DA009 - Electronic Medical Record Review as Source of Follow-Up Data

DA009	Guidelines and Strategies for Electronic Medical Record Review as Source of Follow-Up Data for the TBIMS National Database
Review Committee: Data	Start Date: 10/18/2023
Attachments: None	Last Revised Date: 10/18/2023
Forms: None	Last Reviewed Date: 7/17/2024

Introduction:

The TBIMS uses established procedural steps to collect and code follow-up data for participants enrolled in the TBIMS National Database (NDB).

Purpose:

To demarcate and standardize procedures for collecting follow-up data using the electronic medical record (EMR) as a data source.

Scope:

NIDILRR-funded TBIMS centers and centers designated as TBIMS follow-up centers.

Responsibilities:

TBIMS staff members responsible for collecting and coordinating follow-up data collection (e.g., TBI investigators, study coordinators, research assistants).

Procedural steps:

When to use the EMR to as a data source.
1. The EMR is intended to be a data collection method of last resort after all efforts have been made to locate the participant per SOP DA007.
2. The EMR may be used to populate pre-specified variables within a Form 2 follow-up when that follow-up would otherwise be coded as “lost” but should not be used when the follow-up was refused (unless the participant gives the data collector permission to review the EMR), or when efforts to locate the participant revealed that the participant expired or was incarcerated during the follow-up window. Per SOP DA006, persons under house arrest are considered incarcerated. Per the Data Dictionary, court ordered rehab is considered a form of incarceration.
3. The EMR may be used to populate pre-specified variables within a “followed” but incomplete Form 2 follow-up that contains missing data on the pre-specified variables (e.g., interview interrupted and unable to re-connect with participant to complete the interview). The EMR data must be obtained from records within specified windows of time per variable (see #4 and Table 1, below). In cases where only a few variables are obtained from a participant and substantially more data is available in the EMR but not within the specified time windows relative to the interview, centers may use discretion in recording either the interview data or the EMR data.
Coding the EMR as Primary versus Secondary source of data and method of data collection.
1. If the EMR is the only source of data, code EMR as the primary source and the primary method, and code N/A as the secondary source and method.
2. If multiple sources and methods of follow-up were used (e.g., interview or mailout plus EMR), code the Primary Method and Primary Source as the method/source from which most of the information was collected, per the Data Dictionary.
Identifying the date of follow-up when the EMR is used as a data collection source.
1. If multiple sources and methods of follow-up were used (e.g., interview or mailout plus EMR), the date of the interview or date on the mailout is to be coded as the date of follow-up, regardless of whether the EMR is the primary or secondary source/method.
2. If the EMR is the only source/method of data collection, follow these instructions:
  1. Wait until the last 2 weeks of the follow-up window, after which efforts to contact participants to complete the interview or mailout have likely been exhausted.
  2. Work backwards chronologically from the date of window closing until you find the most recent episode of care that contains information that can be used to populate data on the pre-specified variables appropriate for EMR-based abstraction. The date of the clinical encounter containing codable data becomes the date of follow-up.
Time frame parameters for reviewing ancillary EMR records once date of follow-up has been established.
1. In some instances, data collectors may want to look at multiple clinical encounters (not just the most recent encounter). In those circumstances, EMR data must be taken from records in which the date of the ancillary clinical encounter was within four weeks of the established date of follow-up.
2. There are exceptions to the 4-week time frame parameter. Certain variables (see Table 1) may be abstracted from records outside the 4-week time frame parameter, but they must never originate from records outside the window’s start or end date.
Dealing with multiple records and conflicting or incomplete information.
1. If there are multiple records in the EMR within the appropriate record review time frame that provide different information (e.g., multiple administrations of the PHQ9, GAD7, etc.), choose the record that is closest in time to the date of follow-up and use that record as the source of data.
2. If there are multiple entries in the EMR within the appropriate record review time frame that provide conflicting information and you are not sure which is the best source (e.g., one note says they are married, another note says they are divorced), code as unknown.
3. Do not make assumptions based on incomplete or absent information. For example, if there is no explicit mention of seizures or rehospitalization, this does not mean no seizure or rehospitalization occurred. Unless it is clear that an event did or did not happen, the item should be coded as unknown.
Variables appropriate for EMR-based data collection. The variables considered appropriate for EMR-based data collection are listed below, along with associated guidance (see also Table 1). All other variables (e.g., OSU-TBI-ID, BTACT, NHANES) are inappropriate for EMR-based data collection.

Death: If it is discovered in the EMR that the participant died since last follow-up, then proceed with completing the limited data elements for an expired status per SOP DA006.

Residence [ResF]: Where the participant is living should be obtained from an EMR note within four weeks preceding the follow-up date. The note should specify where they live. Do not assume that a participant lives at home simply because they lived at home at a previous time point, or their spouse resides in a private residence.

Address: For sites that are not permitted by IRB to collect address variables related to the geographic identifiers (GEO-ID), code these variables as authorization not received. Zip code may still be collected. For sites that are permitted by IRB to collect GEO-ID address variables, this data should be obtained from an EMR note within four weeks preceding the follow-up date.

Primary person participant lives with [LiveWhoF]: Who the participant is living should be obtained from an EMR note within four weeks preceding the follow-up date. The note should specify with whom they live. Do not assume that a participant lives with their spouse simply because the participant is married.

Marital status [MarF]: The participant’s marital status should be obtained from an EMR note within four weeks preceding the follow-up date. Do not take from EMR cover sheet/face sheet unless there is a time stamp that clearly denotes that the information is current within 4 weeks of the date of follow-up. The note should specify their current marital status. Do not assume a participant is single (never married) because they live alone. Do not assume a participant is single (never married) if a note specifies that they are currently single unless it is clear they were never married (as opposed to divorced or widowed).

Cultural questions: If this is the first time collecting the Cultural questions, the questions may be abstracted from the EMR at any time point as long as the answers are clearly stated (e.g., EMR clearly states primary language spoken in the home, country of birth, number of years been in the US.) Data collectors may need to compute number of years in the US in relation to recorded follow-up date. For example, if the EMR notes that the participant was living in the US for six years at a timepoint that is two years before the follow-up date, then the data collector would report years in the US as “eight.”

Education [EduYearsF]: The participant’s current educational level should be obtained from an EMR note within four weeks preceding the follow-up date. The language in the note should match the wording of the variable (e.g., obtained a high school diploma… a bachelor’s degree… working toward a bachelor’s degree).

Start date of competitive employment after rehabilitation discharge [EmpFirstF]: Since this variable refers to the first instance that a participant began competitive employment following their index TBI and up until the follow-up interview date, this variable may be abstracted from the EMR at any time point between the date of the most recent follow-up and the date of current follow-up. Per the Data Dictionary, if the exact date is unknown, estimate to the nearest half-month and code the day in the middle of that half month.

Employment status [Emp1F]: The participant’s current employment status should be obtained from an EMR note within four weeks preceding the follow-up date.

Occupational category [OCCF]: Must be from an EMR note within four weeks preceding the follow-up date. Must be sufficient information to code the occupational category per the guidance in the Data Dictionary.

Annual income [EarnF]: The participant’s annual income in the year preceding the follow-up interview should be obtained from an EMR note within four weeks preceding the follow-up date.

Household income [HHIncomeF]: Must be from an EMR note within four weeks preceding the follow-up date. Must be clearly stated as the total family income for the past year, inclusive of the participant’s income and the income of family members living with the participant.

Transportation [TransModeF]: Must be from an EMR note within four weeks preceding the follow-up date.

PART-O family of variables: Must be from an EMR note within four weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview with identical response options.

SWLS family of variables: Must be from an EMR note within four weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview with identical response options.

Rehospitalization family of variables: May be abstracted from an EMR note anytime within the year preceding the follow-up date. Must be clear that the participant was or was not hospitalized anywhere (including outside the internal health care system) to code these variables. If not clear, code as unknown.

Seizures [PastYearSeizF]: May be abstracted from an EMR note anytime within the year preceding the follow-up date. Must be sufficient information on number of seizures that have occurred to code this variable.

Physical measurements [HeightF] and [WeightF]: Must be from an EMR note within four weeks preceding the follow-up date.

General Health [GenHlthF]: Must be from an EMR note within four weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview (“In general would you say your health is…”) with identical response options.

Cigarette smoking [SmkCigF]: Must be from an EMR note within four weeks preceding the follow-up date. Must specify cigarette smoking in same method captured by variable (some days, nearly every day, versus not at all). Do not assume that a mention of current “tobacco” use (without specifying cigarettes) equates to “cigarette” smoking because tobacco may refer to smokeless tobacco.

Drug use family of variables: May be abstracted from an EMR note anytime within the year preceding the follow-up date.

Alcohol use family of variables: Must be from an EMR note within four weeks preceding the follow-up date. Wording must match the wording used in the interview. Do not estimate drinking quantities.

PHQ-9: Must be from an EMR note within two weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview with identical response options.

GAD-7: Must be from an EMR note within two weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview with identical response options.

Suicide attempt [SuicideF]: May be abstracted from an EMR note anytime within the year preceding the follow-up date. Self-injurious behaviors of any severity accompanied by any intent to die meet the definition of attempt. Attempts that are initiated but interrupted (e.g., put gun in mouth, but then put gun away… took pills but then intentionally vomited…) count as an attempt. Suicidal thoughts/ideation without initiation of self-injurious behavior would not meet the definition of suicide attempt. Self-injurious behavior with no intent to die should not be coded as a suicide attempt. Self-injurious behavior for which the intent to die is unclear or undetermined should be coded as unknown.

FIM family of variables: Must be from an EMR note within four weeks preceding the follow-up date. As per the Data Dictionary advice regarding Form 1 data collection, FIM scores may be abstracted from the medical record as long as the notes are specific (e.g., "patient feeding themselves independently"; "patient is unable to ambulate"; "patient needs the assistance of two people for all transfers").

DRS family of variables: Must be from an EMR note within four weeks preceding the follow-up date. Must have been asked in the same manner as in the Form 2 interview (DRS post-acute interview) with identical response options.

GOS-E family of variables: Must be from an EMR note within four weeks preceding the follow-up date. As per the Data Dictionary, data to populate the GOS-E responses may come from various sources as the GOS-E is not intended as a self-perception instrument. Someone who is currently working and driving, for example, is obviously able to follow motor commands. Do not make assumptions for higher level items.

Table 1

Variable	Timeframe Parameter
Death family of variables	Anytime between the prior follow-up date and current follow-up window closing date.
Residence [ResF]	Within 4 weeks of follow-up date.
Address family of variables	Within 4 weeks of follow-up date.
Living with [LivWhoF]	Within 4 weeks of follow-up date.
Marital Status [MarF]	Within 4 weeks of follow-up date.
Cultural questions	Anytime.
Education [EduYearsF]	Within 4 weeks of follow-up date.
Employment status [Emp1F]	Within 4 weeks of follow-up date.
Employment start date [EmpFirstF]	Anytime between most recent follow-up and date of current follow-up.
Occupational category [OCCF]	Within 4 weeks of follow-up date.
Annual income [EarnF]	Within 4 weeks of follow-up date.
Household income [HHIncomeF]	Within 4 weeks of follow-up date.
Transportation [TransModeF]	Within 4 weeks of follow-up date.
PART-O family of variables	Within 4 weeks of follow-up date.
SWLS family of variables	Within 4 weeks of follow-up date.
Rehospitalization family of variables	Up until 1 year from follow-up date.
Seizures [PastYearSeizF]	Within 4 weeks of follow-up date.
Height	Within 4 weeks of follow-up date.
Weight	Within 4 weeks of follow-up date.
General Health [GenHlthF]	Within 4 weeks of follow-up date.
Cigarette smoking [SmkCigF]	Within 4 weeks of follow-up date.
Drug use family of variables	Up until 1 year from follow-up date.
Alcohol family of variables	Within 4 weeks of follow-up date.
PHQ9	Within 2 weeks of follow-up date.
GAD7	Within 2 weeks of follow-up date.
Suicide attempt [SuicideF]	Up until 1 year from follow-up date.
FIM family of variables	Within 4 weeks of follow-up date.
DRS family of variables	Within 4 weeks of follow-up date.
GOS-E family of items	Within 4 weeks of follow-up date.

Training requirements:

Staff persons who are responsible for the Form 2 data collection for TBIMS should be familiar with these criteria. On-going training will be conducted by quarterly data collector teleconferences and in-person data collectors’ meetings.

Compliance:

All follow-up data collectors are required to comply with these guidelines, and attend the quarterly data collector teleconferences and in-person data collectors’ meetings.

References:

N/A

History:

Date	Action
08/5/2023	Creation of SOP
11/30/2023	Approved by Data Committee
7/17/2024	Updated Training and Compliance section, as well as GEO-ID information.

Review Schedule:

At least every 5 years.