Symptom-checker tools

From Clinfowiki
Jump to: navigation, search

I. Introduction: Symptom checkers can be divided into two main categories, clinical decision support systems, and computerized self-triage and self-diagnosis tools. The use of both applications is not infrequent, as the differential diagnosis generator tools (symptom checker for providers) have been used by thousands of providers and medical students across the globe (1). A single self-triage app [i-triage] has reported about 50 million users per year (2).

II. Symptom-Checker tools for Providers (Differential Diagnosis Generators):' A recent study evaluated more than 20 web-based and app-based tools designed as symptom checkers and DDx generators for providers. Such diagnostic systems can be classified into the following three categories. First are the systems that use structured knowledge, such as the Internist-I system that was designed in the 70s and “uses disease-finding relations and disease-disease relations, with associated numbers such as sensitivity — the fraction of patients with a disease who have a finding” (3). The MYCIN system and the DXPlain® Tool, both use structural knowledge. A relatively recent tool that was designed using structured data is the Iliad system. Second are the systems that use unstructured knowledge; such tools started to rise in the 2000s, such as PEPID™ and Isabel©. Third are the systems that use clinical algorithms and roles (3). These applications can also be appropriately divided into 1) general apps and 2) specialty specific apps. Some tools will be used only for a specific age group or in specific disease conditions, such as neurology diseases (4).

A. General Tools: Below are popular symptom checker systems selected by a recent audit study (4)

1. PEPID™ uses its own independent content, this tool generates DDX lists and diagnoses based on “a proprietary scoring system related to the number of selected signs/symptoms consistent with each potential diagnosis. Additionally, each sign/symptom is assigned a unique score/weight relative to its importance in differentiating among specific diagnoses”(4). (3, 4) 2. DXPlain® was developed in the early 1980’s by the Laboratory of Computer Science (LCS) at Massachusetts General Hospital (MGH). It generates a differential diagnosis based on disease prevalence, “the frequency of the finding in the disorder, and how strongly it suggests that disease”(4) using a probability algorithm to analyze different clinical manifestations such as (symptoms, physical exam findings, lab and imaging data (1, 4). 3. Isabel© uses natural language processing (NLP) as one of the unique features of this tool, and its design “uses autonomy information retrieval software and a database of medical textbooks to retrieve appropriate diagnoses given input findings”(3). 4. DiagnosisPro® was seen as “an excellent database of signs, symptoms, and abnormal tests, but is not useful as a diagnostic tool”(5) in a review published in JAMA. However, its website reported more than 60,000 visits per month in 2008 (6).

B. Specialty-Specific Tools Examples (4):

1. Gideon (Infectious diseases)

2. Visual Diagnosis (Disorders with visible symptoms)

3. IDDX (Infectious diseases)

4. Oral Radiographic Differential Diagnosis

5. Brain Expert

6. Bone Expert

7. Dermatology Differential Diagnosis

8. Aphasia DX

9. VisualDX

10. DCSA: Diagnosis of Congenital Syndromes and Anomalies

11. SimulConsult (Pediatric Neurology)

C. Triage Aid Symptom Checker Tools:

Symptom checker systems can also be used as a triage aid tool, and ExpertRN© is a good example of such a tool. It is one of the clinical decision support systems that have been used as a triage program by the Mayo Clinic. It includes about 140 symptom-related algorithms and was used more than 60,000 times in 2011 alone. These algorithms will be activated after a triage nurse receives a phone call from a patient and then decides what algorithm to activate, such as chest pain or abdominal pain. The outcome of the call will result in either calling 911, or specifying the urgency of the care needed for the patient i.e. how soon and also the appropriate location for the care “(ambulance, emergency department, doctor’s office appointment, call to the office, or self-care only)”. It has been shown that this CDS system improved the documentation quality of the telephone triage notes (7).

III. Computerized self-Triage & self-diagnosis tool

A. Introduction: The use of Google and other web-based tools to search for health-related answers is common; this, in addition to other factors, led to the development of online applications that aim to facilitate self-diagnosis of clinical conditions (2). Other tools were developed to help self-diagnosis in addition to facilitating self-diagnosis. The use of such tools is not infrequent, to the extent that a recent article was recently was entitled “Symptom Checkers: Triage of the future?” (8).

B. Accuracy: A recent study looked at 23 symptom checkers, found that such applications were able to list the accurate diagnosis first in the list of differential diagnosis in 34% of the cases. Additionally, more than half of the time, the correct diagnosis will be listed within the top three differential diagnoses. The accuracy was found to be higher for non-emergent and common conditions. It also showed that appropriate triage advice would be given by the symptom-checker more than half of the time (52-61%) with a higher accuracy in emergent cases and for uncommon cases. For both diagnosis and triage accuracy, standardized patient scenarios were used to test accuracy (9).

C. Current status: It’s important to acknowledge the fact that these tools still require more optimization and cannot substitute for consulting health care providers. A recent publication concludes that “symptom checkers can give the user a sense of possible diagnoses but also provide a note of caution, as the tools are frequently wrong and the triage advice overly cautious,” (9).

D. Categories: A different method of categorizing self-triage and self-diagnosis tools are done by looking at the type of organization that provides the symptom checker or deciding whether the symptoms-checker is using standard nurse triage guidelines or not.

E. Examples (9, 10):

1. Tools designed by a general foundation, such as, which was created by the American Academy of Family Physicians provide an online symptom checker. It begins with a single general symptom, uses flowcharts of yes/no questions, and results in a single general diagnosis and triage advice.

2. Institution-based tool: The Mayo Clinic symptom checker starts with general symptoms, and then it uses a checklist of other symptoms and risk factors. The outcome generates a list of potential diagnoses, rather than a single diagnosis and provides educational materials.

3. Insurance provider tools: Itriage is a triage tool designed by Aetna (insurance provider) that mainly provides triage advice and also provides a list of providers based on the consumer’s geographic location.

4. Commercially produced tools, such as WebMD use graphics and suggested questions to ask your doctor. Additionally, it provides a list of diagnoses and action to take, but does not provide self–triage advice.

5. Other tools, such as Esagil, incorporate blood and urine lab results, in addition to symptoms, in order to populate a list of potential diagnoses, however, no triage recommendations are offered.

References: 1. Elkin PL, Liebow M, Bauer BA, Chaliki S, Wahner-Roedler D, Bundrick J, et al. The introduction of a diagnostic decision support system (DXplain) into the workflow of a teaching hospital service can decrease the cost of service for diagnostically challenging Diagnostic Related Groups (DRGs). Int J Med Inform. 2010;79(11):772-7.

2. Wyatt JC. Fifty million people use computerised self triage. BMJ : British Medical Journal. 2015;351.

3. Ferrucci D, Levas A, Bagchi S, Gondek D, Mueller ET. Watson: Beyond Jeopardy! Artificial Intelligence. 2013;199(Supplement C):93-105.

4. Bond WF, Schwartz LM, Weaver KR, Levick D, Giuliano M, Graber ML. Differential diagnosis generators: an evaluation of currently available computer programs. J Gen Intern Med. 2012;27(2):213-9.

5. Aronson AR. Diagnosispro: The ultimate differential diagnosis assistant. JAMA. 1997;277(5):426-.

6. [Available from:

7. North F, Richards DD, Bremseth KA, Lee MR, Cox DL, Varkey P, et al. Clinical decision support improves quality of telephone triage documentation--an analysis of triage documentation before and after computerized clinical decision support. BMC Med Inform Decis Mak. 2014;14:20.

8. Symptom Checkers: Triage of the future?

 [Available from:

9. Semigran HL, Linder JA, Gidengil C, Mehrotra A. Evaluation of symptom checkers for self diagnosis and triage: audit study. BMJ. 2015;351:h3480.

10. Smith CA, Keselman A. Meeting health information needs outside of healthcare: opportunities and challenges. xx, 356 pages p.

Submitted by (Aziz Alhomod)