Testing, Testing Everywhere

A patient lies in a hospital bed – pale, feverish, and obviously in pain. But with such nonspecific symptoms, how can the doctor possibly know what might be wrong? The answer, for many care providers, lies in laboratory tests – and the more, the better, in some cases. Who knows what might turn up in the patient’s blood counts, metabolic panels, or renal function? Why risk not testing for something that might provide an answer, when it’s so easy to tick every box on the requisition sheet?

Pathologists know that (at least when it comes to laboratory testing) more isn’t always better. Unnecessary tests cost the laboratory time and resources. They cost the hospital money. And they carry a cost for the patient, too – anxiety, discomfort, and even more severe consequences, such as hospital-induced anemia arising from too many blood draws. Moreover, benefits are never guaranteed; increased testing may not actually reveal the issues affecting the patient, whereas it can result in unnecessary interventions or feed into a never-ending cycle of tests. So why do doctors continue to order tests their patients don’t need?

We spoke to Kevin Eaton, lead author of a recent study (1) on over-utilization, to find out why recommendations against over-testing aren’t working – and what interventions might result in greater success.

What are the dangers of excessive testing?

Several studies have shown that excessive lab testing can lead to hospital-acquired anemia (2)(3)(4), which in turn can lead to unnecessary blood transfusions and worse patient outcomes. Additionally, labs ordered without a high pre-test probability for a disease state are difficult to interpret and often lead to more unnecessary testing, which further contributes to rising costs and patient harm. It’s a vicious cycle. And, of course, whether one test or a dozen too many, phlebotomy can be a painful experience for patients, so it can lead to patient dissatisfaction. In this era of patient-centered care, our first priority is the physical health of those entrusted to us – but we must also be aware of the psychological stressors of hospitalization, and that includes those that arise from medical testing.

Why do many care providers over-test despite recommendations to the contrary?

Many barriers to reducing excessive lab testing have been cited in the literature, including lack of knowledge of lab costs, provider inexperience, change in clinical status, fear of missing a diagnosis, or diagnostic uncertainty. Often, providers feel more comfortable with more information. In some cases, the practice is actually patient-driven; patients want to know if they have a particular problem, and they request or even pressure doctors to do the testing for them, even in situations where it may not be necessary.

Additionally, I think it is difficult for providers to acknowledge (or sometimes even care about) the downstream effects of excessive testing (such as more unnecessary tests, or hospital-acquired anemia). In daily practice, it’s easy to lose track of the repeat CBC or whole blood hemoglobin ordered to follow up on a lab abnormality that turned out not to be a problem in the first place. More importantly, many providers – like me – can easily be influenced by anecdotal clinical encounters. It only takes one bad patient outcome, even one unrelated to a failure to test, to undo efforts to reduce repetitive lab orders. Also, providers may not even be aware of how many lab tests they are ordering in the first place. Busy clinical days leads to poor recall, and the sheer number of diagnostic lab tests ordered on different patients makes it difficult to have a “big picture” idea of one’s ordering habits.

Because of these numerous barriers, many of which vary from provider to provider, strategies to reduce excessive testing can’t rely solely on educational recommendations. As described in our implementation guide, there needs to be a multimodal strategy that includes educational initiatives, feedback for providers on their ordering patterns, and changes made to the electronic ordering systems that make it harder to order unnecessary tests.

What educational initiatives might encourage physicians to order fewer tests?

The key is to focus on educating providers about the appropriate indications for various laboratory tests. These recommendations can be decided by collaborating with various subspecialty experts to form a unified set of indications for various lab tests (perhaps starting with the most expensive and frequently misused). These, of course, vary between institutions but might include red blood cell folate testing, hepatitis C viral load and genotype, type and screens, or reflexive lupus antibody testing without a positive antinuclear antibody.

It’s also important to educate providers on the downstream effects of unnecessary testing. Ordering lab tests without a clinical reason makes the results difficult to interpret. If there is an abnormal value, additional unnecessary testing is often ordered – only to find out, ultimately, that there wasn’t even a problem in the first place. These “cascade” effects contribute to wasted value, patient dissatisfaction, and possibly unintended complications.

Finally, providers must be given personalized information on their own lab ordering practices. Many studies have shown that peer comparisons of lab orders helps to reduce excessive testing. Often, providers may not be aware of the extent of their ordering. Having the numbers in front of them can lead to an epiphany – and if that alone is not enough, comparing their ordering patterns with those of their more conservative peers could prompt them to reconsider their practices!

And how would you recommend reprogramming computer systems?

The approach to computer system adjustments needs to acknowledge that there are, in my opinion, two main categories of lab tests for hospitalized patients. The first are tests ordered to help make the diagnosis of a specific disease state (for instance, thyroid studies, antinuclear antibody, hepatitis antibodies, and more). The second category includes tests ordered to check and monitor a patient’s health – such as blood cell counts, kidney function (basic metabolic panel), or liver function (hepatic panel).

To target the first category, electronic ordering systems can be programmed in a way that educates users on the appropriate indications for specific tests. These can be done in the form of non-intrusive computer alerts, which have been shown to have varying degrees of efficacy (5)(6); it is true that some physicians may ignore them altogether, but others may take note of a pop-up that provides a list of recommended reasons for ordering a particular test. Potentially more impactful, though, would be to have testing algorithms that simplify decision-making by streamlining the appropriate indications for certain tests. For example, rather than having a provider order both a thyroid stimulating hormone (TSH) and a free T4 test, the provider can select a TSH algorithm reflex. The lab would then only reflexively send the free T4 order if the TSH result were abnormal.

To target the second type of labs (as discussed in our new guidelines), the electronic medical record-based strategies we’ve found most effective are those that simply don’t allow providers to order repeating labs. A major contributing factor to over-testing is that labs like CBCs and BMPs are ordered to repeat daily on patients admitted to the hospital – but patients don’t necessarily need these tests every single day of their hospitalization. Programming computer systems to eliminate the “repeating” lab option (or at least to limit the total number of repeats allowed to a maximum of x days) has been shown to reduce repetitive testing without preventing doctors ordering the tests that are truly needed.

How can pathologists tackle the main testing culprits?

The most common labs (CBC, BMP, CMP, and so on) are the ones often ordered as daily repeating labs for hospitalized patients – but they have the potential for significant downsides if the patient has been clinically stable. In the setting of clinical stability, abnormalities on these lab tests are difficult to interpret and often prompt additional testing, which can lead to excessive phlebotomy and hospital-acquired anemia. Anecdotally, I also feel as though rheumatologic tests (such as ANA, dsDNA, or anti-Smith antibody) are also frequently ordered inappropriately and can lead to further unjustified testing on patients. Both pathologists and primary care providers should pay attention to ordering patterns for these kinds of labs – and speak up when excessive testing carries the potential for harm. Be aware, though, that your mileage may vary. Over-utilized lab tests often differ between institutions depending on patient population and providers’ preferences.

The “Choosing Wisely” campaign has made phenomenal progress in encouraging the conversation about overutilization. These recommendations come from experts in different professional societies and lay the groundwork for institutions to advocate for improvement. Our implementation guidelines expand on the “Choosing Wisely” recommendations (7) to prevent lab test overutilization and help provide institutions with evidence-based strategies to implement initiatives for reducing repetitive laboratory testing in their own hospital systems.

What else can pathologists do? They can collaborate with other subspecialty experts to determine standard indications for different lab tests – perhaps starting by targeting the tests that are most expensive, or most often misused. They can educate providers on appropriate indications for different lab tests to help avoid unnecessary testing. They can even help develop safe testing algorithms – for instance, to build lab test orders that reflex to additional tests if (and only if) the first result is abnormal. These initiatives can potentially save a lot of tests from being ordered in the first place, improving the overall health and happiness of our hospitalized patients.

Kevin Eaton is a third-year internal medicine resident in Longcope Firm at Johns Hopkins Hospital, Baltimore, USA.