Survival from breast cancer: an analysis of Australian data by surgeon case load, treatment centre location, and health insurance status

Multidimensional Machine Learning Personalized Prognostic Model in an Early Invasive Breast Cancer Population-Based Cohort in China: Algorithm Validation Study

Background

Current online prognostic prediction models for breast cancer, such as Adjuvant! Online and PREDICT, are based on specific populations. They have been well validated and widely used in the United States and Western Europe; however, several validation attempts in non-European countries have revealed suboptimal predictions.

Objective

We aimed to develop an advanced breast cancer prognosis model for disease progression, cancer-specific mortality, and all-cause mortality by integrating tumor, demographic, and treatment characteristics from a large breast cancer cohort in China.

Methods

This study was approved by the Clinical Test and Biomedical Ethics Committee of West China Hospital, Sichuan University on May 17, 2012. Data collection for this project was started in May 2017 and ended in March 2019. Data on 5293 women diagnosed with stage I to III invasive breast cancer between 2000 and 2013 were collected. Disease progression, cancer-specific mortality, all-cause mortality, and the likelihood of disease progression or death within a 5-year period were predicted. Extreme gradient boosting was used to develop the prediction model. Model performance was assessed by calculating the area under the receiver operating characteristic curve (AUROC), and the model was calibrated and compared with PREDICT.

Results

The training, test, and validation sets comprised 3276 (499 progressions, 202 breast cancer-specific deaths, and 261 all-cause deaths within 5-year follow-up), 1405 (211 progressions, 94 breast cancer-specific deaths, and 129 all-cause deaths), and 612 (109 progressions, 33 breast cancer-specific deaths, and 37 all-cause deaths) women, respectively. The AUROC values for disease progression, cancer-specific mortality, and all-cause mortality were 0.76, 0.88, and 0.82 for training set; 0.79, 0.80, and 0.83 for the test set; and 0.79, 0.84, and 0.88 for the validation set, respectively. Calibration analysis demonstrated good agreement between predicted and observed events within 5 years. Comparable AUROC and calibration results were confirmed in different age, residence status, and receptor status subgroups. Compared with PREDICT, our model showed similar AUROC and improved calibration values.

Conclusions

Our prognostic model exhibits high discrimination and good calibration. It may facilitate prognosis prediction and clinical decision making for patients with breast cancer in China.

Public health insurance and cancer-specific mortality risk among patients with breast cancer: A prospective cohort study in China

Little is known about how health insurance policies, particularly in developing countries, influence breast cancer prognosis. Here, we examined the association between individual health insurance and breast cancer-specific mortality in China. We included 7436 women diagnosed with invasive breast cancer between 2009 and 2016, at West China Hospital, Sichuan University. The health insurance plan of patient was classified as either urban or rural schemes and was also categorized as reimbursement rate (ie, the covered/total charge) below or above the median. Breast cancer-specific mortality was the primary outcome. Using Cox proportional hazards models, we calculated hazard ratios (HRs) for cancer-specific mortality, contrasting rates among patients with a rural insurance scheme or low reimbursement rate to that of those with an urban insurance scheme or high reimbursement rate, respectively. During a median follow-up of 3.1 years, we identified 326 deaths due to breast cancer. Compared to patients covered by urban insurance schemes, patients covered by rural insurance schemes had a 29% increased cancer-specific mortality (95% CI 0%-65%) after adjusting for demographics, tumor characteristics and treatment modes. Reimbursement rate below the median was associated with a 42% increased rate of cancer-specific mortality (95% CI 11%-82%). Every 10% increase in the reimbursement rate is associated with a 7% (95% CI 2%-12%) reduction in cancer-specific mortality risk, particularly in patients covered by rural insurance schemes (26%, 95% CI 9%-39%). Our findings suggest that underinsured patients face a higher risk of breast cancer-specific mortality in developing countries.

Measuring the Impact of Multidisciplinary Care on Quality for Pancreatic Surgery: Transition to a Focused, Very High-volume Program

Outcomes are superior for pancreatic resection at high-volume centers. To assess the impact of focused high-volume multidisciplinary care, a quality index (QI) was used to review our experience. Data from 1996 to July 2012 were analyzed in three groups: 1) early (1996 to 2007); 2) transition (2008 to 2009); and 3) mature (2010 to July 2012). A total of 239 patients were included with a mean age of 63.4 years and the majority were white (65.7%). The number of patients with Charlson comorbidity index greater than 2 and age older than 80 years increased comparing Group 1 with latter groups. Volume increased over time: Group 1 (n = 93) 7.75/year, Group 2 (n = 51) 25.5/year, and Group 3 (n = 95) 39/year. Overall mortality was 5.9 per cent: Group 1, 4.3 per cent; Group 2, 11.5 per cent; and Group 3, 3.9 per cent ( P = 0.0454). The QI score incorporates documentation, chemotherapy, resection for Stage I/II, time to treatment, margins, lymph nodes, mortality, and surgical volume with a maximum possible score 10. The QI increased over time: 3 in Group 1; 4 in Group 2; and 6 in Group 3. An improvement was noted for the quality indicators: surgical resection ( P = 0.0125) and use of palliative and adjuvant therapy ( P = 0.0144 and < 0.0001). Implementation of a focused multidisciplinary pancreatic surgery program increases quality.

Factors influencing breast cancer outcomes in Australia: A systematic review

PURPOSE

This systematic review evaluates factors influencing breast cancer outcomes for women treated in Australia, facilitating the exploration of disparities in breast cancer outcomes for certain groups of women in Australia.

METHOD

A systematic literature search was performed using MEDLINE and Scopus focusing on breast cancer in Australia with outcome measures being breast cancer survival and recurrence with no restrictions on date. Risk of bias was assessed using Cairns Assessment Scale for Observational studies of Risk factors (CASOR).

RESULTS

Fifteen quantitative studies were included: two were high quality, 11 were intermediate quality, and two were low quality. Traditional risk factors such as invasive tumour type, larger size, higher grade and stage, lymph node involvement and absence of hormone receptors were found to be associated with breast cancer mortality. Being younger (<40 years old) and older (>70 years old), having more comorbidities, being of lower socioeconomic status, identifying as Aboriginal or Torres Strait Islander, living in more rural areas or having a mastectomy were factors found to be associated with poorer breast cancer outcomes.

CONCLUSION

Despite the heterogeneity of the studies, this review identified significant risk factors for breast cancer mortality and recurrence. The use of this data would be most useful in developing evidence-based interventions and in optimising patient care through creation of a prediction model.

PROSPERO REGISTRATION

CRD42017072857.

Hospital effect on infections after four major surgical procedures: outlier and volume-outcome analysis using all-inclusive state data

BACKGROUND

Hospital volume is known to have a direct impact on the outcomes of major surgical procedures. However, it is unclear if the evidence applies specifically to surgical site infections.

AIMS

To determine if there are procedure-specific hospital outliers [with higher surgical site infection rates (SSIRs)] for four major surgical procedures, and to examine if hospital volume is associated with SSIRs in the context of outlier performance in New South Wales (NSW), Australia.

METHODS

Adults who underwent one of four surgical procedures (colorectal, joint replacement, spinal and cardiac procedures) at a NSW healthcare facility between 2002 and 2013 were included. The hospital volume for each of the four surgical procedures was categorized into tertiles (low, medium and high). Multi-variable logistic regression models were built to estimate the expected SSIR for each procedure. The expected SSIRs were used to compute indirect standardized SSIRs which were then plotted in funnel plots to identify hospital outliers.

FINDINGS

One hospital was identified to be an overall outlier (higher SSIRs for three of the four procedures performed in its facilities), whereas two hospitals were outliers for one specific procedure throughout the entire study period. Low-volume facilities performed the best for colorectal surgery and worst for joint replacement and cardiac surgery. One high-volume facility was an outlier for spinal surgery.

CONCLUSIONS

Surgical site infections seem to be mainly a procedure-specific, as opposed to a hospital-specific, phenomenon in NSW. The association between hospital volume and SSIRs differs for different surgical procedures.

Differences in Breast Cancer Survival between Public and Private Care in New Zealand: Which Factors Contribute?

Background

Patients who received private health care appear to have better survival from breast cancer compared to those who received public care. This study investigated if this applied to New Zealand women and identified factors that could explain such disparities.

Methods

This study involved all women who were diagnosed with primary breast cancer in two health regions in New Zealand, covering about 40% of the national population, between June 2000 and May 2013. Patients who received public care for primary treatment, mostly surgical treatment, were compared with those who received private care in terms of demographics, mode of presentation, disease factors, comorbidity index and treatment factors. Cox regression modelling was performed with stepwise adjustments, and hazards of breast cancer specific mortality associated with the type of health care received was assessed.

Results

Of the 14,468 patients, 8,916 (61.6%) received public care. Compared to patients treated in private care facilities, they were older, more likely to be Māori, Pacifika or Asian and to reside in deprived neighbourhoods and rural areas, and less likely to be diagnosed with early staged cancer and to receive timely cancer treatments. They had a higher risk of mortality from breast cancer (hazard ratio: 1.95; 95% CI: 1.75, 2.17), of which 80% (95% CI: 63%, 100%) was explained by baseline differences, particularly related to ethnicity, stage at diagnosis and type of loco-regional therapy. After controlling for these demographic, disease and treatment factors, the risk of mortality was still 14% higher in the public sector patients.

Conclusions

Ethnicity, stage at diagnosis and type of loco-regional therapy were the three key contributors to survival disparities between patients treated in public and private health care facilities in New Zealand. The findings underscore the need for more efforts to improve the quality, timeliness and equitability of public cancer care services.

Predictors for Reconstruction and Mood Disorder Associated With Reconstruction in Patients With Breast Cancer and Mastectomy: A Retrospective Cohort Study

This study used Taiwan's National Health Insurance medical claims to investigate the predictors for operative modes chosen by early-stage breast cancer patients; as well as to assess whether operative modes are associated with risk of mood disorder. We included 36,377 patients with breast cancer who received surgery between 2000 and 2008, and were followed to the end of 2010; they were further classified into 3 groups: mastectomy alone (n = 34,900), along with early reconstruction (n = 1080), and along with delayed reconstruction (n = 397). The results showed that age, insurance premium, urbanization level, and postsurgery chemotherapy and radiotherapy were all significant predictors for the selection of operative modes. Breast cancer patients with mastectomy alone, early reconstruction, and delayed reconstruction showed a cumulative incidence rate of mood disorder of 36.90%, 41.56%, and 33.89%, respectively. The multiple cox proportional model further revealed that early (hazard ratio [HR] = 1.06, 95% confidence interval (CI) = 0.93-1.21) and delayed (HR = 1.17, 95% CI = 0.96-1.42) reconstruction were associated with a slightly higher but insignificant risk of mood disorder, as compared to the patients received no reconstruction.

Analysing risk factors for poorer breast cancer outcomes in residents of lower socioeconomic areas of Australia

OBJECTIVE

To investigate patient, cancer and treatment factors associated with the residence of female breast cancer patients in lower socioeconomic areas of Australia to better understand factors that may contribute to their poorer cancer outcomes.

METHODS

Bivariable and multivariable analyses were performed using the Breast Quality Audit database of Breast Surgeons of Australia and New Zealand. RESULTS Multivariable regression indicated that patients from lower socioeconomic areas are more likely to live in more remote areas and to be treated at regional than major city centres. Although they appeared equally likely to be referred to surgeons from BreastScreen services as patients from higher socioeconomic areas, they were less likely to be referred as asymptomatic cases from other sources. In general, their cancer and treatment characteristics did not differ from those of women from higher socioeconomic areas, but ovarian ablation therapy was less common for these patients and bilateral synchronous lesions tended to be less frequent than for women from higher socioeconomic areas.

CONCLUSIONS

The results indicate that patients from lower socioeconomic areas are more likely to live in more remote districts and have their treatment in regional rather than major treatment centres. Their cancer and treatment characteristics appear to be similar to those of women from higher socioeconomic areas, although they are less likely to have ovarian ablation or to be referred as asymptomatic patients from sources other than BreastScreen. What is known about this topic? It is already known from Australian data that breast cancer outcomes are not as favourable for women from areas of socioeconomic disadvantage. The reasons for the poorer outcomes have not been understood. Studies in other countries have also found poorer outcomes in women from lower socioeconomic areas, and in some instances, have attributed this finding to more advanced stages of cancers at diagnosis and more limited treatment. The reasons are likely to vary with the country and health system characteristics. What does this paper add? The present study found that in Australia, women from lower socioeconomic areas do not have more advanced cancers at diagnosis, nor, in general, other cancer features that would predispose them to poorer outcomes. The standout differences were that they tended more to live in areas that were more remote from specialist metropolitan centres and were more likely to be treated in regional settings where prior research has indicated poorer outcomes. The reasons for these poorer outcomes are not known but may include lower levels of surgical specialisation, less access to specialised adjunctive services, and less involvement with multidisciplinary teams. Women from lower socioeconomic areas also appeared more likely to attend lower case load surgeons. Little difference was evident in the type of clinical care received, although women from lower socioeconomic areas were less likely to be asymptomatic referrals from other clinical settings (excluding BreastScreen). What are the implications for practitioners? Results suggest that poorer outcomes in women from lower socioeconomic areas in Australia may have less to do with the characteristics of their breast cancers or treatment modalities and more to do with health system features, such as access to specialist centres. This study highlights the importance of demographic and health system features as potentially key factors in service outcomes. Health system research should be strengthened in Australia to augment biomedical and clinical research, with a view to best meeting service needs of all sectors of the population.

Impact of geographic location on surgical outcomes of women with breast cancer

BACKGROUND

The aim of this study is to use the Breast Surgeons of Australia and New Zealand Quality Audit to evaluate the diagnostic work up and surgical treatment path for Australian women with early breast cancer in relation to geographic location.

METHODS

We accessed data on patient geographic demographics, diagnostic and surgical procedures and cancer characteristics in 115,872 episodes of early breast cancer reported to the Quality Audit between 1998 and 2012. Tumour size, grade and number, lymph node positivity and lymphovascular invasion are the major prognostic factors adjusted for.

RESULTS

Core biopsy is the most frequently performed diagnostic method and open biopsy the least (P=0.001). Remote locations have the highest proportion of open biopsies and cities the lowest (13.8 versus 9.8%, P=0.001). The outer regional/remote patient group has a higher likelihood of an open biopsy than city patients (odds ratio 1.5; 95% confidence interval 1.1-2.2 P=0.02). Reconstruction rates are low throughout ranging from 3 to 4% (P=0.001). Cities have the highest proportion of no surgery (1.5%, P=0.001). Cities have the highest proportion of breast conserving surgery and the remote the lowest (64.5% versus 29.9%, P=0.001). Outer regional locations have the highest proportion of mastectomies and cities the lowest (38 versus 29.9%, P=0.001). The likelihood of receiving radiotherapy is associated with geographic location of residence and geographic distribution of the radiotherapy facility (P=0.002, P=0.001).

CONCLUSION

Geographic location of residence is associated with the diagnostic work up and surgical treatment of Australian women with early breast cancer.