Extracting cancer mortality statistics from death certificates: A hybrid machine learning and rule-based approach for common and rare cancers.

OBJECTIVE: Death certificates are an invaluable source of cancer mortality statistics. However, this value can only be realised if accurate, quantitative data can be extracted from certificates-an aim hampered by both the volume and variable quality of certificates written in natural language. This paper proposes an automatic classification system for identifying all cancer related causes of death from death certificates. METHODS: Detailed features, including terms, n-grams and SNOMED CT concepts were extracted from a collection of 447,336 death certificates. The features were used as input to two different classification sub-systems: a machine learning sub-system using Support Vector Machines (SVMs) and a rule-based sub-system. A fusion sub-system then combines the results from SVMs and rules into a single final classification. A held-out test set was used to evaluate the effectiveness of the classifiers according to precision, recall and F-measure. RESULTS: The system was highly effective at determining the type of cancers for both common cancers (F-measure of 0.85) and rare cancers (F-measure of 0.7). In general, rules performed superior to SVMs; however, the fusion method that combined the two was the most effective. CONCLUSION: The system proposed in this study provides automatic identification and characterisation of cancers from large collections of free-text death certificates. This allows organisations such as Cancer Registries to monitor and report on cancer mortality in a timely and accurate manner. In addition, the methods and findings are generally applicable beyond cancer classification and to other sources of medical text besides death certificates.

Automatic ICD-10 classification of cancers from free-text death certificates.

OBJECTIVE: Death certificates provide an invaluable source for cancer mortality statistics; however, this value can only be realised if accurate, quantitative data can be extracted from certificates--an aim hampered by both the volume and variable nature of certificates written in natural language. This paper proposes an automatic classification system for identifying cancer related causes of death from death certificates. METHODS: Detailed features, including terms, n-grams and SNOMED CT concepts were extracted from a collection of 447,336 death certificates. These features were used to train Support Vector Machine classifiers (one classifier for each cancer type). The classifiers were deployed in a cascaded architecture: the first level identified the presence of cancer (i.e., binary cancer/nocancer) and the second level identified the type of cancer (according to the ICD-10 classification system). A held-out test set was used to evaluate the effectiveness of the classifiers according to precision, recall and F-measure. In addition, detailed feature analysis was performed to reveal the characteristics of a successful cancer classification model. RESULTS: The system was highly effective at identifying cancer as the underlying cause of death (F-measure 0.94). The system was also effective at determining the type of cancer for common cancers (F-measure 0.7). Rare cancers, for which there was little training data, were difficult to classify accurately (F-measure 0.12). Factors influencing performance were the amount of training data and certain ambiguous cancers (e.g., those in the stomach region). The feature analysis revealed a combination of features were important for cancer type classification, with SNOMED CT concept and oncology specific morphology features proving the most valuable. CONCLUSION: The system proposed in this study provides automatic identification and characterisation of cancers from large collections of free-text death certificates. This allows organisations such as Cancer Registries to monitor and report on cancer mortality in a timely and accurate manner. In addition, the methods and findings are generally applicable beyond cancer classification and to other sources of medical text besides death certificates.

Classification of cancer-related death certificates using machine learning.

BACKGROUND: Cancer monitoring and prevention relies on the critical aspect of timely notification of cancer cases. However, the abstraction and classification of cancer from the free-text of pathology reports and other relevant documents, such as death certificates, exist as complex and time-consuming activities. AIMS: In this paper, approaches for the automatic detection of notifiable cancer cases as the cause of death from free-text death certificates supplied to Cancer Registries are investigated. METHOD: A number of machine learning classifiers were studied. Features were extracted using natural language techniques and the Medtex toolkit. The numerous features encompassed stemmed words, bi-grams, and concepts from the SNOMED CT medical terminology. The baseline consisted of a keyword spotter using keywords extracted from the long description of ICD-10 cancer related codes. RESULTS: Death certificates with notifiable cancer listed as the cause of death can be effectively identified with the methods studied in this paper. A Support Vector Machine (SVM) classifier achieved best performance with an overall Fmeasure of 0.9866 when evaluated on a set of 5,000 freetext death certificates using the token stem feature set. The SNOMED CT concept plus token stem feature set reached the lowest variance (0.0032) and false negative rate (0.0297) while achieving an F-measure of 0.9864. The SVM classifier accounts for the first 18 of the top 40 evaluated runs, and entails the most robust classifier with a variance of 0.001141, half the variance of the other classifiers. CONCLUSION: The selection of features significantly produced the most influences on the performance of the classifiers, although the type of classifier employed also affects performance. In contrast, the feature weighting schema created a negligible effect on performance. Specifically, it is found that stemmed tokens with or without SNOMED CT concepts create the most effective feature when combined with an SVM classifier.

The impact of OCR accuracy on automated cancer classification of pathology reports.

OBJECTIVE: To evaluate the effects of Optical Character Recognition (OCR) on the automatic cancer classification of pathology reports. METHOD: Scanned images of pathology reports were converted to electronic free-text using a commercial OCR system. A state-of-the-art cancer classification system, the Medical Text Extraction (MEDTEX) system, was used to automatically classify the OCR reports. Classifications produced by MEDTEX on the OCR versions of the reports were compared with the classification from a human amended version of the OCR reports. RESULTS: The employed OCR system was found to recognise scanned pathology reports with up to 99.12% character accuracy and up to 98.95% word accuracy. Errors in the OCR processing were found to minimally impact on the automatic classification of scanned pathology reports into notifiable groups. However, the impact of OCR errors is not negligible when considering the extraction of cancer notification items, such as primary site, histological type, etc. CONCLUSIONS: The automatic cancer classification system used in this work, MEDTEX, has proven to be robust to errors produced by the acquisition of freetext pathology reports from scanned images through OCR software. However, issues emerge when considering the extraction of cancer notification items.

Babies born after ART treatment cost more than non-ART babies: a cost analysis of inpatient birth-admission costs of singleton and multiple gestation pregnancies.

BACKGROUND: Currently, about one-third of infants born after assisted reproductive technology (ART) worldwide are twins or triplets. This study compared the inpatient birth-admission costs of singleton and multiple gestation ART deliveries to non-ART deliveries. METHODS: A cohort of 5005 mothers and 5886 infants conceived following ART treatment were compared to 245 249 mothers and 248 539 infants in the general population. Birth-admission costs were calculated using Australian Refined Diagnosis Related Groups and weighted national average costs (2003-2004 euro). RESULTS: ART infants were 4.4 times more likely to be low birthweight (LBW) compared with non-ART infants, translating into 89% higher birth-admission costs (euro2,832 and euro1,502, respectively). ART singletons were also more likely to be LBW compared with non-ART singletons, translating into 31% higher birth-admission costs (euro1,849 and euro1,415, respectively). After combining infant and maternal admission costs, the average cost of an ART singleton delivery was euro4,818 compared with euro13 890 for ART twins and euro54 294 for ART higher order multiples. Findings were not sensitive to changes in casemix. CONCLUSIONS: The poorer neonatal outcomes of ART singletons compared with non-ART singletons are significant enough to impact healthcare resource consumption. The high costs associated with ART multiple births add to the overwhelming clinical and economic evidence in support of single embryo transfer.

The urban-remote divide for Indigenous perinatal outcomes.

OBJECTIVE: To determine whether remoteness category of residence of Indigenous women affects the perinatal outcomes of their newborn infants. DESIGN AND PARTICIPANTS: A population-based study of 35 240 mothers identified as Indigenous and their 35 658 babies included in the National Perinatal Data Collection in 2001-2004. MAIN OUTCOME MEASURES: Australian Standard Geographical Classification remoteness category, birthweight, Apgar score at 5 minutes, stillbirth, gestational age and a constructed measure of perinatal outcomes of babies called "healthy baby" (live birth, singleton, 37-41 completed weeks' gestation, 2500-4499 g birthweight, and an Apgar score at 5 minutes >or= 7). RESULTS: The proportion of healthy babies in remote, regional and city areas was 74.9%, 77.7% and 77.6%, respectively. After adjusting for age, parity, smoking and diabetes or hypertension, babies born to mothers in remote areas were less likely to satisfy the study criteria of being a healthy baby (adjusted odds ratio [AOR], 0.87; 95% CI, 0.81-0.93) compared with those born in cities. Babies born to mothers living in remote areas had higher odds of being of low birthweight (AOR, 1.09; 95% CI, 1.01-1.19) and being born with an Apgar score < 7 at 5 minutes (AOR, 1.63; 95% CI, 1.39-1.92). CONCLUSIONS: Only three in four babies born to Indigenous mothers fell into the "healthy baby" category, and those born in more remote areas were particularly disadvantaged. These findings demonstrate the continuing need for urgent and concerted action to address the persistent perinatal inequity in the Indigenous population.

Trends in hospital service provision.

In this paper, trends in hospital service provision are measured using data on the numbers and nature of hospitals, on hospital expenditure and on hospital activity over recent years. The number of public acute care hospitals was fairly stable, however, bed numbers decreased. Hospital numbers rose for private hospitals, as did numbers of beds, particularly for group for-profit private hospitals. Recurrent health expenditure on hospitals as a proportion of all recurrent health expenditure fell, although it rose for private hospitals, and real increases in expenditure occurred for both public acute and private hospitals. Population rates for separations and patient days rose for private hospitals and were stable and fell, respectively, for public acute hospitals. Average length of stay decreased for both public acute and private hospitals, with increasing numbers of separations occurring on a same day basis. Increasing proportions of procedures were undertaken during same day stays, and in private hospitals. Separation rates varied geographically, with highest rates overall, and for public hospitals and overnight separations, for patients resident in remote centres and other remote areas. Highest rates for private hospitals were for patients resident in capital cities, other metropolitan centres and large rural centres.