Foreground questions can be categorized into 5 types, including:

1.Therapy: determining the effect of interventions on patient-important outcomes (symptoms, function, morbidity, mortality, and costs)

2.Harm: ascertaining the effects of potentially harmful agents (including therapies from the first type of question) on patient-important outcomes

3.Differential diagnosis: in patients with a particular clinical presentation, establishing the frequency of the underlying disorders

4.Diagnosis: establishing the power of a test to differentiate between those with and without a target condition or disease

5.Prognosis: estimating a patient’s future course

Clinical questions often spring to mind in a form that makes finding answers in the medical literature a challenge. Dissecting the question into its component parts to facilitate finding the best evidence is a fundamental skill.

One can divide questions of therapy or harm into 4 parts following the PICO framework: patients or population, intervention(s) or exposure(s), comparator, and outcome. For questions of prognosis, you can use 1 of 2 alternative structures. One has only 3 elements: patients, exposure (time), and outcome. An alternative focuses on patient-related factors, such as age and sex, that can modify prognosis: patients, exposure (e.g., older age or male), comparison (e.g., younger age or female), and outcome. For diagnostic tests, the structure we suggest is patients, exposure (test), and outcome (criterion standard).

You need to correctly identify the category of study because, to answer your question, you must find an appropriately designed study. If you look for a randomized trial to inform the properties of a diagnostic test, you will not find the answer you seek.

Different structures or design of studies can investigate different foreground questions. To answer the foreground question that of interest, one should know these different structure or design of clinical studies. Because different study designs can correspond different type of foreground questions, I arrange following discussion due to the type of foreground questions.

Therapy and Harm

To answer questions about a therapeutic issue, we seek studies in which a process analogous to flipping a coin determines participant’s receipt of an experimental treatment or a control or standard treatment: a randomized trial. Once investigator allocate participant to treatment or control groups, they follow them forward in time to determine whether they have, for instance, a stroke or myocardial infarction – what we call the outcome of interest.

When randomized trials are not available, we look to observational studies in which – rather than randomization – clinician or patient preference, or happenstance, determines whether patients receive an intervention or alternative.

Ideally, we would also look to randomized trials to address issues of harm. For most potentially harmful exposures, however, randomly allocating patients is neither practical nor ethical. For instance, one cannot suggest to potential study participants that an investigator will decide by the flip of a coin whether or not they smoke during the next 20 years. For exposures such as smoking, the best one can do is identify observational studies (often sub classified as cohort or case-control studies) that provide less trustworthy evidence than randomized trials.

Differential Diagnosis

For sorting out differential diagnosis, we need a different study design. Here, investigators collect a group of patients with a similar presentation (e.g., painless jaundice, syncope, or headache), conduct an extensive battery of tests, and, if necessary, follow patients forward in time. Ultimately, for each patient the investigators hope to establish the underlying cause of the symptoms and signs with which the patient presented.


Establishing the performance of a diagnostic test (i.e., the test’s properties or operating characteristics) requires a slightly different design. In diagnostic test studies, investigators identify a group of patients among whom they suspect a disease or condition of interest exists (such as tuberculosis, lung cancer, or iron-deficiency anemia), which we call the target condition. These patients undergo the new diagnostic test and a reference standard (also referred to as gold standard or criterion standard). Investigators evaluate the diagnostic test by comparing its classification of patients with that of the reference standard.


A final type of study examines a patient’s prognosis and may identify factors that modify that prognosis. Here, investigators identify patients who belong to a particular group (such as pregnant women, patients undergoing surgery, or patients with cancer) with or without factors that my modify their prognosis (such as age or comorbidity). The exposure here is time, and investigators follow up patients to determine whether they experience the target outcome, such as an adverse obstetric or neonatal event at the end of a pregnancy, a myocardial infarction after surgery, or survival in cancer.