While on call on the internal medicine service at my hospital, I recently admitted a 47-year old woman overnight, who had increased swelling in her ankles and a fluid collection in her abdomen. After taking a thorough clinical history and performing a complete physical examination, I presumed the cause to be alcoholic liver cirrhosis.
I remembered seeing similar patients as a medical student, and observing the residents supervising me at that time order a battery of tests to ‘rule out’ all other causes of liver cirrhosis. Trying to be a ‘thorough’ physician, I emulated this ‘shotgun’ approach to ordering diagnostic tests for my patient with liver cirrhosis, including an echocardiogram, since heart failure is included on the list of possible causes. I had hoped my staff would be proud of me in the morning that I had provided such comprehensive care for my patient – he was not.
Diagnostic testing, and medical imaging in particular, is becoming increasingly easy for physicians to order . At the same time, the cost of providing care is rising (of course the rise in healthcare costs is multifactorial, with diagnostic testing only one component of total healthcare costs). According to a recent report from the Canadian Institutes for Health Information, the growth in Canada’s total health expenditure increased by an average of $9.2 billion yearly. Canadians now spend $5802/person on healthcare (up from $3213/person in 2000), which accounts for approximately 11% of Canada’s gross domestic product, 40% of provincial budgets, and exceeds the rate of economic growth over the same time period.
Providing cost effective care is challenging, especially given physicians’ discomfort with diagnostic uncertainty. Traditionally, medical students have been taught to create a differential diagnosis of diseases for a patient’s clinical presentation and then address each of these items in a systematic way to eliminate them from their list.
This strategy of ‘shotgun testing’ has been demonstrated by Calderon and his colleagues to be an expensive and ineffective approach to patient care. They examined how effective diagnostic tests (postural vitals, cardiac enzymes, ECGs, telemetry, echocardiograms, electroencephalograms and head CTs) were when routinely ordered to identify the cause of fainting (known as syncope) in elderly patients. The simplest and cheapest test was also the most infrequently ordered (postural vitals; $5/test ordered in 38% of the cases), but the most effective in determining the cause (15%). Compare this to head CTs that were ordered 63% of the time, cost ~$1500/test and only identified the cause 0.5% of the time, never mind the radiation to which these patients were exposed. Yet in the emergency room, I can get a head CT for my elderly patient presenting with a fall as easily as I can get a set of postural vitals!
Thankfully, many thoughtful physicians (including my disappointed staff the next morning) are trying to shift our focus to obtaining valuable information and making responsible decisions with our test ordering. The Choosing Wisely campaign”focuses on encouraging physicians, patients and other health care stakeholders to think and talk about medical tests and procedures that may be unnecessary, and in some instances can cause harm.“ It publishes a list of evidence-based recommendations aimed at ensuring the most appropriate care is individualized to patients, with the hope that they will stimulate discussion about the need—or lack thereof—for many frequently ordered tests or treatments.
For the evaluation of syncope, the American College of Physicians’ Choosing Wisely recommendation is to avoid obtaining brain-imaging studies unless there is an abnormality in their history and physical exam to suggest otherwise – a recommendation supported by Dr. Calderon’s study.
In the case of my patient with alcoholic cirrhosis, unfortunately there was no Choosing Wisely list to tell me what to do (or what not to do!). But I learned a valuable lesson in listening to her, performing a thorough physical exam and heeding the advice of my staff – to utilize good clinical reasoning with regards to her care and avoid ordering tests to rule out diseases that my patient had very little probability of having based on my initial interactions with her. She knew why her liver was damaged; what she really wanted was sound medical advice in how to manage her disease going forward.
So I cancelled the echocardiogram, saved my patient the trouble of an unnecessary test and hopefully saved the system a few bucks at the same time.