TY - JOUR
T1 - Preventing critical failure. Can routinely collected data be repurposed to predict avoidable patient harm? A quantitative descriptive study
AU - Nowotny, Benjamin Michael
AU - Davies-Tuck, Miranda
AU - Scott, Belinda
AU - Stewart, Michael
AU - Cox, Elizabeth
AU - Cusack, Karen
AU - Fletcher, Martin
AU - Saar, Eva
AU - Farrell, Tanya
AU - Anil, Shirin
AU - McKinlay, Louise
AU - Wallace, Euan M.
PY - 2021/3
Y1 - 2021/3
N2 - Objectives: To determine whether sharing of routinely collected health service performance data could have predicted a critical safety failure at an Australian maternity service. Design: Observational quantitative descriptive study. Setting: A public hospital maternity service in Victoria, Australia. Data sources: A public health service; the Victorian state health quality and safety office-Safer Care Victoria; the Health Complaints Commission; Victorian Managed Insurance Authority; Consultative Council on Obstetric and Paediatric Mortality and Morbidity; Paediatric Infant Perinatal Emergency Retrieval; Australian Health Practitioner Regulation Agency. Main outcome measures: Numbers and rates for events (activity, deaths, complaints, litigation, practitioner notifications). Correlation coefficients. Results: Between 2000 and 2014 annual birth numbers at the index hospital more than doubled with no change in bed capacity, to be significantly busier than similar services as determined using an independent samples t-test (p<0.001). There were 36 newborn deaths, 11 of which were considered avoidable. Pearson correlations revealed a weak but significant relationship between number of births per birth suite room birth and perinatal mortality (r2=0.18, p=0.003). Independent samples t-tests demonstrated that the rates of emergency neonatal and perinatal transfer were both significantly lower than similar services (both p<0.001). Direct-to-service patient complaints increased ahead of recognised excess perinatal mortality. Conclusion: While clinical activity data and direct-to-service patient complaints appear to offer promise as potential predictors of health service stress, complaints to regulators and medicolegal activity are less promising as predictors of system failure. Significant changes to how all data are handled would be required to progress such an approach to predicting health service failure.
AB - Objectives: To determine whether sharing of routinely collected health service performance data could have predicted a critical safety failure at an Australian maternity service. Design: Observational quantitative descriptive study. Setting: A public hospital maternity service in Victoria, Australia. Data sources: A public health service; the Victorian state health quality and safety office-Safer Care Victoria; the Health Complaints Commission; Victorian Managed Insurance Authority; Consultative Council on Obstetric and Paediatric Mortality and Morbidity; Paediatric Infant Perinatal Emergency Retrieval; Australian Health Practitioner Regulation Agency. Main outcome measures: Numbers and rates for events (activity, deaths, complaints, litigation, practitioner notifications). Correlation coefficients. Results: Between 2000 and 2014 annual birth numbers at the index hospital more than doubled with no change in bed capacity, to be significantly busier than similar services as determined using an independent samples t-test (p<0.001). There were 36 newborn deaths, 11 of which were considered avoidable. Pearson correlations revealed a weak but significant relationship between number of births per birth suite room birth and perinatal mortality (r2=0.18, p=0.003). Independent samples t-tests demonstrated that the rates of emergency neonatal and perinatal transfer were both significantly lower than similar services (both p<0.001). Direct-to-service patient complaints increased ahead of recognised excess perinatal mortality. Conclusion: While clinical activity data and direct-to-service patient complaints appear to offer promise as potential predictors of health service stress, complaints to regulators and medicolegal activity are less promising as predictors of system failure. Significant changes to how all data are handled would be required to progress such an approach to predicting health service failure.
KW - governance
KW - healthcare quality improvement
KW - incident reporting
KW - obstetrics and gynecology
KW - patient safety
UR - http://www.scopus.com/inward/record.url?scp=85077751054&partnerID=8YFLogxK
U2 - 10.1136/bmjqs-2019-010141
DO - 10.1136/bmjqs-2019-010141
M3 - Article
C2 - 31915180
AN - SCOPUS:85077751054
SN - 2044-5415
VL - 30
SP - 186
EP - 194
JO - BMJ Quality & Safety
JF - BMJ Quality & Safety
IS - 3
ER -