HOPS Attach 6 - Data Submission doc. examp..pdf

NA-ACCORD Data Submissions Document (P2018)

Version 10.2

April 15, 2019

Table of Contents

Changes

Contact Information

Selecting Patients

1.1 Definitions

1.2 NA-ACCORD Inclusion Criteria

1.3 NA-ACCORD Exclusion Criteria

2 Transmitting Data

2.1 Overview

2.2 Procedures for Transmitting Data

2.3 Data Formats

2.4 NA-ACCORD Data Tables

2.5 Coding Data

2.6 Coding Diagnosis Data Source

2.7 Coding Medication Data Source

2.8 Coding Laboratory data source

3 Patient Record

3.1 Patient Record Format

3.2 Coding for Sex Variables

3.3 Coding for Race

3.4 Coding for Hispanic

3.5 Coding for Risk Factor for HIV Transmission

3.6 Enrollment Date

3.7 Last Activity Date

3.8 Death Date

3.9 Coding Source of Death Date

3.10 Sub-site Code

3.11 HIV Negative Patients

3.12 BirthCountry

3.13 Transgendered

3.14 Sample Patient Data Submission

4 Diagnosis Record

4.1 Format

4.2 Submission of codes

4.3 Diagnosis table format

4.4 Sample Diagnosis Data Submission

5 Laboratory Test Results Record

5.1 CD4/Viral Load at Intake

5.2 Viral Load Above and Below levels of quantification

5.3 Normal Min/Max

5.4 Submission of Codes

5.5 Laboratory table fields

5.6 Sample Laboratory Data Submission (note that encounterID and cohortRecordID were not submitted)

6 Medication Table Record

6.1 Data should be as complete as possible

6.2 Historical ARVs

6.3 Submission of Codes

6.4 Sample Medication Data Submission

7 Genotypic Resistance Records

7.1 Coding Rules

7.2 Genotypic Test Results Table Fields

7.3 Sample Genotypic Resistance Data Submission

8 Phenotypic Resistance Records

8.1 Coding Rules

8.2 Phenotypic Test Results Table Fields

8.3 Sample Phenotypic Resistance Data Submission

9 Cause of Death Record

9.1 Mortality Table Fields

9.2 Coding for Cause of Death

9.3 Definitions

9.4 Coding for Causes of Death (‘Type’ Classification)

9.5 Coding Cause of Death Source

9.6 Submitting NDI+ Underlying Cause of Death

9.7 Sample Mortality Table Submission with Cause of Death from Multiple Sources

9.8 Examples of classifying causes of death

10 Geographic Data [geographic Table]

10.1 Send two-character abbreviation for State/Province

10.2 Submission of postal code information [postCode field]:

10.3 Date associated with residence information

10.4 Clinic location as proxy for State/Province

10.5 Homeless patients

10.6 Geographic Table Fields

10.7 Sample Geographic Data Submission

11 Encounter Information [encounter table]

11.1 Encounter Table Record (for HIV Primary Care visits and Interviews/Visits for Interval

Cohorts)

12 Insurance Information [insurance table]

12.1 Insurance Table Record

13 Inpatient Stays [hospitalization table]

13.1 encounterID Field

14 Substance Use Survey Information [substanceSurvey Table]

14.1 Question Field

14.2 Response Field

14.3 Response Date

14.4 Question Label Field

14.5 Response Label Field

15 Procedures [procedure Table]

15.1 Procedure Table Format

15.2 Coding Procedure data source

15.3 Source field codes

16 Discharge Diagnosis Data [DischargeDx table]

16.1 Submission of codes

16.2 Assigning Discharge Diagnosis Date

16.3 EncounterId Field

16.4 Discharge Diagnosis table format

16.5 Sample Discharge Diagnosis Data Submission

Appendix 1

Table of NA-ACCORD Cohorts with Potential Patient Duplication

Appendix 2

Standard 2-character U.S. State and Canadian Province abbreviations

Changes

1. Look for yellow highlighting to indicate changes or important information in the body of the submissions document and in the Standard Codes spreadsheet.

2. Ideally, we would like to receive data from your cohort through December 2018 by June 28, 2019. Please contact me if you anticipate that this timeline will be a problem for your cohort or if your cohort will be unable to submit data through December 2018.

3. Sex Coding: The sex variable has been split into two new variables: Birth Sex and Present Sex

(See Section 3.2 of the Submissions Document). The DMC recognizes that most Hospital

Administrative systems capture the patient’s Present Sex, but if available, cohorts are requested to submit both Birth and Present Sex which will help to more accurately capture patient sex, especially for Transgender patients.

4. Race and Hispanic Coding: The raceEth variable has been split into two separate variables in order to more accurately capture Race and Hispanic ethnicity (See Sections 3.3 and 3.4 of the

Submissions Document).

5. Medications: Consult the “NA-ACCORD P2018 Standard Codes (Ver 10.2) - April 2019.xlsx” spreadsheet for new medications requested for the P2018 submission.

a. Many cohorts provided opioid and opioid treatment data on an ad hoc basis with their

P2017 data. For the P2018 submission, these medications are requested from all cohorts.

6. Discharge diagnoses are only requested from UCSD, JHHCC, KPMAS, KPNC, SAC, and UW. Please be sure to submit an encounterID in both the Inpatient and DischargeDx tables that can be used to link discharge diagnoses to their corresponding hospitalization.

7. Removal of duplicate patient reminder: Cohorts with patients that might appear in other NA-

ACCORD cohorts should coordinate with those cohorts to ensure that duplicate patient data are not submitted; see Section 2.5.2 for instructions and Appendix 1 for a list of overlapping cohorts.

Contact Information If you have questions about this document, please use the following contact information:

Stephen Van Rompaey, PhD NA-ACCORD Data Integration Lead kelpie@uw.edu 206-744-3675

Liz Morton NA-ACCORD Data Manager lmorton@uw.edu 206-744-3683

Justin McReynolds NA-ACCORD Data Systems Administrator mcjustin@uw.edu mailto:kelpie@uw.edu mailto:lmorton@uw.edu mailto:mcjustin@uw.edu

Selecting Patients Use the following information to identify patients in your cohort who are eligible for enrollment in the

NA-ACCORD collaboration. For efficiency, this submissions document can be used by those cohorts that are participating selectively in NA-ACCORD aims and only submitting a subset of data for specific studies.

1.1 Definitions

1.1.1 Clinical cohort

Populations of HIV-infected patients in routine care at academic medical centers and community-based facilities that deliver HIV primary and specialty care, whose data are primarily based upon medical records as part of patient care.

1.1.2 Interval cohort

Populations of HIV-infected patients prospectively enrolled in cohort studies, whose data are derived independently from medical care, although may be confirmed by clinical records.

1.1.3 NA-ACCORD cohorts by cohort type

Clinical Cohorts

• CWRU

• Fenway

• HIVRN

• HOPS

• JHHCC

• Kaiser Permanente North California

• Kaiser Permanente Mid-Atlantic States

• Montreal

• OHTN

• RRC (Puerto Rico)

• SAC

• UA-Birmingham

• UCSD

• UNC-Chapel Hill

• UW-Seattle

• UBC-Vancouver

• VACS/VACS8

• Vanderbilt

Interval Cohorts

• ALIVE

• ALLRT/AACTG

• MACS

• WHIS

• SCOPE

1.1.4 PT0

The date the patient entered the cohort.

• For clinical cohorts this is the date that the patient initiated care at your clinic defined as the initial visit date. (See Section 1.2.3.1)

• For interval cohorts this is the date that the patient was enrolled in your study defined as the enrollment date. (See Section 1.2.3.2)

1.1.5 Historical data

This refers to HIV-related clinical and treatment activity that the patient experienced before (PT0) they initiated care or enrolled in your cohort

1.1.6 Current data

Clinical and treatment activity that the patient experienced on or after (PT0) the date they initiated care or were enrolled in your cohort.

1.1.7 Lost to follow-up

These are patients who are no longer under observation by your cohort. For clinical cohorts, these are patients no longer receiving care at your affiliated clinic(s) because they have moved, changed provider or care setting, or for an unknown reason. For interval cohorts, these are patients who have dropped out of your study or are no longer appearing for regular interviews. Patients who have not had a primary care visit (clinical cohorts) or had an interview (interval cohorts) within 18 months are defined as lost to follow-up.

1.1.8 Study end date

Cohorts should send the most recent data available on each patient. The NA-ACCORD Epi/Biostat core will determine the appropriate censor date for each analysis.

1.2 NA-ACCORD Inclusion Criteria

1.2.1 Confirmed HIV infection

As evidenced by documented positive HIV antibody test, detectable HIV-1 RNA level, clinical event such as an AIDS-defining illness, or confirmation of positive HIV antibody test performed outside the system.

1.2.2 PT0

The date of the patient’s initial visit or enrollment must be known.

1.2.3 Patient follow-up

1.2.3.1 Clinical cohorts

The patient’s initial visit date (PT0) is defined as the first primary care visit at your clinic that was followed by a second primary care visit within 12 months. For clinical cohorts a patient must have at least two primary care visits within 12 months at some point during their care to be enrolled in the NA-

ACCORD. If a patient only has a single primary care visit or if they never have two visits within 12 months they are excluded from the NA-ACCORD project. The inclusion criterion is intended to exclude patients not receiving routine clinical care at a clinic.

1.2.3.2 Interval cohorts

The patient must have at least one follow-up interview within 12 months of the patient’s initial enrollment date. For interval cohorts a patient must have at least an enrollment interview and one follow-up interview. Patients with only an enrollment interview are excluded from the NA-ACCORD project.

1.2.4 Patient’s sex must be known.

1.2.5 Patient’s year of birth must be known.

1.2.6 Historical data must be available for the patient

The cohort must have collected clinical activity on the patient that occurred prior to PT0 that includes all

AIDS-defining conditions (required), all antiretroviral medications (required), and all CD4 and HIV-1 RNA measurements (preferred).

1.2.7 Patient’s vital status must be known

The vital status (active, lost to follow-up, deceased) must be known as of their last visit or interview date.

1.2.8 Patients must be uniquely identified

Patients must not be duplicated in other cohorts (see Section 2.5.1).

1.2.9 Patients must be 18 years of age

Patient must be 18 years at time of enrollment into a cohort. In rare case, a patient will be under care at a clinic for pediatric HIV care and then continue under care as an adult at the same clinic. In these cases, the patient may be included in the NA-ACCORD project, but their enrollment date should be coded as the first visit date following their 18th birthday.

1.2.10 Cohort inception date (Patients enrolled in clinical cohorts only)

Clinical cohorts will only submit patients who are enrolled in care at their clinic on or after their cohort inception date. The cohort inception date is the date that complete clinical data are captured at your clinic in an electronic health record (EHR); this typically means that all demographic, diagnosis, medication, and laboratory test result data are captured electronically on patients. Most NA-ACCORD cohorts have already discussed this issue with Dr. Mari Kitahata and their cohort inception date has been established. If you have questions about your cohort inception date, please contact the NA-

ACCORD Data Management Core for clarification.

1.3 NA-ACCORD Exclusion Criteria

1.3.1 Transgendered, transsexual, or intersexed patients are NO LONGER excluded from

NA-ACCORD data submissions.

2 Transmitting Data

2.1 Overview

Cohorts will submit their data using secure data transfer procedures and standard data formatting.

Please contact Justin McReynolds to coordinate your data transfer or if you have questions about the process (mcjustin@uw.edu).

2.2 Procedures for Transmitting Data

The technical team supports secure web transport for cohort record uploads. This method provides secure transmission of cohort data over the Internet and provides isolation of individual cohort data once transmitted to the NA-ACCORD.

2.2.1 Secure web upload (HTTPS)

Cohorts navigate to a secure web site, login using cohorts-specific user names and passwords, and upload data to the secure web server. These logins are person specific and are not to be shared.

2.3 Data Formats

2.3.1 ASCII files

Cohorts will create a separate CSV file for each of the data tables listed below. Column headers should be present and all values enclosed with double quotations. For the P2015 data submission, the designated naming convention needs to be followed: P2015_TableName.csv. Files will automatically be stamped with a timestamp corresponding to the time of data submission.

2.3.2 Access database template

The Data Management Core (DMC) will provide an Access database that consists of preformatted data tables. A cohort can extract data from their local information system and write it into the appropriate data table. No specific file name is required for the Access Database; however, sites should not rename the tables within the database template. The database will automatically be stamped with a timestamp corresponding to the time of data submission.

2.4 NA-ACCORD Data Tables

This section describes coding procedures for key data elements.

2.5 Coding Data

2.5.1 Assigning unique record identifiers

The NA-ACCORD DMC will use the following two elements in order to uniquely identify the records within the repository: cohortPatientId, and cohortRecordId. The cohortRecordId is a unique row mailto:mcjustin@uw.edu identifier that comes from the source system. In most cases, this will be an auto-increment primary key column from a representative table within the site’s EHR (it is not a HIPAA violation to send this record identifier from your source system to NA-ACCORD). The cohortRecordId is a string value in order to allow for sites that use GUIDs as primary key identifiers on records rather than integers.

Submitting cohortRecordID’s is optional.

2.5.2 Assigning patient identification numbers

Cohorts will de-identify their data and assign their own unique identifier to each patient. Some cohorts will retain the link between the patient’s identifiers and the study ID, while other cohorts will only submit anonymous data.

Some contributing cohorts may have patients who are also members of another NA-ACCORD cohort. In

Appendix 1 we present a table that shows which cohorts may experience this problem. Please examine

Appendix 1 and determine if your cohort is listed. If so, examine rows of the table and use the red ‘X” to identify which other cohort you may share patient information. It is your responsibility to coordinate with that cohort to determine which patients will be sent from your cohort. The columns in the table have been arranged hierarchically; a cohort listed on the left has priority over cohorts to its right in the table for submitting a given patient.

2.5.3 Coding dates

It is critical that each patient possess sufficient information concerning the date at which clinical events occurred to clearly specify the sequence of a patient’s clinical course and outcomes.

Dates will vary in the precision with which they are collected. Some cohorts may collect dates using only

‘month’ precision while other cohorts collect ‘day’ precision. Frequently, historical data (activities that occurred prior to initiating care or study enrollment) are only available as less precise information because they are patient-reported and did not occur while under observation by your cohort.

The following data coding formats are to be used to indicate date precision and when an ‘unknown’ date occurred relative to the patient’s PT0. Please note that we are using “00” to indicate that the month and/or day of a clinical event is unknown.

Precision Meaning Format Example

Day This event occurred on March 16, yyyy-mm-dd 1998-03-16

Month This event occurred in March 1998.

The day of the month is unknown.

yyyy-mm-00 1998-03-00

Year This event occurred in 1998. Month and day are unknown.

yyyy-00-00 1998-00-00 historical unknown This event occurred sometime prior to the date the patient initiated care or was enrolled in a study cohort.

1900-00-00 1900-00-00 current unknown

The event occurred sometime during the period that the patient

1955-00-00 1955-00-00 was under care or enrolled in a study cohort.

Medication duration known, but historical date unknown

The medication was taken sometime prior to the date the patient initiated care or was enrolled in a study cohort, but the duration is known. Use the year ‘0000’ to indicate the start date for this type of medication (See section

2.5.4.1 for instructions on how to show duration.)

0000-00-00 0000-00-00

2.5.4 Coding dates for medication durations

In the event that a cohort knows that a patient received a specific drug for a known duration, but the actual start and stop dates are unknown, then the cohort should use the following guidelines.

2.5.4.1 Medications prior to patients PT0

If the medication was received prior to the patient’s PT0 then the cohort can assign a surrogate start and stop date prior to PT0 for the appropriate duration. For example, if a patient reports that they received

AZT for one year prior to their enrollment date of January 19, 2001, then the cohort could assign a start date of 0000-00-00 and a stop date of 0001-00-00 to indicate the one year duration of the historical medication.

2.5.4.2 Missing start or stop dates while under care or enrolled

Missing start and stop dates for medications that occurred while the patient was under care or enrolled in your cohort (on or after PT0) represent a more difficult problem.

• Interval cohorts should apply the coding rules they use to capture medication activity during any survey period. If your interval cohort does not have coding rules for this situation, then you should code both the start and stop date as ‘1900-00-00’.

• Clinical cohorts may be unable to assign surrogate start and stop dates and these patients may be excluded from the analysis. If the start and stop date is unknown, then you should code both the start and stop date as ‘1900-00-00’.

2.6 Coding Diagnosis Data Source

Diagnoses, medication, and laboratory data must have a data source qualifier that indicates where the data element was obtained. Cohorts must classify each diagnosis according to its source. These codes indicate increasing accuracy of the diagnosis, ranging from an unknown data source to a diagnosis confirmed by medical record review.

2.6.1 Unknown

The code “Unknown” may be used if a cohort possesses no source information on the diagnosis. It is possible that historical circumstances at your clinic resulted in clinical diagnoses within your electronic data system whose source is unknown. In principle, any diagnosis with an unknown source should be examined through medical record review.

2.6.2 Patient reported without supporting outside documentation

This source code should be used for patient reported diagnoses pertaining to care the patient received outside your clinic/study for which there is no supporting outside medical record documentation. This typically occurs when the patient initiates care at a clinic or enrolls in a study cohort and relies entirely on their memory to report an historical diagnosis without outside medical record documentation to confirm the diagnosis. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.6.3 Outside documentation

This should be used for diagnoses that are documented in outside medical records pertaining to care the patient received outside your clinic/study. This typically occurs when a patient initiates care at a clinic and or enrolls in a study cohort and review of outside medical records confirms a diagnosis the patient had in the past or has at the time they are enrolled. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.6.4 Data collected by NA-ACCORD cohort

This code is intended to capture diagnoses generated by your cohort through routine data collection.

2.6.4.1 Clinical cohorts

The “Data collected by NA-ACCORD cohort” code applies to diagnoses documented by a clinician in the electronic health record (EHR) or related system (includes diagnoses directly entered into the EHR by a clinician or clinician diagnoses entered into the EHR by a designated data entry person).

2.6.4.2 Interval cohorts

The “Data collected by NA-ACCORD cohort” code applies to diagnoses obtained during the standard interview process used by interval cohorts.

2.6.5 Verified clinical diagnosis

The code “Verified clinical diagnosis” is intended to capture diagnoses that were verified through medical record review. Cohorts often verify particular diagnoses through ongoing routine medical record review.

Diagnosis Source Codes Data Source Labels Comments

D1 Source unknown How the diagnosis information on the patient was obtained is unknown.

D2 Patient reported without supporting outside documentation

For clinical condition(s) or treatment(s) not experienced by the patient in your clinic/study, for which there is no supporting outside medical documentation for the clinical condition(s) and treatment(s) reported.

D3 Outside documentation For clinical condition(s) or treatment(s) not experienced by the patient in your clinic/study that are documented in outside medical records, lab reports, etc.

D4 Data collected by NA-ACCORD cohort

Use for all clinical data generated at your clinic (EHR-based) or during patient interviews for interval cohorts.

D5 Verified clinical diagnosis A diagnosis verified by medical record review; typically, an ongoing routine review process implemented by your cohort.

2.7 Coding Medication Data Source

Like diagnosis data, medication data has similar source and reliability issues.

2.7.1 Unknown

The code “Unknown” may be used if a cohort possesses no source information on the medication. It is possible that historical circumstances at your clinic resulted in medications within your electronic data system whose source is unknown. In principle, any medication with an unknown source should be examined through medical record review.

2.7.2 Patient reported without supporting outside documentation

This source code should be used for patient reported medications pertaining to care the patient received outside your clinic/study for which there is no supporting outside medical record documentation. This typically occurs when the patient initiates care at a clinic or enrolls in a study cohort and relies entirely on their memory to report an historical medication without outside medical record documentation to confirm the medication. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.7.3 Outside documentation

This should be used for medications that are documented in outside medical records pertaining to care the patient received outside your clinic/study. This typically occurs when a patient initiates care at a clinic and or enrolls in a study cohort and review of outside medical records confirms a medication the patient had in the past or has at the time they are enrolled. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.7.4 Data collected by NA-ACCORD cohort

This code is intended to capture medications generated by your cohort through routine data collection.

2.7.4.1 Clinical cohorts

The “Data collected by NA-ACCORD cohort” code applies to medications documented by a clinician in the electronic health record (EHR) or related system (includes medications directly entered into the

EHR by a clinician or clinician medications entered into the EHR by a designated data entry person).

2.7.4.2 Interval cohorts

The “Data collected by NA-ACCORD cohort” code applies to medications obtained during the standard interview process used by interval cohorts.

2.7.5 Pharmacy dispensing records

The code “Pharmacy dispensing records” is intended to capture courses of therapy that have been computed from data received from a pharmacy dispensing system.

Medication Source Codes

Data Source Labels Comments

M1 Source unknown How the medication information on the patient was obtained is unknown.

M2 Patient reported without supporting outside documentation

For medication(s) not experienced by the patient in your clinic/study, for which there is no supporting outside medical documentation for medication(s) reported.

M3 Outside documentation For medication(s) not experienced by the patient in your clinic/study that are documented in outside medical records, lab reports, etc.

M4 Data collected by NA-ACCORD cohort.

Use for all clinical data generated at your clinic (EHR-based) or during patient interviews for interval cohorts.

M5 Pharmacy dispensing records Course of therapy was derived from pharmacy dispensing data.

M6 Clinical Trial Use this source if the medication record was obtained from clinical trial records.

2.8 Coding Laboratory data source

2.8.1 Unknown

The code “Unknown” may be used if a cohort possesses no source information on the laboratory test result. It is possible that historical circumstances at your clinic resulted in laboratory test results within your electronic data system whose source is unknown. In principle, any laboratory test result with an unknown source should be examined through medical record review.

2.8.2 Patient reported without supporting outside documentation

This source code should be used for patient reported laboratory test results pertaining to care the patient received outside your clinic/study for which there is no supporting outside medical record documentation. This typically occurs when the patient initiates care at a clinic or enrolls in a study cohort and relies entirely on their memory to report a historical laboratory test without outside medical record documentation to confirm the test results. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.8.3 Reported in outside documentation

This should be used for laboratory test results that are documented in outside medical records pertaining to care the patient received outside your clinic/study. This typically occurs when a patient initiates care at a clinic and or enrolls in a study cohort and review of outside medical records confirms a laboratory test the patient had in the past or has at the time they are enrolled. This can also occur if a patient is lost to follow-up and returns to care at your clinic or resumes participation in an interval cohort.

2.8.4 Data collected by NA-ACCORD cohort

This code is intended to capture laboratory test results generated by your cohort through routine data collection.

2.8.4.1 Clinical cohorts

The “Data collected by NA-ACCORD cohort” code applies to laboratory test results from an Electronic

Health Record (EHR) that typically originate in a separate Laboratory Medicine system.

2.8.4.2 Interval cohorts

The “Data collected by NA-ACCORD cohort” code applies to laboratory test results obtained during the standard interview process used by interval cohorts.

Laboratory Source Codes Data Source Labels Comments

L1 Source unknown How the laboratory information on the patient was obtained is unknown.

L2 Patient reported without supporting outside documentation

For laboratory test result(s) not experienced by the patient in your clinic/study, for which there is no supporting outside medical documentation for test result(s) reported.

L3 Outside documentation For laboratory test result(s) not experienced by the patient in your clinic/study that are documented in outside medical records, lab reports, etc.

L4 Data collected by NA-ACCORD cohort.

Use for all clinical data generated at your clinic (EHR-based) or during patient interviews for interval cohorts.

3 Patient Record The patient record will capture information on both the patient’s demographic characteristics and vital status. Transmitting a patient’s race/ethnicity is optional although cohorts should send all available information.

3.1 Patient Record Format

Use the following format to send demographic data on your study subjects:

Required fields are bolded.

Field Description Data Type Example cohortPatientId De-identified patient identification code string birthSex*

Sex at birth string Female Male Intersexed presentSex* Present sex string Female Male Intersexed birthYear

Year of patient’s birth yyyy 1981 race Patient’s self-identified race string White Black hispanic Patient’s self-identified Hispanic origin string Yes No risk1 Patient risk factor for HIV transmission string Male Sex with Male risk2 Additional risk factor of HIV transmission string Injection Drug Use enrollDate Date the patient initiated care or enrolled in your cohort date 2001-07-13 lastActivityDate The date of the last recorded patient activity date 2006-02-28 deathDate

Patient’s death date date 2006-03-19 deathDateSource

Source of patient’s death date information string Clinic reported

SSDI

subSiteID ID indicator for sub-sites within cohorts string 003 hivNegative Indicates that the patient is a member of an HIV negative cohort.

string Y

BirthCountry Country of patient’s birth. string Somalia transgendered Patient is identified as transgendered String Yes, No cohortRecordID Unique record identifier A10000001 any

* Either Birth Sex or Present Sex is required.

3.2 Coding for Sex Variables

Sites that do not collect “Birth Sex” should report patient’s sex using the “Present Sex” field

Field Code birthSex Male birthSex Female birthSex Intersexed presentSex Male presentSex Female presentSex Intersexed

3.2.1 Intersexed Individuals

The sex code “Intersexed” is available for sites that collect this information, but there is currently no requirement that sites classify patients as intersexed.

• Do not confuse the intersexed condition with transgender patients who are transitioning from one sex to another.

• Intersexed is a group of conditions where there is a discrepancy between the external genitals and the internal genitals (the testes and ovaries). Consult the following website for more information on the intersexed condition: http://www.isna.org/faq/what_is_intersex

3.3 Coding for Race

Cohorts have used different approaches for capturing race and ethnicity information for patients. In both Canada and the United States, ethnicity is collected for patients who self-identify as Hispanic. In some cohorts, race and ethnicity are combined, while in other cohorts they have been separated.

Submit the most precise race code from the following table. Sometimes racial groups have been combined, such as “Asian/Pacific Islander” and historically, race coding has included a Hispanic category. If the only race code available for a patient is Hispanic then you will follow the rules in Section

3.5 for coding Hispanic ethnicity. (This is an optional field)

Field Code Description race American Indian American Indian or Alaska Native - a person having origins in any of the original peoples of North, Central, and South America, and who maintains tribal affiliation or community attachment.

race Asian Asian – a person having origins in any of the original peoples of the Far East, Southeast Asia, or the Indian subcontinent including, for example, Cambodia, China, http://www.isna.org/faq/what_is_intersex

India, Japan, Korea, Malaysia, Pakistan, the Philippine

Islands, Thailand, and Vietnam.

race Black Black or African American - a person having origins in any of the black racial groups of Africa.

race Pacific Islander Pacific Islander - Pacific Islander or Native Hawaiian, a person having origins in any of the original peoples of

Hawaii, Guam, Samoa, or other Pacific Islands.

race White White – a person having origins in any of the original peoples of Europe, the Middle East, or North Africa.

race Asian/Pacific Islander Asian/Pacific Islander - For sites that used a combined category that captured individuals of Asian, Hawaiian, or Other Pacific Islands origin.

race Multiracial Identifies with multiple racial groups.

race Other Other.

3.4 Coding for Hispanic

For sites that directly collect Hispanic ethnicity on patients, code patients “Yes” or “No” for

“Hispanic” (This is an optional field). For those sites that have combined Hispanic within their Race categories, the following rules will apply:

• If the individual at your site has a Race category other than Hispanic, then that individual will be unknown for Hispanic ethnicity and no Hispanic data will be sent on this patient.

• If the individual has a Race category of Hispanic, then:

• The individual will be assigned “Yes” for Hispanic ethnicity, and

• The individual will be unknown for Race and no race data will be sent on this patient.

hispanic Yes Hispanic or Latino/Latina hispanic No Not Hispanic or Latino/Latina

3.5 Coding for Risk Factor for HIV Transmission

Risk factor for HIV transmission will be used primarily to identify patients who have a history of injection drug use (IDU). If you have risk factor information for a patient, then you should insure that any risk factor that indicates injection drug use is submitted. You may submit up to two risk factors if multiple risk factors are collected in your cohort.

3.6 Enrollment Date

The date the patient initiated care or was enrolled in your cohort (PT0). (If a patient has a primary care visit and is lost to follow up, but later has two primary care visits within 12 months, then the second primary care visit will be defined as the patient’s PT0 and activity prior to this date will be treated as historical data.) (See Section 1.2.3)

3.7 Last Activity Date

Submit the date of the last recorded patient activity.

3.7.1 For clinical cohorts

Determine the patient’s last clinical activity using the following types of activity: diagnosis, laboratory test result, medication dispense/start date, primary/specialty care visit.

3.7.1.1 For clinical cohorts that conduct ongoing medical record review

Some clinical cohorts conduct ongoing medical record review to confirm the accuracy of antiretroviral medication data and diagnoses, such as AIDS-defining illnesses. This ongoing record review of patient data may occur at intervals scheduled weekly, monthly, or as long as six months. As a result, record-review validated medication and diagnosis data may lag behind data available for laboratory test results, encounters, mortality, etc. If a cohort submits all available data on these patients, then patients will not have observations across all data types for the same period of time, which creates potential problems for analyses using these data and may lead to inaccurate inferences on associations between clinical indicators.

In order to provide the most accurate data for analysis, clinical cohorts that have ongoing medical record review protocols should assign the last medical record review date for a patient as their ‘last activity date’ and then truncate data across all data types at this date for this patient.

3.7.2 For interval cohorts

Use the last date that the patient was interviewed for your study.

3.8 Death Date

Patient’s death date should be submitted even if it occurs after the patient was no longer under care at your site or disenrolled in your study. (The Biostatistical Core will determine when to right censor patients and whether or not it is appropriate to use a patient’s death date in an analysis.)

3.9 Coding Source of Death Date

The codes listed below capture various levels of certainty and information about a patient’s death. A

“Death Certificate” should be considered the most reliable source of information on a patient’s death date. Death certificates are especially relevant for those patients who do not have a Social Security

Number (US), such as foreign nationals residing within the U.S. without documentation. This may also be a problem for Canadian cohorts.

National or Provincial Registries would follow the death certificate in reliability regarding a patient’s date of death. In the US patients sometimes use fraudulent Social Security Numbers and death information may be incorrectly reported to the US Social Security Administration, although this is less likely to be a problem among Canadian cohorts. In the US, there are three national level indices for death data:

• National Death Index – Data compiled by the US Centers for Disease Control from State death certificates that list death date.

• National Death Index Plus – Data compiled by the US Centers for Disease Control from State death certificates that list death date and cause of death.

• Social Security Death Index – Data compiled by the US Social Security Administration that lists death date.

Use “Clinic reported” for those patients whose death was reported to the clinic and have not been verified by a death certificate or National/Provincial death index.

If your cohort uses another source of death date verification, then submit ‘Other’ as the death date source and provide the DMC with a description.

Use the following coding to submit source of death date. You may submit either the numeric or text value from the valid codes list.

Field Name Valid Codes Country deathDateSource 1 = Clinic Reported 2 = Death certificate

3 = NDI

4 = NDI+

5 = Provincial Death Index

6 = SSDI

7 = Other (specify) 9 = Unknown

Both Both

US

US

Canada

US

Both

3.10 Sub-site Code

The subSiteID field should be used by NA-ACCORD cohorts that are comprised of multiple sites (such as

HIVRN, HOPS, and OHTN).

3.11 HIV Negative Patients

Use the hivNegative field to indicate that the patient is a member of an HIV Negative cohort (use ‘Y’ to in this field to indicate that the patient is HIV negative, use ‘S’ to indicate that the patient has sero-converted).

3.12 BirthCountry

Use the BirthCountry field to indicate the country of the patient’s birth. In some cases you may compute this if you know the Province or State of the patient’s birth.

3.13 Transgendered

Patients meeting any one of the following conditions should be classified as transgendered. (This is an optional field) transgendered Yes Yes, if:

a. The patient self identifies as transgendered; OR

b. Patients whose gender identity, expression, or behavior is not traditionally associated with their birth sex and pursue gender expression (masculine or feminine) through external self-presentation and behavior (may include undergoing cosmetic procedures or taking hormones); OR

c. Transsexual individuals who seek some degree of sex reassignment surgery, take hormones, or undergo other cosmetic procedures; OR

d. Anyone who feels gender dysphoria, whether or not they meet criteria for gender identity disorder, such that this dysphoria interferes with their social or occupational functioning.

transgendered No Does not meet conditions for “Yes”.

3.14 Sample Patient Data Submission

Cohort Patient ID birth Sex present Sex

Birth Year

Race Hisp risk1 risk2 enrollDate lastActivity Date

DeathDate deathDate Source subsite ID Hiv Negative Birth Country

Trans gendered

66559901 Male 1980 Asian No Male sex male Injection Drug Use 1999-06-15 2005-07-19 003 US 77369912 Female 1986 Black No Heterosexual 2004-12-01 2006-03-15 021 Somalia Yes 11450929 Male 1972 White Yes Male sex male 1995-04-19 2002-07-07 2002-08-01 Death certificate 034 Y

4 Diagnosis Record

4.1 Format

The implementation of ICD-10 coding across NA-ACCORD cohorts has required the DMC to change our approach to collecting diagnosis data. The new approach uses a combination of submitting standard NA-

ACCORD diagnosis codes for AIDS-defining illnesses, ICD-9 & ICD-10 codes, and ad hoc diagnosis codes that were created for specific NA-ACCORD sites to accommodate a local idiosyncrasy for a small number of diagnoses. Use the following format to send both historical diagnosis data and diagnoses collected in your clinical system or during routine interviews for interval cohorts.

4.1.1 AIDS-Defining Conditions:

1) NA-ACCORD cohorts will continue to send AIDS-defining illnesses (ADIs) using the NA-

ACCORD Standard Codes for these diagnoses. Please note that “PNEUMONIA” and

“SALMONELLA SEPTICEMIA” have been dropped from the list of ADIs and Pneumonia

(bacterial, viral, ventilator associated, and unspecified) are now captured through ICD-9 and

ICD-10 codes. The DMC assumes that sites have review protocols in place to capture both

“PNEUMONIA RECURRENT” and “SALMONELLA SEPTICEMIA RECURRENT” diagnoses. See

“ADI” tab of the attached “P2015 Standard Codes (7.2) - June 2016.xlsx” spreadsheet.

2) In general, NA-ACCORD cohorts are expected to have an ongoing medical review process to verify ADIs.

3) As a service to NA-ACCORD sites the DMC has created a list of ICD-10 codes that can be used to ascertain potential new ADIs as part of your ongoing medical review process. This list can be found in the “ADI ICD-10 Codes” tab of the attached “P2015 Standard Codes (7.2) - June

2016.xlsx” spreadsheet.

4) For some patients, you may have historical data that only indicates that the patient had

AIDS prior to initiating care or enrolling in your cohort, with no information about which

AIDS-defining condition was diagnosed. In this situation, submit the label ‘AIDS defining diagnosis unspecified’. You should use the date that indicates the earliest that this information was known by your cohort. Frequently, this will be the date they initiated care or enrolled in your study.

4.1.2 Submitting ICD-9 and ICD-10 Codes for non-ADI Diagnoses

We are requesting that you send all of the ICD-9 and ICD10 codes listed in the “ICD-9” and “ICD-

10” tabs of the attached “NA-ACCORD Standard Codes (7.2) - June 2016.xlsx” spreadsheet.

1) If your cohort submitted NA-ACCORD Standard Diagnosis Codes for non-ADI diagnoses previously, for the P2015 data submission DO NOT map the ICD-9/10 codes to NA-

ACCORD Standard Diagnosis codes as they will be centrally mapped by the DMC to the appropriate diagnosis code.

2) If you feel that the DMC has omitted relevant ICD-9/10 codes, please submit them and the DMC will determine whether or not they need to be added to our list.

4.1.3 Ad hoc Diagnosis Codes

Some of the NA-ACCORD Standard Diagnosis Codes were created to accommodate an idiosyncratic local diagnosis code for a specific site. If this was done for your site, then continue to submit those NA-ACCORD Standard Diagnosis code(s) as you have done in the past.

4.2 Submission of codes

Submit codes used by your cohort for diagnoses, based on the rules described above.

• NA-ACCORD cohorts are reminded to send all instances of each diagnosis code and not just the one with the earliest diagnosis date; and DO NOT send “rule-out” diagnoses.

• Please continue to send diagnosis data on your patients from all available sources, including both inpatient and outpatient diagnoses.

4.3 Diagnosis table format

Required fields are bolded.

Field Description Example Datatype cohortPatientId De-identified patient identification code string diagnosis The label used by your cohort to specify an AIDS-defining condition.

PCP

string diagnosisDate Diagnosis date of the AIDS-defining condition. Apply date rules in Section 2.5.3 for fuzzy or unknown dates.

YYYY-MM-DD date source Use Section 2.6 to define source of diagnosis.

D1, D2, etc. string encounterID Unique sequence number that links clinical events associated with a record in the hospitalization Table

1234567890 string cohortRecordID Unique record identifier A10000001 any

4.4 Sample Diagnosis Data Submission

cohortPatientId diagnosis diagnosis_date dischargeDx source

345699 PCP 1999-05-16 D2

345699 MAC 2001-08-31 D4

345699 Toxo 2004-06-17 Y D4 345717 Hypertension 2001-11-18 D4 345717 Diabetes II 2003-01-12 D5

In this case, the cohort reports that the patient had a diagnosis of ‘Pneumocystis jiroveci pneumonia

(PCP)’ prior to initiating care at the clinic (historical diagnosis) that was patient-reported without outside supporting documentation. Later the patient had a diagnosis of ‘MAC or M kansasii disseminated or extrapulmonary’ that was diagnoses diagnosed at the NA-ACCORD participating clinic by a provider and

‘Toxoplasmosis of brain’ that was associated with a hospitalization (they did not submit data for the encounterID and cohortRecordID fields).

Another patient has received a diagnosis of Hypertension and Diabetes Mellitus Type 2 while enrolled in the cohort.

5 Laboratory Test Results Record Use the following format to send both historical laboratory test results and current laboratory test results.

5.1 CD4/Viral Load at Intake

In principle, at intake the patient should be asked if they know their CD4 cell count nadir and their baseline viral load (or set point). If you have a historical CD4 cell count that is explicitly labeled "nadir" in your system then you would put "nadir" in the "interpretation" field to distinguish it from other historical CD4 cell count data you might have. The same approach applies for viral load. If you explicitly capture the patient's baseline viral load at intake then you can put "baseline" in the interpretation field to distinguish it from other historical viral load data.

5.2 Viral Load Above and Below levels of quantification

We are asking sites to explicitly interpret their viral load test results and code them as ‘BLQ’ (Below the level of quantification – undetectable) or ‘ALQ’ (Above the level of quantification) in the ‘interpretation’ field when appropriate.

5.3 Normal Min/Max

Included in the laboratory test result table are fields for the expected normal minimum and maximum result value for each laboratory test. Increasingly, the NA-ACCORD is using test result values to characterize the clinical status of patients and it is important that the normal minimum and maximum be submitted when available. Note that if a normal minimum and maximum are not provided, we might ask for that information at a later date.

5.3.1 Clinical cohorts

For clinical cohorts that receive their laboratory test results directly from their laboratory medicine system, this should be straightforward to obtain. If you are not receiving this information electronically then you are not expected to provide it.

5.3.2 Interval cohorts

Unless this information is already stored in your data system, you will not be expected to submit it.

5.4 Submission of Codes

Submit codes used by your cohort for laboratory tests, based on the rules described above. These would be the codes you provided Liz Morton when you completed your data survey spreadsheet. We will map them to NA-ACCORD standard codes centrally.

5.5 Laboratory table fields

Required fields are bolded.

Field Name Brief Description Examples Datatype cohortPatientId De-identified patient identification code string testName Name of the laboratory test. Albumin string result

Result 4.5 string units Units that test are measured in at your cohort.

g/dl, mMol. etc. string normalMin Normal expected minimum result value

3.5 string normalMax Normal expected maximum result value

5.2 string interpretation Use to indicate an interpretation for the test result if appropriate. (Also used to indicate CD4 cell count nadir and baseline HIV- 1 RNA for historical data)

Hi, Low, Normal, Abnormal, nadir, baseline, BLQ, ALQ resultDate

Local result date. See Section 2.2 date source Source of the laboratory test L4 string

5.6 Sample Laboratory Data Submission (note that encounterID and cohortRecordID were not submitted) cohortPatientId testName result units normalMin normalMax interpretation resultDate source

345699 Abs CD4 305 copies/uL .773 2.25 nadir 1999-12-22 L3 345699 Albumin 4.5 g/dL 3.5 5.2 2000-10-08 L3 345699 Abs CD4 0.276 THOU/uL .773 2.25 2000-01-31 L4

345699 HIV-1 RNA RT-

PCR Ultra 2500 copies/mL <50 >75,000 2000-01-31 L4

345699 Abs CD4 0.402 THOU/uL .773 2.25 2001-04-15 L4

345699 HIV-1 RNA RT-

PCR Ultra <50 copies/mL <50 >75,000 BLQ 2001-04-15 L4

345699 Hemoglobin 14.4 g/dL 13.0 18.0 2001-05-07 L4

6 Medication Table Record Use the following format to send both historical and current medication data. Complete historical and current medication data on antiretroviral (ARV) medications is required for all study subjects; when the specific medication is not known then several less precise coding options are described below. For non-

ARV medications complete data is expected while the patient is under care at your clinic, but there is no requirement for historical data on non-ARV medications.

6.1 Data should be as complete as possible

It is essential that the most complete data possible are collected on antiretroviral medications. Use the rules listed in sections 2.5.3 and 2.5.4 to submit start and stop dates for these medications.

6.2 Historical ARVs

For some patients, you may have historical data that only indicates that the patient had prior experience with an antiretroviral mediation before initiating care or enrolling in your cohort and no information about which medication they were treated with. You should use the date that indicates the earliest that this information was known by your cohort. Frequently, this will be the date they initiated care or enrolled in your study. In this situation, use the following rules:

6.2.1 Hx ARV Treatment, Med unknown

If you know that they patient was treated with an antiretroviral medication, but have no information about the class, then use the code ‘Hx ARV Treatment, Med unknown’.

6.2.2 Hx PI Treatment, Med unknown

If you know that they patient was treated with a Protease Inhibitor (PI), but have no information about the specific medication, then use the code ‘Hx PI Treatment, Med unknown’.

6.2.3 Hx NNRTII Treatment, Med unknown

If you know that they patient was treated with a Non-nucleoside Reverse Transcriptase Inhibitor

(NNRTI), but have no information about the specific medication, then use the code ‘Hx NNRTII

Treatment, Med unknown’.

6.2.4 Hx NRTII Treatment, Med unknown

If you know that they patient was treated with a Nucleoside Reverse Transcriptase Inhibitor (NRTI), but have no information about the specific medication, then use the code ‘Hx NRTII Treatment, Med unknown’.

6.2.5 Hx EI Treatment, Med unknown

If you know that they patient was treated with an HIV Entry Inhibitor, but have no information about the specific medication, then use the code ‘Hx EI Treatment, Med unknown’.

6.2.6 Hx Fusion Treatment, Med unknown

If you know that they patient was treated with an HIV Fusion Inhibitor, but…