Reducing variation in surgical care

Variations in surgical care have been recognised since the early 1980s and are generally interpreted as evidence of uncertainty among practitioners regarding optimal care. The prescription for remedying variations in surgical practice has generally included development of better medical evidence to identify best practices, dissemination of medical evidence to surgeons, and use of practice guidelines and care pathways to streamline care. More than 20 years later, there is still abundant evidence that surgical care varies substantially. Why do variations in surgical care persist? And what can be done about them?

In this issue of BMJ, Lassen et al (p 1420) report the results of a survey of lead surgeons in five north European countries regarding processes of care for colorectal surgery.¹ These processes included use of preoperative bowel preparation, routine postoperative nasogastric decompression, and use of epidural analgesia in the postoperative surgical ward. The survey identified substantial international variation in the use of such perioperative interventions. In many cases, this variation occurred in spite of abundant, high quality medical evidence.

Surgery and evidence based medicine have had turbulent relations of late. There is broad recognition that much of current surgical practice is not informed by solid medical evidence² and that the application of methods such as randomised controlled trials to surgical questions is often difficult or impractical.³ However, there is growing uniformity of opinion in the surgical community that the quality of evidence supporting surgical care must be improved and that we need innovative methods for disseminating evidence into practice.⁴

Surgery is disadvantaged when it comes to health research. Unlike trials of new drug treatments, research on surgical procedures has no natural sponsor. The lack of strict regulatory mechanisms for the approval of surgical procedures and devices in most countries leads to a situation where large randomised trials are not necessary for surgical interventions to be adopted. How surgical procedures and perioperative care are provided by different surgeons or hospitals varies enormously, leading to an “expertise bias” that may make a valid randomised controlled trial impossible in many circumstances⁵ and in others may limit the generalisability of results to other care settings.

Problems with evidence based surgery notwithstanding, we would argue that the immediate challenge to improving the quality of surgical care is not discovering new knowledge, but rather how to integrate what we already know into practice. Lassen et al¹ and others remind us that much may be achieved simply by raising the quality of surgical care according to existing evidence. In many ways, this is a far more difficult task than simply doing more research. Translating best evidence into surgical practice will require the engagement of large numbers of individual practitioners, in a multitude of healthcare contexts, and using a variety of techniques.

Evidence from before and after studies in several countries shows that surgeons' behaviour can indeed be changed and the quality of surgical care can be improved. In Norway, surgeons concerned with poor outcomes of surgery for rectal cancer initiated a multi-faceted intervention in 1994 to improve the quality of care according to best practices. This initiative led to a voluntary registry of treatment for rectal cancer with feedback to hospitals as well as postgraduate courses for surgeons and pathologists on optimal surgical techniques and on interpreting pathological tests, reorganisation at specific regional hospitals and specialist gastrointestinal surgeons of surgery for rectal cancer, and access to an opinion leader for surgeons both locally and regionally.⁶ After the intervention, the rate of local recurrence of rectal cancer fell from 28% to 8%, and five year survival increased from 55% to 71%.⁷

In the United States, the Northern New England Cardiovascular Disease Study Group initiated a regional intervention in 1990 to improve the outcomes of coronary artery bypass graft surgery. This intervention comprised anonymous feedback to surgeons of risk adjusted outcome data, training in continuous quality improvement, and a series of site visits by surgeons to observe processes of care in other settings. After the intervention, mortality from coronary artery bypass graft surgery decreased by 24%.⁸

A key factor underlying successful programmes for quality improvement in surgery seems to be the engagement of individual surgeons both locally and regionally, through developing communities of practice.⁹ We now know more about what it takes to translate evidence into knowledge for practising surgeons. Innovative interventions that bring surgeons together may require substantial investment, but they will be worth while if they deliver evidence based surgery.