Data sources

 

HIV Datamart

PHOL conducts centralized HIV diagnostic and viral load (VL) testing for Ontario, and maintains databases that contain information on the vast majority of such testing in the province. These databases were integrated to form the HIV Datamart. In the datamart, a person’s diagnostic and VL test records are linked using patient identifiers. However, it is not possible to link non-nominal HIV-positive diagnostic tests (coded, anonymous) to VL tests, as no identifying information is available to facilitate linkage.

All information in the HIV Datamart is confidential, and only de-identified aggregate data is shared with OHESI partners.

The HIV diagnosis and VL databases used to create the datamart are described in further detail below:

HIV diagnosis database (1985 to 2015)

The HIV diagnosis database contains records for all individuals who have had an HIV-negative and/or HIV-positive diagnostic test result in Ontario. HIV-positive diagnoses include people who were diagnosed with HIV for the first time in Ontario, as well as people who were diagnosed elsewhere and migrated to Ontario and tested again. PHOL conducts the vast majority of HIV testing in Ontario. The database does not include information on tests conducted by other laboratories for the purposes of life insurance and screening blood donations and organ/tissue donors. In Ontario, individuals testing for HIV may use either their full name (nominal) or a code assigned in specific primary care settings (coded), or test without a code or name at designated HIV testing clinics (anonymous). Coded and anonymous testing are both forms of non-nominal testing.

HIV viral load database (1996 to 2015)

VL testing was implemented in 1996 and the database at PHOL contains records for all individuals who have had a VL test in Ontario. In addition to VL test results, the database contains information from the VL test requisition form (completed by the provider), including most recent CD4 count and whether the patient is on ART at the time of testing. Providers complete the information on ART on approximately 80% of VL test requisition forms. All VL tests in the database were conducted nominally.

 

Ontario HIV Laboratory Cohort

A cohort of people with diagnosed HIV living in Ontario, referred to as the Ontario HIV Laboratory Cohort, was created using the integrated HIV Datamart. Individuals are included in the cohort if they have either;

  1. a nominal HIV-positive diagnostic test (1985 to 2015), and/or
  2. at least one viral load test (1996 to 2015)

Non-nominal HIV-positive diagnostic tests (i.e., tests conducted anonymously or using coded identifiers) are excluded from the cohort. The lack of identifying information on non-nominal tests means that it is not possible to link these tests to subsequent VL tests and monitor engagement in HIV care. However, individuals diagnosed non-nominally are included in the cohort when they receive a nominal HIV diagnostic or VL test at entrance to care.

Individuals with record of a VL test only (no linked nominal HIV-positive diagnostic test) are included in the cohort, with one exception. Individuals with no nominal HIV-positive diagnostic test and all undetectable VL test results do not enter the cohort if they have evidence of being HIV-negative (i.e. record of a nominal HIV-negative diagnostic test after, on the same day as, or within 30 days before their last undetectable VL test); these individuals are likely HIV-negative people receiving a VL test for diagnostic purposes.

Overall (1985 to 2015), the HIV Datamart includes 40,372 HIV-positive diagnostic test records and 23,851 individuals with record of at least 1 VL test. Of the HIV-positive diagnostic tests, 18,683 (46.3%) are non-nominal and were excluded from the cohort (note: the proportion of HIV-positive diagnostic tests that are non-nominal has decreased over time, from 50% in 1996 to 15% in 2015). A further 947 individuals were excluded because they had no nominal HIV-positive diagnostic test, all undetectable VL tests, and evidence of being HIV-negative. A total of 29,587 people in the HIV Datamart have record of a nominal HIV-positive diagnostic test and/or at least one VL test and were included in the cohort.

3. and have not been administratively lost to follow-up after two years.

After entering the cohort, individuals are removed if they have no record of a VL test for more than two consecutive years, and no VL test at a later date (referred to as administratively LTFU). These individuals are removed to account for potential death or migration out of the province. If an LTFU individual has a subsequent VL test in later years, they re-enter the cohort and are counted as being a diagnosed person living in Ontario during the years in which they were in a gap in care. The two-year LTFU criteria was selected given that this would include most individuals who are in a known gap in care (i.e. individuals with no VL test in one or more consecutive years, but a VL test in later years). Between 2000 and 2014, the median duration of known gaps in care was 1.4 years. After application of the two-year LTFU criteria, there were 16,110 diagnosed people living with HIV in the cohort at the end of 2015.

Cohort individuals are included in the analysis of annual cascade indicators until administratively LTFU. For example, if a diagnosed individual had a VL test in 2011, but no test for the next four years, they are included in the analysis for two years after their last VL test (ie. 2012 and 2013), but not afterwards (i.e. 2014 and 2015). However, if this individual re-engages in care in the future (e.g. 2018) and has another VL test, they would be included in the analysis for all the years in which they were in a gap in care (i.e. 2012 to 2017). During this gap in care, it is assumed that the individual is not in care, not on ART, and not virally suppressed.

Note: In sensitivity analyses, non-nominal diagnoses were included in the cohort and the LTFU criteria extended to three years.
 

Newly HIV diagnosed sample

The Ontario HIV Laboratory Cohort was used to create a sample of individuals who were diagnosed with HIV for the first time in Ontario (ie. were not individuals who were diagnosed outside the province and migrated to Ontario and tested again). This sample is used to measure time from HIV diagnosis to linkage to care and viral suppression. Individuals are defined as newly diagnosed if they have record of a nominal HIV-positive diagnostic test and no evidence of being previously diagnosed (ie. no detectable VL test or CD4 count before their diagnosis date, first VL after diagnosis not suppressed). Unlike the inclusion criteria for the Ontario HIV Laboratory Cohort, individuals with a VL test only (no linked HIV-positive diagnostic test) are not included in the newly diagnosed sample, as a diagnosis date is needed to measure time to care and suppression.

Overall, 13,410 individuals had a nominal HIV-positive diagnostic test between 1996 and 2015. Of these, 3,797 (28.3%) were excluded because they had evidence of being previously diagnosed. The remaining 9,613 are included in the newly diagnosed sample.

The number of newly diagnosed individuals included in the analyses of time to linkage to care and viral suppression is likely an underestimate of the actual number of newly diagnosed individuals in the province. The newly diagnosed sample excludes non-nominal HIV-positive diagnostic tests and does not include individuals with record of a VL test only (no linked nominal HIV-positive diagnostic test). This means individuals diagnosed non-nominally are not included in these analyses unless they received a nominal HIV diagnostic test at entrance to care.