Discover Portal

patient being looked after following surgery

NIHR Signal Better care of deteriorating patients has reduced US mortality after surgery

Published on 24 April 2019

doi: 10.3310/signal-000764

Improved management of deteriorating patients with surgical complications has reduced the number of deaths in US hospitals rather than it being due to fewer complications. Over the past 10 years, complication rates have remained fairly similar. It is the reduction in 'failure to rescue' that has made the main difference in mortality. It is unclear if this is because of earlier detection of patients who are deteriorating due to complications, or improved response and treatment.

This large observational study tracked mortality, rates of serious complications, and failure to rescue of over 700,000 older adults following major surgery in Medicare hospitals.

These US findings support the recently introduced early warning system in the NHS called NEWS2. This aims to detect which patients are deteriorating so that management can be escalated. This is a hospital-wide initiative for all adult patients, not just following surgery.

Share your views on the research.

Why was this study needed?

Surveys estimate that a third of potentially preventable deaths in the UK are due to inadequate clinical monitoring. Patients recovering from major surgery are at particularly high risk of sudden deterioration.

An earlier review suggested that junior staff may take physiological measurements but may not be aware of their meaning and that hospital hierarchy was a barrier to calling rapid response teams.

Though overall deaths after major surgery have reduced in recent years, there is still some variation between hospitals. This study aimed to identify which factors could help improve mortality rates at poorer performing hospitals.

What did this study do?

This retrospective cohort study used US national data from Medicare, the federal health insurer. It included 702,268 people older than 65 years who had specific surgery in 3,404 hospitals from 2005 to 2014. Operations of interest were repair of abdominal aortic aneurysm, or removal of part or all of the colon, lung or pancreas. These types of surgery were chosen because they are common operations with high risk of adverse events.

The study looked at changes over time in mortality within 30 days of surgery and complications. If patients died after having at least one major complication, this was called failure to rescue.

This large study used a robust set of national data. However, differences in the demographics and healthcare provision in the US and UK may limit its applicability.

What did it find?

  • Over the period studied, the proportion of patients who died within 30 days of surgery fell by 16% from 7.7% to 6.5%; the proportion of patients with serious complications fell by 3% from 14.8% to 14.4%; and the proportion of people who died because of failure to rescue fell by 11% from 22.9% to 20.3%.
  • When hospitals were split into five groups depending on their level of improvement over the period, the best performing fifth of hospitals reduced mortality by 37% from 9.0% to 5.7%; reduced serious complications by 11% from 15.2% to 13.5%; and reduced failure to rescue by 25% from 25.2% to 18.9%.
  • The worst performing fifth of hospitals increased mortality by 12% from 6.9% to 7.7%; increased serious complications by 5% from 14.6% to 15.4%; and increased failure to rescue by 4% from 21.5% to 22.3%.
  • Overall, 64% of the variation in mortality over time was explained by rates of failure to rescue. Rates of serious complications explained 4–7% of the variation in mortality. About 30% of the variation in mortality was unexplained and may be due to factors not measured in this study.

What does current guidance say on this issue?

NICE’s 2007 guideline on recognising and responding to deterioration in acutely ill adults will be amended to note that the NEWS2 early warning score has been endorsed by NHS England and the Royal College of Physicians.

NEWS2 assesses respiration rate, oxygen saturation, systolic blood pressure, pulse rate, level of consciousness or new confusion, and temperature. Each measure is scored and combined to give an overall assessment of how ill a patient is.

Higher scores trigger increasingly urgent clinical responses and increased frequency of monitoring. NEWS2 should now be in use by all hospitals in England since the end of March 2019.

What are the implications?

This study supports current recommendations to monitor patients for signs of deterioration in their clinical condition and to respond adequately, as reducing failure to rescue was the main driver for the improved mortality.

The study did not explain how the improvements were made and did not mention any national US early warning systems. This differs from the UK where NEWS2 is becoming the standardised process, which should improve communication and response. Adopting the same system across hospitals will also be beneficial considering staff movement, high usage of locums and bank staff.

Citation and Funding

Fry BT, Smith ME, Thumma JR et al. Ten-year trends in surgical mortality, complications, and failure to rescue in Medicare beneficiaries. Ann Surg. 2019; Jan 23. doi: 10.1097/sla.0000000000003193 [Epub ahead of print].

No funding information was provided for this study.

Bibliography

Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med. 2009;361(14):1368-75.

McGaughey J, O'Halloran P, Porter S, Blackwood B. Early warning systems and rapid response to the deteriorating patient in hospital: a systematic realist review. J Adv Nurs. 2017;73(12):2877-91. 

NICE. Acutely ill adults in hospital: recognising and responding to deterioration. CG50. London: National Institute for Health and Care Excellence; 2007.

RCP. National Early Warning Score (NEWS) 2. Standardising the assessment of acute-illness severity in the NHS. London: Royal College of Physicians; 2017.

 

Why was this study needed?

Surveys estimate that a third of potentially preventable deaths in the UK are due to inadequate clinical monitoring. Patients recovering from major surgery are at particularly high risk of sudden deterioration.

An earlier review suggested that junior staff may take physiological measurements but may not be aware of their meaning and that hospital hierarchy was a barrier to calling rapid response teams.

Though overall deaths after major surgery have reduced in recent years, there is still some variation between hospitals. This study aimed to identify which factors could help improve mortality rates at poorer performing hospitals.

What did this study do?

This retrospective cohort study used US national data from Medicare, the federal health insurer. It included 702,268 people older than 65 years who had specific surgery in 3,404 hospitals from 2005 to 2014. Operations of interest were repair of abdominal aortic aneurysm, or removal of part or all of the colon, lung or pancreas. These types of surgery were chosen because they are common operations with high risk of adverse events.

The study looked at changes over time in mortality within 30 days of surgery and complications. If patients died after having at least one major complication, this was called failure to rescue.

This large study used a robust set of national data. However, differences in the demographics and healthcare provision in the US and UK may limit its applicability.

What did it find?

  • Over the period studied, the proportion of patients who died within 30 days of surgery fell by 16% from 7.7% to 6.5%; the proportion of patients with serious complications fell by 3% from 14.8% to 14.4%; and the proportion of people who died because of failure to rescue fell by 11% from 22.9% to 20.3%.
  • When hospitals were split into five groups depending on their level of improvement over the period, the best performing fifth of hospitals reduced mortality by 37% from 9.0% to 5.7%; reduced serious complications by 11% from 15.2% to 13.5%; and reduced failure to rescue by 25% from 25.2% to 18.9%.
  • The worst performing fifth of hospitals increased mortality by 12% from 6.9% to 7.7%; increased serious complications by 5% from 14.6% to 15.4%; and increased failure to rescue by 4% from 21.5% to 22.3%.
  • Overall, 64% of the variation in mortality over time was explained by rates of failure to rescue. Rates of serious complications explained 4–7% of the variation in mortality. About 30% of the variation in mortality was unexplained and may be due to factors not measured in this study.

What does current guidance say on this issue?

NICE’s 2007 guideline on recognising and responding to deterioration in acutely ill adults will be amended to note that the NEWS2 early warning score has been endorsed by NHS England and the Royal College of Physicians.

NEWS2 assesses respiration rate, oxygen saturation, systolic blood pressure, pulse rate, level of consciousness or new confusion, and temperature. Each measure is scored and combined to give an overall assessment of how ill a patient is.

Higher scores trigger increasingly urgent clinical responses and increased frequency of monitoring. NEWS2 should now be in use by all hospitals in England since the end of March 2019.

What are the implications?

This study supports current recommendations to monitor patients for signs of deterioration in their clinical condition and to respond adequately, as reducing failure to rescue was the main driver for the improved mortality.

The study did not explain how the improvements were made and did not mention any national US early warning systems. This differs from the UK where NEWS2 is becoming the standardised process, which should improve communication and response. Adopting the same system across hospitals will also be beneficial considering staff movement, high usage of locums and bank staff.

Citation and Funding

Fry BT, Smith ME, Thumma JR et al. Ten-year trends in surgical mortality, complications, and failure to rescue in Medicare beneficiaries. Ann Surg. 2019; Jan 23. doi: 10.1097/sla.0000000000003193 [Epub ahead of print].

No funding information was provided for this study.

Bibliography

Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med. 2009;361(14):1368-75.

McGaughey J, O'Halloran P, Porter S, Blackwood B. Early warning systems and rapid response to the deteriorating patient in hospital: a systematic realist review. J Adv Nurs. 2017;73(12):2877-91. 

NICE. Acutely ill adults in hospital: recognising and responding to deterioration. CG50. London: National Institute for Health and Care Excellence; 2007.

RCP. National Early Warning Score (NEWS) 2. Standardising the assessment of acute-illness severity in the NHS. London: Royal College of Physicians; 2017.

 

Ten-year Trends in Surgical Mortality, Complications, and Failure to Rescue in Medicare Beneficiaries

Published on 1 January 2019

Fry, Brian T.,Smith, Margaret E.,Thumma, Jyothi R.,Ghaferi, Amir A.,Dimick, Justin B.

Annals of Surgery Volume Publish Ahead of Print , 2019

Mini This study examines how reductions in postoperative complications and improvements in failure to rescue have contributed to improvements in surgical mortality over the past decade. Improvements in rescue explained the majority of observed improvements in surgical mortality, whereas decreased complication rates explained a small proportion of this improvement. Objective: To evaluate how changes in complication and failure to rescue rates influence hospitals??? postoperative mortality rates. Summary Background Data: Surgical mortality has declined over the last decade, but the mechanisms underlying these improvements are unknown. Specifically, the relative impact of reducing postoperative complications versus improving ???failure to rescue??? remains unclear. Methods: Using Medicare claims data, we performed a retrospective study of abdominal aortic aneurysm repair, pulmonary resection, colectomy, and pancreatectomy patients. We examined risk-adjusted 30-day mortality, serious complications, and failure to rescue for these patients in from 2005 to 2014 (n = 702,268 patients in 3404 hospitals). Hospitals were then stratified into quintiles by their change in mortality over time. Results: After stratifying by reductions in mortality from 2005 to 2014, the top 20% of hospitals decreased mortality by 37% (9.0%???5.7%, P < 0.001), decreased serious complications by 11% (15.2%???13.5%, P < 0.001), and decreased failure to rescue by 25% (25.2%???18.9%, P < 0.001). In contrast, the bottom 20% of hospitals increased mortality by 12% (6.9%???7.7%, P < 0.001), increased serious complications by 5% (14.6%???15.4%, P < 0.001), and increased failure to rescue by 4% (21.5%???22.3%, P < 0.001). Partitioning of variance demonstrated that decreased failure to rescue explained 64% of improvement in hospitals??? mortality over time, whereas decreased serious complications accounted for only 5% of this improvement. Conclusions: Hospitals with the largest reductions in surgical mortality achieved these improvements primarily through reducing failure to rescue rates and not by reducing serious complication rates. This suggests that hospitals aiming to reduce surgical mortality should engage in efforts focused on improving rescue. Reprints: Brian T. Fry, MS, Center for Healthcare Outcomes and Policy, 2800 Plymouth Road, Building 16, Ann Arbor, MI 48109. E-mail: brianfry@med.umich.edu. J.B.D. receives grant funding from the National Institutes of Health, the Agency for Healthcare Research and Quality, and is a cofounder of ArborMetrix, Inc, a company that makes software for profiling hospital quality and efficiency. The company had no role in the study herein. A.A.G. is supported through grants from the Agency for Healthcare Research and Quality and a Patient Centered Outcomes Research Institute Award. B.T.F. is supported by National Institutes of Health grant 1TL1TR002242 through the Master of Science in Clinical Research program at the University of Michigan. J.R.T. and B.T.F. had access to all of the data in this study and take full responsibility for the integrity of the data as well as the accuracy of the data analysis. The authors report no conflicts of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.annalsofsurgery.com). Copyright ?? 2019 Wolters Kluwer Health, Inc. All rights reserved.

Expert commentary

This paper is of great importance to surgical teams throughout the UK. It uses data on 702,268 patients in 3,404 hospitals, in four types of surgery under US Medicare, 2005-14.

The main driver was failure to rescue measured by the death of patients with complications. Decreased failure to rescue explained 64% of the difference in mortality. The top hospitals also decreased serious complications, but this was less important.

The paper recommends actions to improve “timely recognition and management of post-operative complications.” Factors such as interprofessional communication and teamwork may be more important than hospital volume for these procedures.

Nick Bosanquet, Emeritus Professor of Health Policy, Imperial College London

The commentator declares no conflicting interests

Expert commentary

When a patient develops complications after surgery and then dies, we describe this as ‘failure to rescue’.

This research suggests that while some American hospitals have done well in preventing failure to rescue, others have got worse. We work hard to prevent failure to rescue in the NHS but may still have similar variation between hospitals. This paper also suggests the number of surgical patients experiencing complications has changed little in ten years.

Complications are impossible to completely prevent in older patients with multi-morbidities. This is worrying because we offer surgical treatments to older patients much more frequently than ten years ago.

Rupert Pearse, NIHR Research Professor & Consultant in Intensive Care Medicine, Clinical Director for Research & Development, Queen Mary University of London and Barts Health NHS Trust

The commentator declares no conflicting interests