Difficult transitions are simply part of life and unfortunately it is often impossible to anticipate and prepare for consequent changes.  In my experience the conversion from a paper to an electronic medical record (EMR) certainly qualifies as one of the sea changes in health care, and while I anticipated electronic order communication would bring a variety of challenges I never saw a particular issue related to laboratory results reporting coming.  To help frame the issue it is helpful to consider some of the advantages of the paper record that some of us took for granted, e.g., inherent contextual information.  The paper record provided an obvious context for information largely because source documents with their visual cues as to their origin and relative reliability were always available even if a document(s) was misfiled in the record, i.e., included in an inappropriate tabbed section of the physical chart.  With regard to laboratory data it was readily apparent whether one was reviewing laboratory reports from the local hospital laboratory or from a reference laboratory when presented with telltale document headers and footers, a logo, etc.  Laboratory observations entered and subsequently found in a physician’s or nurse’s note or a letter that accompanied a patient admission or referral was readily understood to be a secondary or tertiary source and may or may not have included sufficient metadata with the laboratory observation, e.g., units of measure, to allow appropriate interpretation of the observation. 

While the EMR is not without contextual cues the facility to display information stored in a clinical repository and in a variety of contexts unrelated to the origin and source of the information creates the potential to misinterpret the meaning, reliability and/or significance of the information.   Let me illustrate my point with an example that came to my attention while reviewing a console-like utility designed to enhance a physician’s ability to organize and manage information in an efficient manner.  The utility provides a concise console or dashboard-like view of various observations including diagnostic laboratory data, e.g., observation organized by trimester for use by an obstetrician.  The console supports manual entry (free-text) of laboratory observations as well as displaying data received from a primary laboratory via an interface.  As currently designed the console does not provide for or inherently require entry of all metadata appropriate to many laboratory observations.  Reporting laboratory, ordering provider and date can be specified in the console header but the date field in the header cannot be configured as a mandatory field.  However, the date of collection can be required for each individual observation entered via the console.   (The console can also be used to place orders for a laboratory test(s) to be referred to a primary source laboratory; a convenience that creates a problem.)   Unfortunately laboratory observations manually entered via the console generate an electronic order and file the data in the EMR repository thus making it possible to display secondary (e.g., data transcribed from paper laboratory records presented at admission, transfer or referral) or even tertiary (e.g., data transcribed from letters containing lab observations or information gathered as part of a patient history) source data in the same context as primary source data (CLIA licensed, accredited, interfaced lab results approved by local Pathologist ).    Date and performing laboratory will display with the laboratory observation if entered via the utility; ‘Unknown Lab’ would be displayed with the result(s) if the source lab is not specified.  

The greatest risk for misinterpretation of manually entered numerical laboratory observations is in the Results View utility of this EMR (spreadsheet-like view).  Entry of a laboratory observation from Lab B in the field defined for Lab A will result in the data from Lab B appearing in the same row as results from Lab A.  Even if the methodology between the laboratories is similar it is still possible that units of measure may differ resulting in a potentially significant misrepresentation of the information.  Hopefully careful review of results in light of the clinical status of the patient would alert the provider to the discrepancy.

Additionally, laboratory observations that result in complex, multipart reports (e.g., results from genetic testing) can be defined in the console but due to space and other issues a free-text entry would be limited to an abstracted version of the report.  While this last issue may not arise if the utility is not pushed beyond its design limits (a console-like view for organizing succinct data) the fact that idiosyncratic abstractions of complex reports could be displayed in results view and chart view is disturbing.

Proposed Next Steps
In light of the foregoing comments it would be helpful to delineate and adopt industry guidelines for entry and presentation of laboratory observations in the EMR.  The following is suggested as a classification scheme and guideline to begin a dialog among medical informatics professionals in Pathology and Medicine and with purveyors of EMR products.

The full integrity of laboratory observations should be maintained in the electronic medical record (EMR) through the inclusion of date of collection, units of measure, appropriate reference ranges and identification of the reporting laboratory with each observation.  Laboratory observations should be classified and presented in a class appropriate context to provide an indication of the relative reliability of the observations (classification defined below). 

Primary Source Data:  Laboratory observations transmitted via interface or provided on paper that originate in the performing laboratory and are readily traceable to the source lab, i.e., transmitted via lab-specific interface, a copy of a formatted report in the style of the reporting laboratory that contains key patient identifiers, test description, date and time of observation, units of measure, appropriate reference ranges, an interpretation where appropriate and information defining the limitations and specificity of the test methodology.   

Secondary Source Data:  Transcribed information taken from a primary source document, correspondence or other documents containing laboratory observations that is not the original or a copy of the original laboratory report prepared and distributed by the reporting laboratory. 

Tertiary Source Data:  Verbally transmitted data regarding a laboratory observation(s) provide by the patient, patient family member or other individuals with such knowledge during a formal or informal conversation, e.g., elicitation of Medical History. 

Using this or another relevant classification system it should be feasible to by policy or design or both to specify an appropriate context for each class of laboratory observation.  

Is this author’s opinion that laboratory observations presented in a chart view and results view context (spread-sheet like view) should be restricted to the presentation of information from a Primary Source.  Restricting these contexts to Primary Source Data helps maintain a consistent level of data integrity for the physician consulting the EMR. 

Notes transcribed by the physician or other authorized clinical staff, or other formats for organizing and recording data, e.g., console-like forms may contain Primary, Secondary or Tertiary Source data describing prior laboratory observations as long as the data is restricted to display in the original electronic context and only that context.  In general the manual entry of laboratory observations should always include sufficient contextual information and relevant metadata to allow appropriate interpretation of the observation by a provider viewing the record.  An exception to a strict application of the classification and proposed policy might be possible if manually entered laboratory observations with appropriate metadata were displayed in other electronics contexts if these entries were sufficiently differentiated from Primary Source data.

Hopefully these observations and suggested guidelines and classifications will stimulate a dialog and will ultimately result in the establishment of industry practices or even standards that will help maintain the integrity of medical information managed by an EMR.