Atherosclerosis Risk in Communities (ARIC) Cohort

General Instructions

The patient’s address is a confidential data item abstracted from the hospital chart and recorded on the ARIC Confidential Data Form. Addresses will be converted to a code defined by longitude and latitude that will then be used under procedures that safeguard the confidentiality of this information, for two primary purposes. One is to provide quality assurance for the event eligibility based on the (coded) place of residence, and the other is to include aggregate information of the area of residence of the patient for the analyses of the ARIC data.

Each of these analytic steps uses geocoded data as opposed to the actual address. To avoid error in geocoding the addresses however, a complete and accurate transcription of the patient address is needed. The purpose of these question-by-question instructions is to make it easier for the ARIC personnel to accomplish this goal.

Hospital records almost uniformly provide the patient’s address and the address of a guarantor. Often the addresses of an immediate relative, a proxy, or contact also are found on the record. The immediate relative is often identified as the guarantor, unless it is “self.” The location of this information in the hospital record is somewhat variable between hospitals, as is the format.

Our goal is to transcribe the address that corresponds to the place of residence of the patient. Only if the patient’s address is not available or incomplete should the address of the relative or that of the guarantor (if different) be transcribed. Even if a full address for the patient is not available, city-state-zip, state-zip, and city-zip combinations are valuable for ARIC to have (especially when zip is available in the 5+4 format).

Thus, the address of a proxy should only be abstracted if a city-state-zip, state-zip, or city-zip combination is not available for the patient.

Some scenarios to consider are shown below; they illustrate the use of the address entry panel on the form (see the corresponding examples of completed forms in the Appendix).

Special reference needs to be made to abbreviations, since these are frequently used in addresses and often reflect idiosyncratic rather than standard abbreviations. A look-up table of street and unit name abbreviations is provided in the Data Entry System to help abstractors record the standard U.S.P.S. abbreviations that are used for geocoding. This look-up table also translates commonly used (but non-standard) abbreviations into the U.S.P.S. abbreviations we need. This table is used as other look-up tables in the DES: typing the initial letters of the name or abbreviation should place the cursor over the appropriate row in the table (if the name / abbreviation is included in the table). Once the abstractor has confirmed that the appropriate line is highlighted, double-clicking the mouse or pressing the ‘Enter’ key will then import the approved abbreviation into the data field on the form.

Numbered highway addresses can be challenging to abstract and require careful attention. Consider “1098 US Hwy 15-501 S”. It should be entered in three fields as follows: [4.B.3] Number = 1098, [4.B.6] Name = US HWY 15 501, and [4.B.8] Name Suffix = S. Please note that with this address, [4.B.6] Name may contain numbers (in this case, 15 501), and that [4.B.7] Name Type remains blank (although we know 15 501 is a highway). In other words, [4.B.6] Name is best defined in such cases by including both components (“US HWY” and “15 501”) in this single field.

In closing, we also want to mention that you will occasionally encounter addresses that seem to defy attempts to divide them without making assumptions. In our experience, such addresses are thankfully uncommon.

This post was last modified on December 11, 2024 6:06 am