## Observation Statistics

OPERATION: Observation Statistics

The official URL for this operation definition is:

`http://hl7.org/fhir/OperationDefinition/Observation-stats`

The Statistics operation performs a set of statistical calculations on a set of clinical measurements such as a blood pressure as stored on the server. This operation is focused on Observation resources with valueQuantity elements that have UCUM unit codes.

The set of Observations is defined by 4 parameters:

• the subject of the observations for which the statistics are being generated ( `subject`)
• which observations to generate statistics for ( `code` and `system`, or `coding`)
• the time period over which to generate statistics `'duration` or `period`)
• the set of statistical analyses to return ( `statistic`)

Possible statistical analyses (see StatisticsCode):

• average ("Average"): The mean of N measurements over the stated period
• maximum ("Maximum"): The maximum value of N measurements over the stated period
• minimum ("Minimum"): The minimum value of N measurements over the stated period
• count ("Count"): The [number] of valid measurements over the stated period that contributed to the other statistical outputs
• totalcount ("Total Count"): The total [number] of valid measurements over the stated period, including observations that were ignored because they did not contain valid result values
• median ("Median"): The median of N measurements over the stated period
• std-dev ("Standard Deviation"): The standard deviation of N measurements over the stated period
• sum ("Sum"): The sum of N measurements over the stated period
• variance ("Variance"): The variance of N measurements over the stated period
• 20-percent ("20th Percentile"): The 20th Percentile of N measurements over the stated period
• 80-percent ("80th Percentile"): The 80th Percentile of N measurements over the stated period
• 4-lower ("Lower Quartile"): The lower Quartile Boundary of N measurements over the stated period
• 4-upper ("Upper Quartile"): The upper Quartile Boundary of N measurements over the stated period
• 4-dev ("Quartile Deviation"): The difference between the upper and lower Quartiles is called the Interquartile range. (IQR = Q3-Q1) Quartile deviation or Semi-interquartile range is one-half the difference between the first and the third quartiles.
• 5-1 ("1st Quintile"): The lowest of four values that divide the N measurements into a frequency distribution of five classes with each containing one fifth of the total population
• 5-2 ("2nd Quintile"): The second of four values that divide the N measurements into a frequency distribution of five classes with each containing one fifth of the total population
• 5-3 ("3rd Quintile"): The third of four values that divide the N measurements into a frequency distribution of five classes with each containing one fifth of the total population
• 5-4 ("4th Quintile"): The fourth of four values that divide the N measurements into a frequency distribution of five classes with each containing one fifth of the total population
• skew ("Skew"): Skewness is a measure of the asymmetry of the probability distribution of a real-valued random variable about its mean. The skewness value can be positive or negative, or even undefined. Source: Wikipedia
• kurtosis ("Kurtosis"): Kurtosis is a measure of the "tailedness" of the probability distribution of a real-valued random variable. Source: Wikipedia
• regression ("Regression"): Linear regression is an approach for modeling two-dimensional sample points with one independent variable and one dependent variable (conventionally, the x and y coordinates in a Cartesian coordinate system) and finds a linear function (a non-vertical straight line) that, as accurately as possible, predicts the dependent variable values as a function of the independent variables. Source: Wikipedia This Statistic code will return both a gradient and an intercept value.

If successful, the operation returns an Observation resource for each code with the results of the statistical calculations as component value pairs where the component code = the statistical code. The Observation also contains the input parameters `patient`, `code` and `duration` parameters. If unsuccessful, an OperationOutcome with an error message will be returned.

The client can request that all the observations on which the statistics are based be returned as well, using the include parameter. If an include parameter is specified, a limit may also be specified; the sources observations are subsetted at the server's discretion if count > limit. This functionality is included with the intent of supporting graphical presentation

URL: [base]/Observation/\$stats

Parameters

 Use Name Cardinality Type Binding Documentation IN subject 1..1 uri The subject of the relevant Observations, which has the value of the Observation.subject.reference. E.g. 'Patient/123'. Reference can be to an absolute URL, but servers only perform stats on their own observations IN code 0..* string The test code(s) upon which the statistics are being performed. Provide along with a system, or as a coding. For example, the code = "45667-2" will evaluate all relevant Observations with this code in Observation.code. For LOINC codes that are panels, (e.g. 55284-4) this means to include all the individual codes that are part of a panel IN system 0..1 uri The system for the code(s). Or provide a coding instead IN coding 0..* Coding The test code upon which the statistics are being performed, as a Coding IN duration 0..1 decimal The time period of interest given as hours. For example, the duration = "1" represents the last hour - the time period from on hour ago to now IN period 0..1 Period The time period over which the calculations to be performed, if a duration is not provided IN statistic 1..* code average|max|min|count The statistical operations to be performed on the relevant operations. Multiple statistics operations can be specified. These codes are defined here IN include 0..1 boolean Whether to return the observations on which the statistics are based IN limit 0..1 positiveInt If an include parameter is specified, a limit may also be specified to limit the number of source Observations returned. If the include paramter is absent or equal to "false" the limit parameter SHALL be ignored by the server OUT return 1..* Observation A set of observations, one observation for each code, each containing one component for each statistic. The Observation.component.code contains the statistic, and is relative to the Observation.code and cannot be interpreted independently. The Observation will also contain a subject, effectivePeriod, and code reflecting the input parameters. The status is fixed to `final`. OUT source 0..* Observation Source observations on which the statistics are based

This operation cannot be performed on observations that the user is not authorized to see. It is assumed that the server has identified and secured the context appropriately, and can either associate the authorization context with a single patient, or determine whether the context has the rights to the nominated patient, if there is one. If there is no nominated patient (e.g. the operation is invoked at the system level) and the context is not associated with a single patient record, then the server should return an error. Specifying the relationship between the context, a user and patient records is outside the scope of this specification.