​Choosing diagnostic tests wisely: Doing the little things well


There was such a wonderful response to the Choosing Surgery Wisely paper from Dr Roland Grad and medical students Nicholas Meti and Mathieu Rousseau, that they have submitted another!

Dr Grad's poster at PODC2015

Dr Grad's poster at PODC2015

You may remember Dr Grad, a family physician and researcher at McGill University, from his poster on harnessing InfoPOEMS to find potential topics for the Choosing Wisely Campaign, which he also presented at the Preventing Overdiagnosis conference this year (PODC2015). [click to view the more recent poster in PDF format]

Again, Rousseau and Meti worked with Dr Grad to extend this work and look at InfoPOEMs that dealt with three topics in diagnostic testing: stable TSH measurements, screening mammography, and mid-stream urine collection. Guided by clinical questions pertaining to these topics and the best available evidence, they make a clear case that we need to choose very wisely when considering 'routine' testing. There are some apt qualitative insights provided by physicians reflecting on the practice-changing POEMs (Patient-Oriented Evidence that Matters) included in this research, which will undoubtedly help it to resonate with readers.

Please feel free to leave questions or comments below or contact the authors directly. If you would like to submit a guest-post for consideration, email lessismoremedicine@gmail.com.

Choosing diagnostic tests wisely: Doing the little things well

Rousseau, M., Meti, N., Grad, R. Faculty of Medicine, McGill University, Montreal, Canada.


As clinicians, do we challenge the appropriateness of our diagnostic test ordering? To achieve shared decision-making in health care, it is up to clinicians to communicate both the harms (as well as the benefits) of diagnostic testing. Within the concept of shared-decision making, there are three core practices: 1) Identifying that a decision must be made; 2) Communicating the potential benefits and harms of options to patients; and 3) Incorporating what is important to patients within the decision. The latter may require us to consider other questions: What would be the impact of this test on the patient’s quality of life? What about the interval between follow-up tests? What is the impact on the economy when diagnostic tests and follow-ups are considered at scale? When clinicians think about ordering a test (or not), we suspect their decision is based on “routines” and “experience”. The point of this post is not to argue against the “art of medicine”, but to raise awareness of new research that can inform decisions about diagnostic testing.

In this post, we highlight the findings of three recent diagnostic test studies. Study findings were disseminated to Canadian physicians as ‘POEMs’. For those unfamiliar with this acronym, POEMs are tailored synopses of primary research or systematic reviews, selected in a process that involves searching over 100 journals. [1] Since 2005, the Canadian Medical Association (CMA) delivers one POEM to their members by email on weekdays. As described in a prior guest post (Choosing Surgery Wisely), we identified the following POEMs by analyzing the ratings of all daily POEMs (n=255) collected from physician members of the CMA in 2014.

CLINICAL QUESTION: “How much do seemingly stable thyroid tests vary over time? / POEM Title: Stable TSH can be rechecked in 2 years”

In a cohort study, the authors asked how frequently do patients with treated hypothyroidism need to have their TSH measured. [2] From a sample size of over 700 persons treated with levothyroxine, they were able to identify a subgroup that would benefit from less frequent TSH monitoring based on their dose of levothyroxine. They report that patients receiving less than 125 micrograms per day could have their TSH rechecked in two years instead of annually. Importantly, this study highlights that once TSH has normalized, the frequency of subsequent monitoring can be stratified based on dosing.

Monitoring frequency is a relevant issue in the clinic setting. In the absence of evidence, many clinicians assume default rates for all manner of diagnostic test and treatment plans. We read the free-text comments submitted by CMA physicians about this POEM. Some of these physicians expressed surprise at the association between dose and frequency of monitoring. Others reported the following: had they known about this approach, they would have spread out the visits for their healthier patients. This would save time and provide costs savings for the healthcare system. Although not addressed by this study, one physician even raised the question of whether we need to be checking TSH levels at all in an asymptomatic patient.


CLINICAL QUESTION: “What are the trade-offs of benefits and harms for women considering a mammogram to screen for breast cancer? / POEM Title: Numbers to help women understand the benefits/ harms of screening mammography”

Welch et al. believe primary care physicians should have more balanced discussions with their patients about the benefits and harms of screening mammography. [3] Their premise is that the majority of discussions focus on the possibility of avoiding death from breast cancer, and do not include a discussion of false alarms nor overdiagnosis. The authors used currently available data from trials of screening mammography to give a range of estimates for harms and benefits with the hope that this information would help decisions about screening. Their results are summarized in this table. Note that the numbers are per-one-thousand women, screened yearly for 10 years:

Figure 1: Estimates of harms and benefits of screening mammography

We received mixed feedback from physicians who read this POEM. Some physicians were grateful to have empiric data to help them in their discussions with patients. One wrote it is “helpful to have the actual numbers presented in such a way that I can share info with the patient when discussing mammograms and screening - always easier when there are numbers that we can look at”, and these numbers “make discussion around breast cancer more objective”. However, others wrote that even though “it is much easier to communicate this information to a patient by simply selecting the age group she falls into, and presenting the numbers for that group [...], I have not yet had a patient who didn't just simply choose the mammogram”. It seems that numbers do not tell the entire story… “because this is an emotional issue, most women we counsel opt for the regular screening”.

The importance of this topic to primary care is high, because as one CMA member wrote “the harms of false positives are seen first-hand in primary care”.

CLINICAL QUESTION: “How accurately does a midstream urine culture predict the results of a catheterized urine culture? POEM Title: Interpretation of midstream urine cultures in healthy young women with suspected UTI”

What about the practice of empirically treating suspected urinary tract infection in otherwise healthy women without relying on culture? In a diagnostic test evaluation study, midstream urine cultures with any evidence of E. coli or K. pneumoniae strongly suggested a true infection, while the presence of enterococci or group B streptococci had little predictive value. [4]

Feedback from physicians who read this POEM showed appreciation for the findings and included comments such as this one: “As a walk-in clinic doctor, urinary symptoms are a very common reason for visits. I routinely treat women on spec for these UTI's and don't send their urine for culture unless it is a complicated UTI, the patient has significant comorbidities, or the patient has recently been on antibiotics.”

The practice of empirically treating suspected urinary tract infection in otherwise healthy women without relying on culture was recommended in a recent review by Grigoryan et al. [5] Her group reviewed the optimal approach for treating acute cystitis in young healthy women and analyzed studies totalling 259 397 patients. This showed that “immediate antimicrobial therapy is recommended rather than delayed treatment or symptom management with ibuprofen alone”.

This choosing-wisely-approach to a common infection was perfectly summarized in this comment submitted by another physician: “great info [in this POEM]. Sometimes we just do too much testing”.


As we reflect on all this, we see that even if one test “can’t hurt”, at scale the impact can be large for publicly funded health care systems. This point has been made by others. For example, Kale et al showed how “routine” diagnostics tests cost large sums of money. [6] Primary health care faces a big challenge in reconsidering how diagnostic testing is used, to ensure better value for all.


1. Grad RM, Pluye P, Tang DL, Shulha M, Slawson DC, Shaughnessy AF. 'POEMs’ suggest potential clinical topics for the Choosing Wisely Campaign. Journal of the American Board of Family Medicine 2015;28:184-189. http://www.jabfm.org/content/28/2/184

2. Pecina J, Garrison GM, Bernard ME. Levothyroxine dosage is associated with stability of thyroid-stimulating hormone values. Am J Med 2014;127(3):240-245 http://www.amjmed.com/article/S0002-9343(13)01021-8/abstract

3. Welch HG, Passow HJ. Quantifying the benefits and harms of screening mammography. JAMA Intern Med 2014; Dec 30 http://archinte.jamanetwork.com/article.aspx?articleid=1792915

4. Hooton TM, Roberts PL, Cox ME, Stapleton AE. Voided midstream urine culture and acute cystitis in premenopausal women. N Engl J Med 2013;369(20):1883-1891 http://www.nejm.org/doi/full/10.1056/NEJMoa1302186

5. Grigoryan L, Trautner BW, Gupta K. Diagnosis and Management of Urinary Tract Infections in the Outpatient Setting. JAMA. 2014;312(16):1677-1684. http://jama.jamanetwork.com/article.aspx?articleid=1917443

6. Kale MS, Bishop TF, Federman AD, Keyhani S. "Top 5" lists top $5 billion. Arch Intern Med 2011;171(20):1856-1858