Consent to Recontact for Future Research Using Linked Primary Healthcare Data: Outcomes and General Practice Perceptions from the ATHENA COVID-19 Study

Setting

The ATHENA program (Australians Together Health Initiative) is a Queensland Health funded concept involving the integration of primary, secondary and other healthcare datasets across Queensland. The vision is to deliver a state-wide registry containing the healthcare information and biospecimens of several million Australians across Queensland using dynamic consent to connect patients, researchers and the clinical trials industry.

As part of this, the ACV19 study was set up in 2020 in response to the COVID-19 pandemic to create a cohort of all people diagnosed with COVID-19 in Queensland with linked primary, secondary and registry data. There were two parts to the study. Part 1 (completed) linked Queensland COVID-19 hospital and administrative data including notifiable conditions and deaths, did not require informed consent and is described in detail elsewhere.²² Part 2 (described here) links data from the Part 1 study to patients’ healthcare information held within general practices and requires patient consent. All patients who had COVID-19 in Queensland in the period January 1^st, 2020, up until December 31^st, 2020, were contacted to gain consent to have their primary healthcare data extracted from their general practice into a Queensland Health database and linked to other datasets for ethically approved COVID-19 research. Patients were also asked to consent to be re-contacted about possible participation in future COVID-19 related studies.

In Australia, the vast majority of general practices use an electronic medical records system to keep coded records of patients.²³ Multiple vendors exist of which Best Practice and Medical Director are the two most commonly used and have 90-95% of market share.²⁴ The ACV19 study engaged with these two systems specifically because of their innate capacity to export data in XML format. Conversely, systems lacking this functionality, as well as practices unable to export data in XML format, were not eligible for inclusion in the study.’

Patient recruitment

Patients identified as having had COVID-19 from the Queensland Health’s Notifiable Conditions System were telephoned by the ACV19 team to ask permission to email or post an information pack about the study. After 2-5 days the patient was recontacted to answer any further questions and guide them through the patient information consent form. Patients then electronically signed the form (DocuSign) or manually signed-and-returned it by post. Patients nominated up to a maximum of three general practices with whom they were now or previously registered. In Australia, a person can be registered with multiple practices. Patients were informed that their consent was being verbally recorded.

General practice interaction

Prior to patient recruitment, the study was advertised in Primary Health Network newsletters and a website created containing study information, with supporting letters from the RACGP, Chief Health Officer and all seven Primary Health Networks in Queensland.²⁵ An electronic clinical trials management system was used to manage and record all patient and general practice interactions. The liaison team consisted of a general practitioner (lead), two registered nurses, a Primary Health Network practice support officer and an allied health worker. An introductory call to the patient’s general practice checked practice details and confirmed they had one of the two electronic record systems suitable for data transfer. After five days, a second call confirmed information receipt, answered questions and confirmed the patient was registered with the practice. Here, it was explained their patient had consented to participate which required export of the identifiable whole patient file to the Queensland Health. It was explained that no financial reimbursement for data release was available.

Data extraction from general practices

An email was sent to the practice manager’s inbox containing a single-use, secure file, transfer-hyperlink (Kiteworks) a standard method of file transfer used by Queensland Health and appended with the general practitioner’s provider number. No other patient or practitioner identifiers were contained in the message. A formal patient ‘records request’ letter and the patient’s completed patient information and consent form were also sent via secure electronic messaging (Medical Objects) to the nominated inbox within the general practice electronic medical record system. This contained documents pre-tagged with the patient’s name and date of birth to allow direct ingestion into the correct patient file within the system. Medical Objects is a widely used, secure method for information exchange between general practices and Queensland Health. A small minority of practices requested facsimile. The general practice exported the whole patient file to Queensland Health using Kiteworks. If no reply had been received within 5-7 days, the general practice was recontacted to guide them through the process. If the practice had any concerns, they were referred to the general practitioner lead who contacted the practice to resolve any issues. A small minority of practices requested facsimile. Contacting patients and general practices by the study team occurred between January 3^rd, 2021 -August 27^th, 2021. The ACV19 steering committee decided that once file transfer occurred and data was extracted, the file content would be deleted.

Data linkage with Queensland health data

Primary healthcare data were linked probabilistically, using name, date of birth and address by the Statistical Services Branch, to Queensland COVID-19 hospital and administrative data including COVID-19 notifications and deaths from Part 1, applying established protocols.²⁶

Data analyses

General practice geographical area was classified using the Modified Monash Model which defines whether a location is a city, rural, remote or very remote (Modified Monash 1 is a major city, 7 very remote).²⁷ To identify recurring responses and themes from interactions with the general practices, a thematic analysis was conducted by the project lead (KG) using a standard phased approach as a guide.²⁸ Members of the general practice liaison team reviewed and familiarised themselves with the records of their general practice interactions and completed a questionnaire whereby they broadly listed practice types of responses. Based on these responses and known general practice attitudes to sharing data published in the literature, a list of categories was drawn up for focus group discussion by the project lead.^{14, 15} The discussion involved the whole team, was audio recorded, and conducted by the project lead, whereby each category and experience per team member was discussed in detail. The number of general practices who had responded in each category was recorded. The audio recording of the discussion was transcribed verbatim, and the transcripts reviewed over multiple inductive cycles from which broader general themes were developed.²⁸

Using the consent to recontact to recruit to a research study

A research group from University of Queensland studying long-COVID wished to use the recontact process to recruit additional participants from the ATHENA COVID-19 cohort into their ethically approved study. The research group had used social media to advertise nationally and recruited 80 patients. Details of their study can be found elsewhere.²⁹ After protocol approval by the ACV19 steering committee, a proportion of the cohort who had consented to recontact were contacted and provided with an outline of the long-COVID study. Recontacted participants were asked if interested in hearing more about the study and, if so, gave consent to having their contact details passed onto the long COVID study team, who then contacted them to discuss the project in detail.

Patient and public involvement

Between 2017-2019 a proof-of-concept study gauged the willingness of the public to link their healthcare data and received a positive response. One-on-one interviews were also undertaken with members of the public to understand the barriers and enablers to sharing and linking of primary healthcare data for research. The concept was also discussed with key stakeholders, including culturally and linguistically diverse community group leaders, and Australian and Torres Strait Islander representatives, to determine culturally sensitive approaches for recruitment of minority groups. This resulted in the Australian and Torres Strait Islander cohort recruitment being undertaken by a Queensland Health staff member identifying as Australian and Torres Strait Islander.

Ethics approval

The ACV-19 study was granted ethics approval by the Gold Coast Human Research Ethics Committee 24^th April 2020, # HREC/2020/QGC/63555.

Consent to recontact to recruit to a research study

Of those patients who consented to recontact, 416 were contacted to discuss participation in the long-COVID study. Of these, 376(90%) consented to have their contact details passed onto the long-COVID team with 32(8%) declining, and 8(2%) were already enrolled. The long-COVID team approached 365 patients and 192(53%) consented to take part in their study.

Interactions with general practice representatives grouped by theme

Interactions with general practices were grouped into five themes: 1) support for study concept and sharing of health information for research; 2) patient consent to allow healthcare data release; 3) trust, legal and cybersecurity issues; 4) reimbursement, time constraints and technology issues; 5) corporate-owned general practice. Details of the interactions are provided in Table 3 and brief outline is provided here.

Theme 1. Support for study concept and sharing of health information for research

The majority response was “generally speaking, quite positive,” regarding “the idea of the study and contributing to the knowledge of COVID.” Twelve practices felt the study was not worthwhile. Ten thought they had to actively participate with concerns about lack of research experience and insurance cover. Four practices asked about ethics approval, and six about healthcare information storage (interstate or internationally) and all eventually took part. One practice wanted to know who had access to and could use the data.

Theme 2. Patient consent for healthcare data release

Fifteen practices were concerned we had approached the patient and gained consent without the practice knowing or being involved. The perception by the liaison team was that practices believed they should have had the opportunity to advise the patient before we spoke to them. All practices made the eventual decision to release the data but two required further discussion with our general practitioner lead. All practices wanted a copy of the signed consent form before patient file export. The vast majority were comfortable with accepting the electronic consent form and electronic signature as a legitimate form of consent. Several general practices that did not accept electronic consent phoned the patient to confirm that they had signed the consent form.

Theme 3. Trust, legal issues, cybersecurity

The majority of general practices had a high level of trust for Queensland Health to act as custodian of data and manage the research study. One team member stated: “I think it was a key factor that it was a government-led initiative.” Most practices at the start of the study had not heard of the ACV19 project. Nineteen practices thought the team caller was a scam when first contacted. This was managed by directing the general practice staff to the Queensland Health ACV19 website. One practice sought independent legal counsel about participation in the study and then took part. Some practices (n=23) had concerns about exporting the whole patient file to the Queensland Health database as this included non-coded documents such as scanned letters and imaging reports. Once made clear that the whole file would be deleted once the coded information has been extracted, the practices were comfortable with exporting files. Nine practices contacted the patients and recommended against participating. When these patients were contacted as part of the study follow up protocol, six patients changed their mind and participated.

Theme 4. Reimbursement, time constraints and technology issues

Nine practices initially requested payment ranging from $80-$150 to export the data, however eight practices subsequently decided they did not require payment. The team reported that most practices were initially concerned they did not have enough time to export the data: However, once explained that the project team would assist, they agreed to proceed. From a technology perspective, although most practices were familiar with using Medical Objects, most had never used the export file function available in their electronic record system and required training. The majority of practices had not heard of Kiteworks and were concerned about using it. However, once explained it was a standard method used for file transfer by Queensland Health, practices were reassured.

Theme 5. Corporate-owned general practices

Nine practices were part of corporate entities (two) with 12 patients collectively. File transfer needed authorisation, and only one corporate entity eventually allowed practice participation.

Study limitations

A main limitation is that questionnaires were not used to directly assess general practice staff attitudes to sharing primary care data. Rather, we recorded interactions with staff over the course of the study and reported on them. Although all interactions included written documentation of comments made by general practice staff, subjective impressions were made by the liaison team and at risk of interpretation bias. Previous studies focused on the general practitioner’s attitude alone and did not encompass the opinions of all staff working in the practice. This should be considered when comparisons are made. Our study was conducted during the unique situation of a pandemic, and it may be suggested that once normal conditions resume, the results of this study will no longer apply. However, whilst the pandemic has caused a major shift in work patterns, the IT methods used in this study were already in standard use, and any new adaptations such as e-consent, are likely to remain permanent. Finally, 30% of patients and their associated practices did not participate in the study, which may be a source of selection bias. Despite these limitations, this study was able to elicit useful information from Australian general practices regarding their impressions on the export of health data for research. The extraction of general practice patient data was a time-consuming process and not suitable to be undertaken at scale. At the start of the pandemic this process was felt to be the most straightforward to set up, and whilst it has provided valuable information on processes and attitudes of general practices involved, a faster automated version of primary care data extraction is in development.

Conclusions

This study showed that the majority of patients and their general practices consented to share their health information for research and recontact. The consent-to-recontact method proved successful in recruiting to a separate research study. The study also described the methods by which primary healthcare data was successfully extracted from general practices and linked to other health data for the purposes of research using informed consent. This knowledge will inform implementation of the larger ATHENA study, as well as those intending to link primary healthcare data for research and use for recruitment to clinical trials and other studies in future.