Two terms you may hear in discussions of electronic health record data are “inpatient” and “outpatient.” From a computational perspective, these terms are useful to know since they can suggest certain characteristics of the data.
The Definitions
- Inpatient: “in” the hospital
- Outpatient: “out” of the hospital (e.g. family medicine clinic, pediatrics clinic, ambulatory surgical center)
These terms can be used to refer to a type of visit:
“Jane sees her primary care physician” refers to an “outpatient visit” (since the visit takes place in the primary care physician’s clinic)
“Jane has open heart surgery” refers to an “inpatient visit” (takes place in the hospital)
Note that even if Jane’s primary care physician is physically located inside of a hospital building, it is still considered an “outpatient” visit because the unique capabilities of a hospital environment aren’t actually being used for that checkup with her primary care physician.
These terms can also be used to refer to a patient – e.g. “Jane is an outpatient” (when she’s seen by her primary care physician in their clinic.)
General characteristics of inpatient vs. outpatient health record data
| Inpatient | Outpatient | |
| Timing | “Denser”: More data over less time
e.g. A patient may have multiple notes, diagnoses, procedures, vitals, and medications entered into their chart every day by doctors, nurses, physical therapists, pharmacists, and other healthcare providers throughout a hospitalization |
“Sparser”: Less data over more time
e.g. A patient may only see their family physician once every few years – e.g. a note from the physician with some diagnosis/medication updates one day every 3 years |
| Diseases | Sudden worsening of chronic diseases (e.g. hypoglycemic coma in a diabetic patient or heart attack in a patient with previous heart disease)
Sudden onset of more dangerous new diseases (e.g. severe appendicitis, sudden unexplained bleeding, whole-body infections). Treatment for accidents (e.g. serious car crash, industrial accident) It is common for patients to require treatment for more than one medical condition while they are hospitalized |
Long-term management of chronic diseases (e.g. updated insulin prescription for a diabetic patient)
Management of less dangerous “new” diseases (e.g. a sinus infection) Patients seen in outpatient clinics can still have many diseases at the same time, but they are controlled enough that the person doesn’t have to be in the hospital. (Note I didn’t say “well-controlled”…there are many patients seen in outpatient clinics whose chronic diseases may be poorly-controlled for a variety of reasons, but they’re not hospitalized because their condition is stable, e.g. a patient with diabetes whose blood sugar is higher than the goal, but they feel fine for now) |
| Procedures | Procedures/surgeries that require an overnight hospital stay, e.g. coronary artery bypass graft (a type of open-heart surgery) | Procedures/surgeries that do NOT require an overnight hospital stay[1].
All of these are “same-day” procedures, and can range from non-invasive procedures like sewing up small cuts to more invasive procedures like an uncomplicated hernia repair or cholecystectomy (gallbladder removal) |
| Medications | Potentially more frequent medication adjustments, since the patient is sicker and being monitored daily | Less frequent medication adjustments, since the patient is only monitored when they come in for a visit.
Many patients may be taking no medications at all, or only a few. The average number of unique outpatient prescription purchases in 2006 was four per person (for a population of patients with at least one prescription.)[2] Some patients will be taking a lot of medications (e.g. patients sometimes come in toting an entire grocery bag of 20 different pill bottles with meds they’re taking for various conditions; often in these more extreme cases the doctor will pare down the number of meds to a more reasonable level.) |
| Made-up examples of common patients | Patient A is hospitalized for a myocardial infarction (heart attack) and also has diabetes and hypertension (high blood pressure)
Patient B is hospitalized for cancer chemotherapy and also has anemia and stomatitis (painful mouth inflammation) (both of these can be chemo side effects) |
Patient A has a cold (and also happens to have chronic high blood pressure)
Patient B cut their finger when peeling vegetables and needs a couple stitches. Patient C has a stomach ache that is subsequently determined to be a common, temporary viral infection. |
The public MIMIC data[3] is an example of inpatient data, since the data come from hospitalized patients.
A lot of private electronic health record data sets will include both inpatient and outpatient data, which is interesting because that provides a more comprehensive picture of each patient.
I’ll be posting future articles with more details on common inpatient vs. outpatient diseases, procedures, and medications.
Glass Box 
