% of patients having had MRI of the pelvis (which is routinely performed in Dunedin for G1 adenocarcinomas for preoperative assessment of depth of infiltration) and 36% of them receiving CT of the pelvis. In terms of cancer characteristics, the majority of the cancers were of low grade at grade 1 (66%) and stage IA (52 %), and of adenocarcinoma in origin (91%). Table 2. Diagnostic methods and cancer characteristics Variables N % Biopsy method Pipelle biopsy Hysteroscopy and D&C Both Unknown 22 18 3 1 51% 42% 7% 2% Radiological imaging CT MRI Both CT & MRI Unknown 16 18 7 3 36% 41% 16% 7% Ultrasound scan of pelvis Before FSA After FSA Unknown 27 9 8 61% 21% 18% FIGO grading G1 G2 G3 29 10 5 66% 23% 11% Revised 2010 FIGO Staging IA IB IIA IIIA IIIC IVB 23 10 4 1 5 1 52% 24% 10% 2% 12% 2% Histological diagnosis Adenocarcinoma Endometrioid Clear cell Serous Carcinosarcoma 40 32 5 3 4 91% 80% 13% 8% 9% Figure 2 shows the summary of the time taken between different components of the diagnostic pathway for endometrial cancer in patients included in this audit. Numbers represent the median time taken in days with ranges shown in brackets. Table 3 shows the definitions of each timeframes. Comparing the results of this audit to the MoHNZ recommendations, excess delay was present for time taken from referral to treatment initiation (93 days vs. 62 days), and a lesser delay was present from referral to FSA (21 days vs. 14 days). The waiting time for staging scans after confirmed diagnosis of cancer also appeared long with a median time of 22 days, contributing to the length of delay in FSA to treatment decision timeframe. The time taken between treatment decisions to treatment initiation was shown to be shorter in Dunedin Hospital compared to MoHNZ cancer proposals. (19 days vs. 31 days). Figure 2. Median times in days (range) between different components of the diagnosis and treatment pathway Table 3 shows that 75% of patients were treated within the MoHNZ best practice indicator 3 (time duration between waiting listed to first definitive treatment). In contrast, only 20% and 32% of patients were seen within the best practice timeframe for indicator 1 (time from referral to first treatment initiation) and indicator 2 (time from referral to FSA) respectively. Table 3. Proportion of patients meeting/not meeting MoHNZ best practice recommendations Indicator Recommendation met N (%) Recommendation not met N (%) Information unavailable N (%) Indicator 1 (62 days) Indicator 2 (14 days) Indicator 3 (31 days) 9 (20%) 14 (32%) 33 (75%) 29 (66%) 25 (57%) 6 (14%) 6 (14%) 5 (11%) 5 (11%) Figure 3. Comparison of times between pipelle biopsy vs. hysteroscopy and D&C groups When timeframes between patients who received pipelle biopsy for diagnosis as opposed to hysteroscopy and D&C were compared (Figure 3), timeframes were longer in the hysteroscopy and D&C group for REF-FSA (38.5 days vs. 19.5 days) and FSA-DIAG (53 days vs. 0 days). As a result of these initial differences, the Hysteroscopy and D&C group had considerably longer timeframes for FSA-TREAT and REF-TREAT when compared to the pipelle biopsy group. DIAG-SCAN, WAIT-TREAT, and SCAN-TREAT timeframes were similar for both groups. Ten patients with longest timeframes were further selected and their clinical history was looked into order to identify factors which contributed towards delays. A systemic aspect contributing to the delay within this patient group was the waiting time for hysteroscopy and D&C, where it took between 53 days to 136 days in the five longest outliers despite the hysteroscopy priority score (80-100) indicating the urgency for the procedure (80-100 out of 100 according to Clinical Priority Assessment Criteria score system). An important patient-related factor was multiple did-not-attends (DNAs) identified in two outliers, which likely mirrored patients denial of a potential cancer diagnosis. Time taken for patient to consider and agree to a treatment plan was also another significant factor with one patients case taking 103 days for treatment to be initiated after the staging scan. The most important clinical factors were insufficient or inconclusive histological Pipelle samples for diagnosis and documented difficult examination at FSA, which were present amongst five out of ten longest outliers. Other possible factors identified from these patients were rural referrals (2 out of the 10 longest outliers), multiple medical comorbidities requiring lengthy medical review prior to operation, and low clinical suspicion on FSA referral with 9 out of 10 longest outliers having been triaged as semi - urgent (to be seen within 6-12 weeks) or even lower priority. Discussion This audit aimed to quantify the duration of time taken for diagnosis and treatment of endometrial carcinoma at Dunedin Hospital New Zealand, and to identify areas where significant delays may be present. When compared to the findings from similar audit done at Christchurch Womens Hospital in 2009, times from referral to diagnosis were considerably longer in Dunedin Hospital (93 days vs. 26.6 days), but similar in treatment decision to treatment initiation (19 days vs. 18.2 days). It is important to note however that the Christchurch audit included all gynaecological cancers, not solely endometrial carcinomas; hence it is difficult to make a direct comparison of results. Symptom duration experienced by patients before presenting to their GP or private gynaecologist was seen to vary between 1 to 24 months. It is worth keeping mind that the symptom duration is not included in the calculation timeframes. Patient related delay therefore may contribute significantly towards overall delay in diagnosis and treatment of endometrial cancer. The delays seen compared to MoHNZ recommendations could be due to factors specific to Dunedin Hospital which is part of the Southern District Health Board (DHB). The Southern DHB has a land area of over 62,356 sq km making it geographically the largest DHB region in New Zealand. 7.1% of the national population live within the Southern DHB catchment area, with 59% of the catchment population residing outside Dunedin City.8 The nature of distribution of population and the large geographic catchment area means effective health resource distribution is challenging. Although referrals from Southland Hospital were excluded in this audit, patients referred from Oamaru, Dunstan, and other rural areas were included which may have contributed towards the delays seen in some patients. Attempts were made to compare time delays between referrals from urban Dunedin and rural areas; however results were not statistically significant, likely due to low power. As shown previously in Figure two, median times were much longer in the hysteroscopy and D&C group compared to the Pipelle biopsy group in REF-FSA time (38.5 days vs. 19.5 days), FSA-DIAG time (53 days vs. 0 days), FSA-TREAT time (91 days vs. 51.5 days), and REF-TREAT time (105.5 days vs. 76 days). Understandably, the longer duration in FSA-DIAG timeframe in the hysteroscopy and D&C group is likely to have been due to patients having to wait until the procedure could be done, while pipelle biopsies were able to be taken either at GP clinics prior to FSA or on the day of the FSA, which explains the median time of 0 days in FSA-DIAG timeframe for Pipelle biopsy group. These results suggest the waiting time for hysteroscopy and D&C in a patient for histological diagnosis is a prominent factor in timely diagnosis and treatment of endometrial cancer. Measures such as supporting the approach of GPs performing pipelle biopsies by increased training of general practitioners and streamlining of referral processes for hysteroscopy and D&C in suspected endometrial carcinoma may prove to be effective in reducing the delays seen. Inability to perform pipelle biopsy due to tight cervical os was also seen to be a common barrier, with 50% of the patients who had hysteroscopy and D&C requiring it for this reason. Difficulty in insertion of the pipelle is often encountered in endometrial biopsy with one study showing that even after excluding women with cervical stenosis, difficulty in passing pipelle through the cervix was experienced in 41.7% of the patients.9 Some have suggested the use of misoprostol for cervical ripening prior to pipelle biopsy as a means to improve the success rate. However there has been only one small randomized controlled trial to date,10 which showed little benefit of 400mcg oral misoprostol prior to biopsy, and even noted an increased incidence of pain and cramping during the procedure compared to controls. A larger randomised controlled trial on the topic is hence warranted. Hysteroscopy and directed biopsy, given its superior diagnostic potential compared to other endometrial biopsy methods, has been promoted to provide precise diagnosis of intrauterine pathologies even when provided in an office setting with a narrow scope.12 The option of an assessment of patients with post-menopausal bleeding at their first FSA appointment in an office hysteroscopy setting has recently become available in Dunedin hospital, and may help to shorten the timeframe to diagnosis as ideally a hysteroscopy and curettage can a be performed at the first FSA appointment. Shortages in access to theatre and imaging for patients to yet be diagnosed with cancer are additional aspects of delayed diagnosis and treatment of cancer patients. The new best practice standards of the Ministry of Health might improve the situation for cancer patients as they are likely to receive higher priority in accessing these resources, however it does not address the actual overall shortage in these areas. Factors identified from the 10 longest outliers have also been identified by other studies within the literature. A study by Robinson et al11 investigated the association between socioeconomic factors and diagnostic delays among all gynaecological cancers, and found that rural patients were at greater risk of experiencing delays compared to urban counterparts in time taken from GP referral to first specialist appointment (odds ratio [OR]=2.20). The presence of medical comorbidities was also associated with increased risk of experiencing secondary-care related delays (OR=1.66). Interestingly, younger patient age was associated with overall increased risk of experiencing delays, and the authors have suggested a few reasons for this; the main points being younger women feeling less urgency to contact their medical practitioner with initial onset of symptoms, less suspicion of cancer from referring doctors due to the patient demographics, and the fact that younger women are more likely to be working hence may find difficulty in scheduling a time for an appointment. While factors such as inconclusive histological samples and difficult gynaecological examinations may not be modifiable, factors such as multiple DNAs may be due to the patients poor understanding of the significance of symptoms such as post-menopausal bleeding. No studies to date have specifically explored New Zealand publics perception of risk in post-menopausal bleeding. Hence, such research may be helpful in determining whether a stronger emphasis on patient education is required in order to reduce impact of patient-related factors such as multiple DNAs in diagnostic delay of endometrial carcinoma. Internationally, specific cancer timeframe recommendations such as NICE guidelines in United Kingdom have demonstrated improvement in faster cancer diagnosis since its implementation13 and some centres are now adopting multi-centred accreditation process14 to standardise care for cancer patients. The Faster Cancer Treatment programme of the Ministry of Health NZ aims to improve services by standardising care pathways and timeliness of services for cancer patients throughout New Zealand. The four main focus areas of this programme are: faster cancer treatment indicators, patient pathway coordination, tumour specific standards and multidisciplinary meetings. The planned collection of consistent data throughout the country will provide up-to-date information about the appropriateness of our current target timeframes as well as aid in its development, and hopefully improve in the context of work done in parallel within the other three main foci of work. This audit is a first assessment of time taken for referral, diagnosis and treatment of endometrial cancer in Dunedin Hospital. Although seemingly small in size, this single-centre based audit involved comparable caseload of endometrial carcinomas to the diagnostic delay study carried out by Vandborg et al5 as well as the aforementioned 2009 Christchurch Womens Hospital Audit. Given its retrospective design however, this audit is affected by bias from missing data and it is possible that some patients with endometrial cancer were lost to follow up, therefore affecting the final result. The results are specific to Dunedin Hospital, therefore is not readily generalizable to other regions in New Zealand, however it is likely that the problem of delay in diagnosis and treatment does not involve Dunedin Hospital alone. This audit specifically looked at patients diagnosed with endometrial carcinoma; hence patients who may have been investigated for PMB but were subsequently found not to have endometrial cancer were not included. Therefore, further research looking into timeframes for patients presenting with post-menopausal bleeding may provide more robust information regarding our current PMB diagnostic pathway performance. Conclusion The results of this audit show that the best practice times as proposed by MoHNZ for timely cancer diagnosis and treatment have not been met for patients with endometrial cancer, specifically with regards to time from initial referral to treatment, and to a certain extent from initial referral to first specialist appointment. Once the diagnosis was made, treatment was usually within the appropriate timeframe. The main factor contributing to the delays which could be identified was waiting times for hysteroscopy and D&C for histological diagnosis. Better education of patients about post-menopausal bleeding, training of general practitioners in pipelle biopsy, expedited hysteroscopy and D&C in an office gynaecology setting, and reduced waiting times for imaging and surgery could be possible ways to address the delays.