Need for shift in focus in research into quality of intensive care

• Published in: The Lancet (British edition) 2002-01, Vol.359 (9301), p.95-96
• Main Authors: Pollack, Murray M, Patel, Kantilal M
• Format: Article
• Language: English
• Subjects: Birth weight; Clinical outcomes; Critical care; Critical Care - standards; Health Services Research - organization & administration; Hospital Mortality; Humans; Intensive care; Intensive Care Units - standards; Medical research; Mortality; Mortality risk; Needs Assessment - organization & administration; Nosocomial infections; Nursing; Outcome Assessment (Health Care); Personnel Staffing and Scheduling - standards; Quality; Quality of Health Care; Research Design - standards; United Kingdom - epidemiology; Workforce planning; Workload; Workloads
• ISSN: 0140-6736; 1474-547X
• DOI: 10.1016/S0140-6736(02)07334-8

Third, the definition of variables may have biased the study against finding significant associations with volume and with staffing variables. Categorisation of volume into high, medium, and low such that each group accounted for a third of all very-low-birthweight admissions limited the ability to detect a minimum threshold effect of volume. Previous studies have indicated that under some circumstances once a minimum volume threshold is met there is no additional advantage of increased experience. This point may be most important when there is a very specialised and technical component to delivery of care. For instance, this threshold effect is found in paediatric cardiovascular surgery and paediatric intensive care.2,3 The care of infants of extremely or very low birthweights, who make up the highest-risk neonatal groups, is sufficiently specialised to be prone to a similar effect. Even drugs may have different effects in individual patients depending on patient-related factors, such as severity of illness.4 It would not be surprising that a threshold effect between volume and mortality exists for neonatal intensive care and was undetected because of the definition of high, medium, and low volume units. Finally, the definition of high and low availability of consultant or nurse may have limited the generalisability of the results outside of the UK. Importantly, the number of staff may not be an adequate proxy for the important issue of how much staff time is spent at the bedside delivering care. The suggestion in this study that measured nursing workload is related to risk-adjusted mortality supports the clear need to relate both nursing and physician staffing to actual bedside care.