IDdx is a decision-support software tool developed to assist in the diagnosis of infectious diseases. Having a software tool is not the same as having an infectious disease specialist with years of education and years of experience diagnosing infectious diseases. IDdx is like an electronic book that you can query. It was not designed to replace the human mind, but to assist the professional with lists and intersection of lists.
For example, one can make a list of all infections that cause abdominal pain and another list of all infections caused by animal bites. The intersection of those two lists is all infections that fit both criteria. This is what IDdx was designed to do--first to do the research to make those lists and then to design the software that makes the querying of those lists as easy as possible.
Each finding (signs & symptoms) and epidemiological factor in IDdx required a lot of research to make these lists as complete and accurate as possible. My method in refining those lists is just plain old iteration, defined by Merriam-Webster as "a procedure in which repetition of a sequence of operations yields results successively closer to a desired result."
I look for the best and most recent sources of information, and my starting point is the lastest edition of Control of Communicable Diseases Manual. The latest edition of Infectious Diseases edited by Mandell et al will be available in September 2009 and will have full-text online access. In the last few years, many searchable electronic resources have become available. These make the job easier to build and continuously improve lists for differential diagnosis.
See my website at www.outbreakid.com for more about zoom-intersection and other concepts used in this decision-support software. Also see the bibliography page there for references used in OutbreakID and IDdx. In December 2009, I updated references to PPID, CCDM, CDC Travel, and 5MCC latest editions. These revisions are available by AutoUpdate when you sync your PDA.
The first video is from www.pda4peds. It shows the reviewers "click-through" of the the IDdx PDA application. This application became available for sale in April 2008 at USBMIS, the company that developed the PDA user interface.
The first case is from The New England Journal of Medicine. A 62 year-old male living in South Korea complained of fever and abdominal pain for 7 days. He had recently eaten raw smelt from a pond. He had abnormal liver function tests. CT of the abdomen showed dilatation of the common bile duct. Worms found at the time duodenoscopy were identified as Clonorchis sinensis.
The second case is also from The New England Journal of Medicine. In this case a 34-year old male came to the emergency room after 3 days of high fever. Laboratory tests showed hemolytic anemia with a hemoglobin of 8.6 g per deciliter. He had a splenectomy in 1994 for treatment of Hodgkin's lymphoma. Recent travel was in South Africa, Costa Rica, and within the United States. In the blood smear, 3% of red blood cells harbored the parasite causing babesiosis.
The third case is from Partners Infectious Disease Images from January 2009. A boy in his mid-teens presented with an ulcer of his right hand and lymphadenopathy of the right elbow and axilla associated with intermittent fever for 5-6 days. He had been exposed to pet dogs, cats, and deer. Also, he hunted regularly and had handled deer and rabbit carcasses within the last month. He had also scratched his hands on hay, briars, and inside beaver dams. The WBC count was 16,600 with 47% neutrophils and 13 % bands. Considerations in the differential diagnosis were tularemia, sporotrichosis, blastomycosis, cat scratch disease, and staph cellulitis. Culture and serology were positive for Francisella tularensis.
The fourth case is from Partners Infectious Disease Images for August 2008. A 70-year old female was admitted to the hospital with an 8 week history of sinus and respiratory symptoms. The patient responded to antibiotic treatment, but a CT of the chest showed 2 calcified nodules. Past history was significant for extensive world travelling including the Caribbean and North Africa. Considerations in the differential diagnosis included pulmonary harmatoma, non-tuberculous mycobacteria, dirofilaria immitis, coccidioides, and echinococcus. Wedge resection of the pulmonary nodules showed the presence of a nematode consistent with Dirofilaria immitis.
Some infectious disease cases are difficult to diagnose. Signs and symptoms may be misinterpreted, mixed up, or just missed. Here are some of the difficulties that I found when reviewing cases on the web: 1. An ophthalmologist misnamed a case of optic neuritis as papilledema. 2. A patient was found to have scleral icterus by the emergency room doctor, but the admitting labs showed a normal bilirubin. 3. A rash may be called maculopapular, vesicular, and pustular by different observers at different times. 4. A rash may be caused by a complication of the disease, e.g., Stevens Johnson syndrome, rather than by the disease itself. 5. More than one disease may be causing the signs and symptoms. 6. The disease may be non-infectious such as a metastatic tumor or plaque emboli. 7. Insignificant findings may be magnified while important ones are ignored.
Because of these difficulties, situation awareness is critical to diagnosing with or without decision-support software. This video shows some of the definitions of findings and epidemiological factors that are available in IDdx. Just click the "?" next to the word. For example, "speech, impaired" is defined as "Difficulty speaking; Dysphonia (difficulty using voice); Dysarthria (difficulty articulating words);" It is not defined as aphasia.
What are some strategies and pitfalls in querying the IDdx database?
1. Think of "zoom-intersection" instead of making a shopping list of symptoms. An "AND" query is the same as two intersecting circles in a Venn diagram. You are trying to zoom-in to a differential diagnosis that fits the unique criteria for this case. Common symptoms shared by most infectious diseases, e.g., fever and fatigue, are not very helpful in limiting the differential diagnosis.Are there any unique criteria in this case like eosinophilia, pleural effusions, or petechiae? Are there any epidemiological factors that you can add from the history of this patient? If the patient lives in the United States and has not traveled, limit your query to diseases endemic to North America (assuming that you are not considering a bioterrorism event). 2. If your query gives you a long differential diagnosis, you may need to scroll to the end of the list. For example, the criteria of fever, fatigue, and pharyngitis results in 25 diseases. The diseases are listed in alphabetical order. 3. You may limit the results by selecting in the Acuity drop-down list one of the following: acute-severe, acute-moderate, or subacute/chronic. Be aware that these categories overlap, and diseases may present in different ways. For example, while tuberculosis is classified as subacute/chronic, it may sometimes present as either a acute-severe (emergency admission), or acute-moderate (recent onset of symptoms treated in a clinic). 4. Clear all criteria before each new query. There is a "Clear" button at the bottom of the screen, and this should be clicked before starting a new query. For example, if you search by a category other that "Category (All)," then this will filter the database for any subsequent queries until it is cleared or changed to Category (All). If the category picked is Zoonoses, then your query result will not include any diseases in other categories.When the category Zoonoses is picked, then a query with the criteria of fever, fatigue, and pharyngitis will give you only 5 diseases in the differential diagnosis.