No abstract

Background: Lung cancer management is characterised by a high disease burden, poor survival and substantial variation in management and outcomes. Service redesign provides opportunities for quality improvement (QI) and this improvement may be leveraged across multiple sites using QI collaboration.Aim: This initiative targeted Quality Improvement (QI) in lung cancer management, engaging a QI collaborative using service redesign methodologies in five Victorian hospitals. QI targets included timeliness from referral and diagnosis to treatment, multi-disciplinary meeting (MDM) presentation and supportive care screening. Redesign strategies targeted process sustainability through enhanced team capability.Methods: This study engaged a prospective quality improvement cohort design targeting newly diagnosed tissue confirmed lung cancer with 6-month pre-intervention period and 6-month redesign implementation period, between September 2016 and August 2017, evaluated using Interrupted Time Series Analysis. Hospital sites included three regional and two metropolitan hospitals in Victoria. QI redesign targeted time intervals from referral to first specialist appointment (FSA), referral to diagnosis, diagnosis to first treatment (any intent), MDM documented in medical records and Supportive Care Screening Tool documented in medical records.Results: There was a marked reduction in referral to FSA interval across all sites, with median (interquartile range) falling from 6 (0-15) to 4 (1-10) days, and proportion seen by a specialist within 14 days increased from 74.3% to 84.2%. The interval between diagnosis and treatment was not substantively changed in the 6-month implementation period. The proportion of subjects with documented presentation to the MDM increased from 61% to 67%. The proportion for which Supportive Care Screening documentation remained low at 26.3% post-intervention.Conclusions: Data-driven redesign initiatives enable identification and analysis of clinical practice variation and may be utilised to enhance timeliness of cancer care and improve local data service capabilities.

No abstract

This article presents an estimation and analysis of the Hurst exponent for Australian stocks using the wavelet technique. Consistent with Mulligan's (2004) study of US technology stocks, we find that the Hurst exponent varies over the cross-section of stocks. We also analyse Mulligan's (2004) and our data and find that beta can explain some of the cross-sectional variation in the Hurst exponents. However, we find that our results are not robust to filtering out the short range dependence in the data.