1. 
   

   What are the causes of inappropriate test ordering?

   
   
2. 
   

   Why is it important to address the problem of inappropriate test ordering?

   
   
3. 
   

   How do you quantify the diagnostic accuracy of a test?

   
   
4. 
   

   With any procedure that entails more than negligible risk, the decision to perform the study must take into account what factors?

Each field of medicine challenges clinicians to recommend a course of action for a specific patient at a particular time. To efficiently and safely obtain the right information through testing, physicians need to ask the right questions. Clinical expertise assumes that the physician has the ability to

• Comprehend the information that the test will provide relevant to the decision at hand.
  
• Appreciate the situation and consequences of each approach.
  
• Utilize the information rationally in the context of a coherent set of care goals and ethical values.
  
• Explicitly communicate choices regarding care to the patient or surrogate so that there is informed consent.

Prior to ordering any examination, clinicians should first question whether they require additional data collection to optimize medical decision-making. Clinicians need to weigh (1) medical indications, (2) patient preferences, (3) quality of life, and (4) contextual features to reach a decision that is right for the patient. Deciding whether or not to perform a diagnostic study involves balancing its risks and costs versus the information it could provide and the benefits and costs of having that information. As a general principle, clinicians should select the least invasive imaging examination that provides the needed information with the least amount of risk, including the smallest dose of ionizing radiation, and at the least cost.

There are times when the physician’s most important job is to know when to ask for assistance and where to go to get that assistance. Initial decisions regarding the care of an acutely ill hospitalized patient must be continually reassessed in light of both new data and the patient’s ongoing course.

The process of clinical reasoning—data collection, problem formulation, and the generation of a hypothesis—ideally results in pursuing the most viable hypotheses. One of the most important first steps in approaching a patient is to frame the problem either in term of a diagnosis or syndrome. The process of weighing the probability of one disease versus that of other diseases possibly accounting for a patient’s illness leads to the creation of a differential diagnosis. This process requires both a pathophysiologic knowledge of the potential causes of a problem and an understanding of the effect of new data on the probability of each potential cause. It is important to consider all possible diagnoses but especially those that are either life-threatening or for which effective treatment exists.

Physicians generally develop a working hypothesis that is based on partial information. This is particularly true in the hospital when the patient is acutely ill and urgent decisions must be efficiently made, often without access to outside ambulatory records. Furthermore, elderly patients commonly have multiple comorbidities and their admission may be precipitated by a decreased ability to perform ADLs or nonspecific symptoms such as lethargy, confusion, incontinence, or falls (Table 101-1). Physicians must be willing to reassess initial decisions when there is new data and the patient’s ongoing course runs counter to expectations.

Inappropriate test ordering refers to both under and over utilization. Physicians commonly repeat studies due to failure to properly review the medical record or because these tests are readily available and easier to obtain than outside information. We hope that these chapters will help the clinician avoid inappropriate testing due to not requesting the best test, underestimating the risk of testing, ordering tests “too soon” after the initial test, or ordering tests that are unlikely to affect patient management (Table 101-2).