Planning your project and completing your Data Access Request (DAR) is similar to completing a major grant funding application — it takes time, planning, and meticulous attention to detail. A major obstacle to timely access to data is an inadequate, or incomplete, application.
The Data Steward(s) require an exact description of the data required for a project in order to evaluate a request, and Population Data BC (PopData) needs a detailed description in order to prepare the appropriate data. A properly completed DAR will specify, in detail, the data needed for the research project, including a study population definition and a complete list of data files, fields and years of data.
Requests are assessed according to legislated requirements in the BC Freedom of Information and Protection of Privacy Act (FIPPA), which only allows the disclosure of personal information if certain conditions are met, including: the exact nature of the data being requested, and how it will be used to fulfill the aims of the research project. The level of information in a standard research protocol is not sufficiently detailed to meet these two requirements. The DAR meets those needs and ensures that Researchers are provided with the data they need to fulfill their research project objectives.
The application process requires that Researchers:
- Become familiar with all material outlining access requirements, data holdings, study population definitions, and privacy considerations.
- Ensure that the BC Ministry of Health (MoH) consent form requirements have been met (if applicable).
- Ensure that the ethics and peer review requirements of the Data Steward(s) have been met.
- Submit a completed DAR to PopData.
- Timelines
-
Completion of a DAR, the associated approval process and finally data preparation, can be a lengthy process. While we are working on reducing the timelines associated with an application, sometimes application to data delivery may take over a year. Researchers should be aware of this timeline when working on funding proposals and planning the course of their research. PopData has guides and tips on how to reduce timelines for application processing, and routinely publishes the time frames for recent projects on its website. See application processing.
- Feasibility and cost estimates
-
Please note that PopData strongly recommends that Researchers request a Feasibility and Cost Estimate Letter for the purposes of grant applications. Researchers may request this online. PopData is usually able to provide the letter within 1 to 2 weeks. Please see our fees and charges page, or talk to a Data Access Unit (DAU) staff member for more information.
In order to learn what data is available, you may also request access to Metadata Central, our online data documentation system. In Metadata Central you will be able to view file and field level summary statistics. For more information, including how to get access, please visit https://www.popdata.bc.ca/resources/MetadataCentral.
- Applications involving external data
-
A Researcher may request linkage to data not administered by PopData. There are two likely types of data sources to which a Researcher would seek linkage:
- Data from a public body which is not covered by an existing Information Sharing Agreement with PopData
- Researcher-collected data
Data from a public body which is not covered by an Information Sharing Agreement with PopData
For linking to data from a public body, the review process would be substantially similar to that for data already managed by PopData, with Data Stewards performing their adjudication function, and the Data Access Unit coordinating the application and review process. PopData would additionally require an Information Sharing Agreement which may be limited to the single research project in order to have data transmitted to PopData.
Researcher-collected data
For linkage to Researcher-collected data, there will be additional requirements for review:
- Compliance with relevant legislation(s) for Researcher-collected data must be considered prior to application. For example, data originally collected from a physician who is not employed by a public body, will need to be in compliance with the Personal Information Protection Act (PIPA), in addition to meeting the requirements of FIPPA, upon approval from the public body for data linkage.
- An assessment of privacy risks related to linkage of Researcher-collected data to additional data. The Researcher will need to describe the potential privacy risks in performing the proposed linkage. This will be done through an appendix to the Data Access Request which includes items common in a Privacy Impact Assessment.
- Informed consent or its waiver: the Researcher will need to provide evidence of written, informed consent to link the Researcher-collected data for the specified research purposes. If consent has not been obtained, the Researcher will need to provide an explanation and a justification for why consent should be viewed as impracticable. For more information, visit our Consent form requirements page.
It is recommended that Researchers submit the aforementioned additional requirements to PopData prior to data collection or pursuit of the research project, so that PopData, together with the Data Stewards, can ensure the relevant information is collected according to a suitable process for subsequent Data Access Requests.
If consents and relevant legislation allows, and the data are of sufficient interest, in the above cases PopData may choose to initiate discussions with the public body or the Researcher in order to make these data available for access by other Researchers via PopData. This would involve negotiation of Information Sharing Agreement(s).
- Study populations and extracts
-
A study population is the group of subjects that the Researcher wishes to include in their analyses. The subjects usually share at least one common characteristic, which may be as simple as age, or more complex, such as the set of people over the age of 65 with a certain health condition. For example, a cohort may be defined as:
- All residents of BC in calendar year 1999
- All residents of BC aged 65 years and over in calendar year 1999
- All residents of BC aged 65 years and over diagnosed with Type 2 diabetes in calendar year 1999
- All residents of BC aged 65 years and over diagnosed with Type 2 diabetes who also receive treatment for kidney failure in 1999
The study population is not the same as the data extract. An extract is the actual data requested for a study population that will help to answer a research question. For example, a researcher may be interested in the number and average length of hospital stays of all individuals in a study population, or in their use of physician services over the preceding two years.
A project may also include more than one study population for comparative purposes, similar to cohort and control groups.
Defining the study population
In the DAR, the Researcher is requested to provide specific and detailed study population information in addition to a rationale or justification for the study population(s) requested.
Data Access Unit staff at Population Data BC (PopData) work with Researchers to define the study population for a project. This interaction is necessary to ensure the desired study population can be technically produced by PopData programming staff (or external programmers, depending on the details of the project). In addition to the text description, a technical definition is also required, but will be produced in most cases by Data Access Unit staff as a result of this interaction.
For example, a Researcher might want to look at the use of health services by adolescents over a certain time period. The specifics of this could be very simple, for example, all individuals in BC aged between 13 and 18 in calendar year 2000. A more complex example might be a study population defined as all heart attack patients who received bypass surgery in calendar year 1995. The specifics required here are:
- How is a heart attack defined?
- If isolating the condition requires diagnostic codes in the hospital separations file, what are the specific codes?
- Does the code need to appear as the first (primary) diagnosis?
- How is bypass surgery defined?
- If it requires procedure codes, what are the specific codes?
- If it requires the use of MSP fee item codes, what are those fee item codes?
Data Access Unit staff at PopData then work with the Researcher to complete the technical definition for the study population(s), seeking confirmation with the Researcher. This support is also built into the process for the DAR which requires that the technical definition be completed by PopData prior to DAR submission. The Data Steward(s) use both the text and the technical definitions in their adjudication.
Data extracts
Researchers will download and complete the appropriate data file checklist(s) and select the fields to be used for analysis.
If a project involves more than one study population each may require a different set of data to be extracted. In this case, separate data extract checklist(s) for each study population may be completed, as applicable.
It is important to remember that the Researcher needs to request the data for their extract, and not the data required to construct a study population, when filling out the data file checklists.
- How much data to request
-
Researchers can request as much data as required to complete a research project. However, the more complex a request, the longer it will take for the Data Steward(s) to review and approve it, the longer it will take Population Data BC (PopData) staff to prepare data, and the higher the cost.
According to the principles of the BC Freedom of Information and Protection of Privacy Act, the Data Access Request (DAR) has been designed to help Researchers determine the minimum amount of data necessary to complete their research. The DAR specifically addresses what research data are needed, given a particular cohort definition.
Where there is some ambivalence about the design of a cohort, Researchers are encouraged to request something reasonably broad so that they can test the effect that different assumptions have on the composition of their study group.
The selection of data fields in the extract, however, must be restrictive. There is no reason, for example, to request a data field for physician numbers if it is clear that physician-specific analysis is not required to answer the research question.
