Natural language processing to identify ureteric stones in radiology reports.

INTRODUCTION: Natural language processing (NLP) is an emerging tool which has the ability to automate data extraction from large volumes of unstructured text. One of the main described uses of NLP in radiology is cohort building for epidemiological studies. This study aims to assess the accuracy of NLP in identifying a group of patients positive for ureteric stones on Computed Tomography - Kidneys, Ureter, Bladder (CT KUB) reports. METHODS: Retrospective review of all CT KUB reports in a single calendar year. A locally available NLP tool was used to automatically classify the reports based on positivity for ureteric stones. This was validated by manual review and refined to maximize the accuracy of stone detection. RESULTS: A total of 1874 CT KUB reports were identified. Manual classification of ureteric stone positivity was 36% compared with 27% using NLP. The accuracy of NLP was 85% with a sensitivity of 66% and specificity of 95%. Incorrect classification was due to misspellings, variable syntax, terminology, pluralization and the inability to exclude clinical request details from the search algorithm. CONCLUSIONS: Our NLP tool demonstrated high specificity but low sensitivity at identifying CT KUB reports that are positive for ureteric stones. This was attributable to the lack of feature extraction tools tailored for analysing radiology text, incompleteness of the medical lexicon database and heterogeneity of unstructured reports. Improvements in these areas will help improve data extraction accuracy.

A multifaceted programme to reduce the rate of tongue-tie release surgery in newborn infants: Observational study.

OBJECTIVES: A programme was introduced in Canterbury, New Zealand to evaluate the diagnosis and treatment of frenulum releases in newborn infants with suspected tongue-tie (ankyloglossia). The primary goals were to support breastfeeding and ensure that unnecessary surgery was avoided. METHODS: Local healthcare professionals reached consensus on a pathway for improving management of infants with tongue-tie and breast-feeding difficulties. This embedded an expert breast-feeding review and assessment of lingual function using a validated method, the Bristol Tongue-tie Assessment Tool (BTAT). Infants with breastfeeding problems related to tongue-tie had a frenotomy at a hospital outpatient clinic. An education programme was developed to support introduction of the new clinical pathway and included seminars and online information for healthcare professionals and the general public. RESULTS: Frenotomy intervention rate reduced markedly from 11.3% in 2015 to 3.5% by mid-2017. Feeding methods were not different before or after surgery between infants who received a frenotomy and those who did not. Initially, the BTAT threshold for frenotomy was set at ≤5, however the final clinical pathway combined a breastfeeding assessment and a BTAT threshold of ≤4. The education programs assisted with the changes in practice, while increased use of the clinician guidance and public health information websites confirmed growing awareness of tongue-tie and community breastfeeding support. CONCLUSIONS: Establishing consistent multidisciplinary assessment of tongue-tie in infants with feeding difficulties led to a marked reduction in frenotomy intervention rate. 23% of the frenotomy group in the 2016 audit showed a significant improvement in the ability to breastfeed, but overall there was no difference in the feeding pattern of infants who either received or were declined a frenotomy. The development of a supportive education programme and availability of online information about tongue-tie for health professionals and consumers contributed to successful uptake of the new clinical pathway.

Unmet need for referred services as measured by general practice.

INTRODUCTION Unmet needs are a key indicator of the success of a health system. Clinicians and funders in Christchurch, Canterbury, New Zealand were concerned that unmet health need was hidden. AIM The aim of this survey was to estimate the proportion of patients attending general practice who were unable to access clinically indicated referred services. METHODS The survey used a novel method to estimate unserviced health needs. General practitioners (GPs, n = 54) asked their patients (n = 2135) during a consultation about any health needs requiring a referred service. If both agreed that a service was potentially beneficial and not available, this was documented on an e-referral system for review. The outcomes of actual referrals were also reviewed. RESULTS The patient group was broadly representative of the Canterbury population, but over-sampled female and middle-aged people and under-sampled Maori. Data adjusted to regional demographics showed that 3.6% of patients had a GP-confirmed unserviced health need. Elective orthopaedic surgery, general surgery and mental health were areas of greatest need. Unserviced health needs were significantly (P ≤ 0.05) associated with greater deprivation, middle-age, and receiving high health-use subsidies. DISCUSSION To our knowledge, this is the first survey of GP and patient agreement on unserviced referred health needs. Measuring unserviced health needs in this way is directly relevant to service planning because the gaps identified reflect clinically indicated services that patients want and need. The survey method is an improvement on declined referral rates as a measure of need. Key factors in the method were using a patient-initiated GP consultation and an e-referral system to collect data.

Does the addition of saline infusion sonohysterography to transvaginal ultrasonography prevent unnecessary hysteroscopy in premenopausal women with abnormal uterine bleeding?

This observational case series in 65 premenopausal women with abnormal uterine bleeding evaluated whether transvaginal ultrasound followed by saline infusion sonohysterography (SIS) prevented unnecessary hysteroscopy. Although SIS indicated that hysteroscopy was unnecessary in eight women, this benefit was offset by the invasive nature of the scan, the number of endometrial abnormalities falsely detected by SIS and the cost of the additional investigation.