Primary Care Physicians’ and Patients’ Perspectives on Equity and Health Security of Infectious Disease Digital Surveillance

Recruitment and Data Collection

We recruited primary care physicians from March through June 2021 via mass WeChat groups of national professional organizations such as China General Practice Teachers’ Association and the Medical Prevention Society General Practice section, which link hundreds of primary care physicians across the country. A research team member advertised the study through the WeChat groups, answered queries, and referred potential participants to the study team. The research team consisted of academics focused on ID, sexual minorities, and aging. Most are registered primary care physicians and nurses, as well as experts in artificial intelligence and information technology. Primary care physicians were recruited purposively across sex, age group (aged ≤45 or >45 years), and region (northern or southern China), to provide a diversity of views.²⁰ Each primary care physician nominated at least 2 adult patients for inclusion in the study. If the patient met the study inclusion criteria (adult and able to communicate in Chinese), they were sampled across sex, age, educational attainment, and occupation. WeChat was the main means of communication with primary care physicians and patients. Video teleconferencing software, such as Zoom (Zoom Video Communications) and Dingtalk (Alibaba Group), were used to conduct the interviews. Informed verbal consent was obtained and confidentiality was assured to all participants before the interviews, which were audio recorded. Incentives (gift cards equivalent to US $28) for participation were provided and documented.

Semistructured interview guides were developed separately for primary care physicians and patients based on the systematic literature review conducted by the research team (Supplemental Appendix 2).⁹ Pilot interviews were conducted with 7 participants (4 primary care physicians, 3 patients) and refinements were made to the guides before the data collection. All interviews were conducted in Putonghua; each lasted 45-60 minutes. All audio recordings were stored with access restricted to the study team. The research team met regularly to review the interview codes and themes until no new topics were emerging, indicating data saturation was achieved.²¹

Data Analysis

The digitally recorded interviews were transcribed verbatim by an independent research assistant. The post-analysis translation method was utilized for retaining explicit and implicit meanings embedded in the language for culturally specific expressions and concepts.²² Thematic analysis followed Braun and Clarke’s guidelines.²³ Two researchers independently studied the data, reading, rereading, assigning codes, then sorting, grouping, and merging approximately 390 codes. These were further categorized into patterns or themes.²³ An audit of 15% of the coded segments was conducted by a third researcher (W.C.W.W.) who discussed findings with both I.Y.Z. and Y.X.M. to ensure inter-rater agreement. Disparities were discussed until agreement was reached.

Summary of Demographic Data

A total of 16 primary care physicians (8 of each sex) aged 29 to 65 years completed the interviews. Of these, 11 had additional training in general medicine (most physicians in China start in general medicine then become primary care physicians via on-the-job training). Thirteen of the physicians had more than 10 years of professional experience and concurrently held titles of Associate Chief Physician or above.

A total of 24 patients (13 females, 11 males) aged 21 to 70 years completed the interviews. Patients had a variety of occupations and levels of educational attainment (Table 1). Five key themes and exemplar quotes from primary care physicians and patients are summarized in Table 2.

Potential Impacts on Trust and Informed Consent

Chinese primary care physicians and patients were aware of health codes being used as a form of DDS and contact tracing in response to the COVID-19 pandemic. Many of the primary care physicians and patients credited the Chinese government’s efficient response and success with COVID-19 pandemic control to China’s centralized administrative system. While primary care physicians generally held positive views on DDS, they had considerable reservations. Some primary care physicians believed that the clinical decision support system that allowed use of telehealth and an automated diagnosis program helped address a workforce shortage and offered timely warnings of ID outbreaks, yet it fell short of replacing in-person medical assessments. Primary care physicians were also concerned that the implementation of DDS might adversely impact their patient relationships. Some suggested a lack of transparency about use of patients’ data could potentially disrupt the already fragile trust-based doctor-patient relationship in China.

During health care reform, patients were led to believe that all information provided to the hospital would be kept confidential, but there are exceptions. Medical confidentiality is subordinate to active public health threats. An option to opt out of DDS would defeat its purpose. Some primary care physicians were under the impression that it was their responsibility to seek the patients’ consent. This was at odds with mandatory ID reporting that does not allow for sufficient consumer understanding of the implications of engagement with DDS. Some physicians felt this could cause future doctor-patient conflict.

Need for Autonomous Decision Making

Digital detection surveillance negatively impacted both primary care physicians’ and patients’ sense of autonomy in decision making. Primary care physicians expressed a sense of diminished autonomy in their decision making. Some patients expressed concern that clinical decision support systems could de-skill physicians despite DDS being intended as an auxiliary clinical tool. Patients also stated that their engagement with DDS meant they had no ability to ensure personal privacy and no formal channels to express their concerns. Several primary care physicians and many patients agreed that national institutions in China hold considerable power, and as a result local governments and primary institutions apply strict control strategies. The fear of interfacing with DDS during clinical consultations meant some patients were reluctant to go to hospitals when they had symptoms of potential IDs because they feared compulsory quarantine. Others admitted to lying about their symptoms, such as fever or cough, in order to obtain a green health code.

Potential Discrimination Associated With ID

The DDS was seen as forcing citizens to adopt the traffic-light health code system based on infection risk. Of note, a few primary care physicians described how their communities discriminated against them after they had been in contact with COVID-19 patients. Patients also outlined fears about the risk of ID disclosure, being identified as yellow or red code holders, and being treated differently in the community. Some patients questioned the accuracy of the coding within DDS (eg, coincidental fever, or not including a district with known COVID-19 cases) which could lead to unnecessary quarantine and significant stress on families. Both primary care physicians and patients recommended public awareness education on the integrity of ID and DDS as being a useful solution to reduce public discrimination and fear.

Risk of Inequity and Divided Care

Certain groups (ethnic or sexual minorities) were regarded as being vulnerable to ID disclosure and resultant shame or exploitation. Older people and young children, who might have difficulties in using smart phones or scanning health codes, were seen to be particularly at risk of falling victim to disparate care. This concern also applied to persons with lower literacy levels as they might not be aware of their data protection rights and could be more prone to online fraudsters. Some primary care physicians thought that the use of DDS in smaller towns could easily lead to the identification of ID patients. Some physicians also raised concerns about the accuracy of the predictions based on models using artificial intelligence, and felt most patients would not understand this implication.

Health Data Security: Whose Responsibility Is It?

While participants stressed that de-identified information from patients diagnosed with an ID should be shared globally for optimal public health outcomes, they felt the current lack of security and protection in the DDS system and regional variations in epidemic control could make information sharing a national security issue. Both primary care physicians and patients recommended that only disease-related information be extracted (not personal data), and that this information should be decrypted with restricted access. They worried that some countries would blame one another for poor performance in the prevention and control of the ID, further imposing considerable security threats beyond the mandates of China’s Center of Disease Prevention and WHO. Furthermore, there was widespread distrust in security of the transnational data, attributed to the loose state control of democracies and high cost of medical care for international students, tourists, and expatriates. Both primary care physicians and patients indicated that an independent international organization reporting to WHO should oversee the preparation and implementation of a universal ethical and legal governance framework. Primary care physicians suggested each country should establish a dedicated team to maintain personal privacy and prevent cyberattacks with inter-governmental agencies bearing the ultimate responsibility and legal regulation of the information.

Key Findings

This study reports, for the first time, qualitative data from primary care physicians and their patients on how they experienced DDS in primary care consultations and specific ethical concerns in China. The potential for DDS to disrupt the essential but fragile trust in doctor-patient relationships in China was a primary concern, especially after hard-won gains based on updated medical training that emphasizes trust-based relationships.^24,25 Other studies suggest that DDS introduced into primary care physician consultations without obtaining informed consent could reignite public fear and distrust of health care professionals.^26,27 Trust is a foundation for medical consultations and highly relevant to predicting satisfactory patient outcomes.^28,29

Ambiguity about obtaining informed consent prior to DDS engagement in China has been highlighted.^30,31 According to Article 13 of the “Personal Information Protection Law of the People’s Republic of China” passed in August 2021, personal information can be collected and processed without informed consent in response to public health emergencies, or for the sake of protecting citizens’ lives, health, and property.³² This is in contrast with the recent national policy changes in the hospital health care delivery system that reflect the increasing importance of engaging consumers in their own care.^26,33 The process of seeking and obtaining fully informed consent for DDS has had vastly different outcomes globally.³⁴ Yet, the more often consent is withheld, the lower the accuracy of DDS. There is a need for a balance between an individual’s right to not give consent for DDS and public health-focused ID identification.³⁵

China’s current traffic light system of health coding, recommended by WHO for COVID-19 pandemic control, could discriminate against citizens with a yellow or red code. Engagement with DDS can contribute to social inequity by creating digital divides, particularly for some groups (eg, older people and children). The DDS accuracy is unavoidably influenced in the process of structuralizing the text data, coding the prescription data, and entering specific details of the presenting symptoms. Some cases of ID might be missed due to the limitations of artificial intelligence. This finding has also been noted in the use of digital health systems internationally, where contact tracing has exposed or exacerbated health inequities by failing to reach specific geographical areas and populations due to lack of internet access and smartphones.³⁵

Developing an Ethical and Regulatory Framework

An initial priority would be to unify data protection processes within China through regulations on data sourcing and shared responsibility for personal data by the use of strong encryption processes.³⁶ Citizens’ rights for autonomous decision making on the use of their personal information, data co-ownership, and participation in the oversight of their data must be respected.³⁷ Cloud-based data storage and sharing is the most frequently used technology to date, however a significant limitation is that patients’ rights are often ignored.³⁷ Next-generation data exchange solutions based on block chains have the potential to clarify access rights and ownership, protect data authenticity, and retain privacy. However, data processing rules such as those based on real-name authentication or facial recognition for data access must be disclosed, and security protection measures adopted to safeguard public interests should not hinder normal personal and social lives.

Adopting patient-centered primary care models that carefully manage communication on the use of DDS³⁷ would further strengthen the doctor-patient relationship and establish transparent pathways from end users through to policymakers.⁷ Digital detection surveillance developers, data analysts, public health officials, and policymakers must also take responsibility for engaging collectively with certain groups to address the deep-rooted inequities in data set indenisation.³⁸ Furthermore, education on consumer rights, data security, and fraud detection needs to be prioritized within primary care physicians’ medical education and public awareness campaigns.

Limitations

This study has several limitations. First, involving allied stakeholders with DDS through co-design approaches could have enabled a more collective and real time development of an ethical framework than possible in the current approach. Second, some primary care physicians failed to recommend 2 patients due to high workload, thus response biases might exist which could have impacted data saturation. Third, although robust efforts have been applied to ensure translation quality, nuanced meanings in Chinese language interviews could be missed in the translation to English.