Evidence Atlas and The POCUS Atlas answer different questions. The image atlas helps learners examine the appearance of ultrasound findings; Evidence Atlas summarises evidence about selected applications. Recognising a finding and understanding how it changes diagnostic probability are related, but neither task can safely stand in for the other.
This comparison is published by iatroX and includes iatroX as a supplementary route for investigating clinical questions. It draws on public documentation checked on 23 September 2026, not an independent validation study of either atlas or a test of iatroX's image interpretation.
Two resources, two starting points
The Evidence Atlas App Store description, checked on 23 September 2026, describes a free iPhone reference with visual summaries of POCUS evidence, sensitivity and specificity, likelihood ratios, example clips and links to The POCUS Atlas. Its stated comparison is with a reference or gold-standard test for the application concerned.
The POCUS Atlas, checked on the same date, presents contributed ultrasound educational material. Its image examples are a natural starting point when the question is what a finding looks like. Evidence Atlas becomes more relevant when the learner asks what that finding contributes to a diagnostic assessment.
Neither description means that every application has equivalent evidence, that all summaries are current to the app's latest software update, or that a novice will reproduce a study's results.
Translate the statistics before using them
Sensitivity asks how often a test is positive among people who have the target condition. Specificity asks how often it is negative among people who do not. Positive predictive value instead asks what proportion of positive results represent the condition in the population being tested. These are different denominators, not alternative labels for a single accuracy score.
Likelihood ratios connect a test result with pre-test odds, which can then be converted to post-test probability. The MSD Manual's explanation of diagnostic tests, reviewed in April 2026, explains these distinctions and why results require clinical context.
The practical reading question is therefore not "Is this a good number?" It is "Which result, population and clinical question does this number describe?" An impressive-looking statistic without those details is difficult to apply responsibly.
A fictional numbers exercise
Consider an invented teaching example, not data from either atlas or an ultrasound study. Suppose a hypothetical test is used in a group of one hundred people: twenty have the condition and eighty do not. It correctly identifies sixteen of the twenty affected people and correctly returns a negative result in seventy-two of the eighty unaffected people.
Its sensitivity is sixteen divided by twenty, or 80%. Its specificity is seventy-two divided by eighty, or 90%. However, there are twenty-four positive results in total: sixteen true positives and eight false positives. The positive predictive value is therefore sixteen divided by twenty-four, approximately 67%, not 80% or 90%.
These figures are deliberately fictional and have no clinical threshold attached. The exercise demonstrates why "I recognise a positive finding" does not mean "the diagnosis is certain". It also shows why the population and the purpose of testing belong in the discussion.
Read the conditions around the result
When reviewing an evidence summary, inspect the target population, definition of a positive finding, reference standard, timing and handling of inconclusive examinations. Ask whether the operators and setting resemble the situation you are considering. These are appraisal questions, not assumptions that the app has already resolved them.
The reference standard also deserves attention. A study comparing one examination with another is not necessarily asking the same question as a study using follow-up outcomes. Before combining conclusions, establish what each study counted as the condition being present.
For teaching, have learners write a one-sentence applicability statement: "This evidence concerns this application, in this kind of population, using this definition." If the summary does not supply enough detail, the next step is the underlying paper, not a stronger conclusion.
Keep software maintenance separate from evidence currency
The Evidence Atlas listing checked on 23 September 2026 includes recent compatibility and performance updates. Those release notes describe software maintenance. They do not, by themselves, establish that every evidence summary has been newly reviewed or that a particular paper has been incorporated.
A reader should therefore check the dates attached to the cited evidence and any content-specific update information. An app may run well on a current phone while containing an older study that remains useful, or one whose applicability now needs reconsideration.
The converse is also important: an older publication is not automatically invalid. The relevant questions concern methods, subsequent evidence and the decision being made, not simply whether a date appears recent enough.
Combine the tools in a teaching session
A useful proposed session begins with an educational clip and a written description. After the learner commits to an interpretation, discuss what the image supports and what remains unseen. Only then move to the evidence question.
A fictional trainee might recognise the intended finding but be unable to explain what a negative examination would mean. That learner does not necessarily need more copies of the same positive image. They may need a discussion about the study population, the definition of an adequate examination and the limits of a negative result.
Another learner may understand the statistical vocabulary but misidentify the structure in the clip. For that person, more evidence summaries will not fix the immediate visual gap. Return to image orientation and supervised explanation first.
Where iatroX fits, and where it does not
Per iatroX product information, September 2026, Ask-iatroX provides free linked-source clinical reference, and the Socratic Tutor supports follow-up questioning around attempted examination questions. These are possible routes for clarifying a reasoning or evidence-interpretation gap, not a substitute for either atlas's specialist visual material.
The published iatroX methodology describes retrieval, ranking, citation grounding, checking and uncertainty handling. Those are system-design features, not proof that every answer or statistical explanation is correct. The same scrutiny of sources and applicability should be applied to iatroX.
For a visual-recognition problem, begin with the image atlas and appropriate teaching. For a question about published test performance, begin with Evidence Atlas and follow its sources. For applying the finding in practice, add clinical context, local pathways and supervision. There is no meaningful single winner across these different tasks.
Frequently asked questions
Is Evidence Atlas an alternative to POCUS Atlas?
It is better understood as a complementary evidence resource. Its published design focuses on test characteristics, while POCUS Atlas provides ultrasound examples for visual study.
Does high sensitivity mean a positive result proves the diagnosis?
No. Sensitivity concerns positive results among people with the condition; the meaning of a positive result also depends on false positives and the population being tested.
Does an app update mean its evidence has been updated?
Not necessarily. Compatibility or performance release notes do not establish a content review, so inspect the dates and sources of the relevant evidence summary.
Work through the reasoning behind a clinical evidence question →