Dutch public health professionals’ perspectives and needs regarding citizen involvement in COVID-19 contact tracing through digital support tools: an exploratory qualitative study

A cross-sectional exploratory semi-structured interview study was performed. For the most part, we followed the COREQ-checklist for interviews and focus groups to enhance this study’s methodological quality and transparency [16]. Note that some items from this checklist (e.g., member checks, repeat interviews) were not implemented due to resource and/or time constraints with the research team, as well as with the research participants.

CT in the Netherlands is traditionally executed by PHPs (qualified public health nurses and -doctors). During the COVID-19 pandemic, many temporary CT-employees (e.g., medical students) were hired for the execution of CT to overcome personnel shortages, shifting PHPs more towards the coordination and organization of CT. However, in contrast to PHPs, temporary CT-employees are usually not specialized in public health, have very limited (or no) experience with CT in general, and have limited oversight over the CT-process in its entirety. Therefore, in this study, we primarily focused on PHPs.

To investigate PHPs’ perspectives and needs regarding more actively and autonomously involving cases and contact persons in CT for COVID-19 through DCTS-tools, we developed a semi-structured interview guide based on the reasoned action approach (RAA). The RAA assumes that an individual’s intention to perform a certain behavior (in this study a PHP’s intention to involve cases and contact persons more actively and autonomously in CT through DCTS-tools) is determined by a set of behavioral, normative, and control beliefs, that lead into three proximal determinants: attitude, perceived norm, and perceived behavioral control [17]. In this study, we developed interview questions based on the RAA’s proximal determinants to elicit PHPs’ beliefs that underpin their perspectives and needs regarding more actively and autonomously involving cases and contact persons in CT of COVID-19 through DCTS-tools. The interview guide (translated to English) can be found in Additional file 1.

Interviews were structured into four main sections. Section one contained general questions regarding PHPs’ experiences with CT for COVID-19, how they felt about the role and participation of cases and contact persons in CT, and their perspectives on digitalization of CT. Section two contained a step-by-step presentation of Fig. 2, to introduce interviewees to the DCTS-tools of interest to this study (delivered through PowerPoint V.2102; figures were created using Microsoft Visio Professional V.2102). The DCTS-tools are based on previous research that we conducted on the application of respondent-driven detection, a snowball-like sampling method for case finding, in the context of pathogens that transmit via close (physical) contact between individuals [18, 19]. DCTS-tools were presented to interviewees in a generic fashion, to allow interviewees to deliberate freely. Section three contained questions based on the RAA’s sub-constructs, which were separately asked for all three DCTS-tools. Section four contained questions regarding interviewees’ final suggestions and thoughts regarding DCTS-tools in general.

The interview guide and the introduction of the DCTS-tools were extensively pilot tested among PHPs from the RIVM. Based on the pilot interviews, only small changes to the order of the questions were made.

Between October 2020 and February 2021, semi-structured in-depth one-on-one interviews were conducted with PHPs involved in the execution of CT for COVID-19, from various PHS in the Netherlands.

There are 25 PHS in the Netherlands, all of which serve different geographic regions. At the time that this study was conducted, the execution of CT for COVID-19 differed between - as well as within - PHS over time, mainly due to (regional) fluctuations in caseload (as illustrated in Fig. 1). We anticipated that the degree to which PHPs’ respective PHS would execute ‘full’, or ‘scaled down’ CT, and other PHS-specific factors related to the execution of CT, could influence PHPs’ perspectives and needs regarding actively and autonomously involving cases and contact persons through DCTS-tools. To account for this, we aimed to include at least one PHP from a PHS from each of the twelve provinces in the Netherlands, over a period of several months. In addition, we aimed to include PHPs with diverse characteristics, including age, gender, years of experience with CT (also prior to the COVID-19 pandemic), and professional background (nurse or doctor).

PHPs were identified and invited through the professional network of the RIVM, using snowball sampling and referrals (i.e., a combination of convenience and purposive sampling was used). For example, we asked regional PHS-coordinators to distribute an invitation to different PHS (for further internal distribution to PHPs) and/or to directly refer us to potential research participants. PHPs who received an invitation through their respective PHS and contacted the executive researcher (YBH) to express interest in participating in the study, and PHPs whom we were directly referred to, first received an in-depth information letter about the study via e-mail. PHPs who then wanted to participate in the study were asked to agree to the terms of the study through an online informed consent form. Finally, after having agreed to the informed consent form, PHPs were automatically redirected to a short online questionnaire through which we asked for their age, gender, professional background, province of employment, years of experience with CT, and months of experience with CT for COVID-19. The online informed consent form and questionnaire were distributed using software for respondent-driven sampling that was jointly developed by the RIVM, UMC Utrecht (the Netherlands), and the Karolinska Institute (Sweden).

Due to the social distancing measures implemented at that time, interviews were conducted online, using the web-conference software Cisco WebEx Meetings (V.40.2.14.19). Interviews were conducted by one male interviewer (YBH) and lasted approximately one hour. Interviewees received a small token of appreciation (worth approximately €10) for their participation.

Interviews were audio recorded, using the recording tool available in Cisco WebEx Meetings, and transcribed verbatim. A thematic analysis (making use of open, axial, and selective coding) was conducted in MAXQDA Plus 2022 (Release 22.0.0) [20]. Coding was inductive and was guided by the research questions presented earlier. We started by analyzing if and why (research question 1), and how (research question 2) PHPs would want to involve cases and contact persons through DCTS-tools 1, 2, and 3, separately. From this, we identified overarching themes related to PHPs’ perspectives and needs regarding the application of DCTS-tools in general. To reduce subjective interpretation of the data, random selections of 13 and 5 transcripts were also coded by a second (AM) and by a third (RB) researcher, respectively. Coding was discussed until consensus was reached.

This study was reviewed by the Medical Ethical Committee of the University Medical Centre Utrecht, who exempted this study from the need for a full medical ethical review (reference number: 20–662/C).

Seventeen PHPs were interviewed, representing PHS from all 12 provinces in the Netherlands. Interviewees had a median age of 39 years (IQR: 32–50) and a median of 4 years (IQR: 2.5–9) experience with CT in general. The median CT-experience for COVID-19 was 8 months (IQR: 5–9). Most interviewees were female (76.5%) and worked as a PHS-nurse (76.5%). At the time that the interviews were conducted, most interviewees’ PHS facilitated full CT (52.9%). See Table 1 for an overview of the sample characteristics.

Although we did not perform an in-depth analysis in this regard, we did not encounter (signs of) relationships between sample characteristics and interviewees’ perspectives and needs towards DCTS-tools, except for ‘PHS CT-phase’, which is discussed in more detail under ‘Circumstances in CT for COVID-19 that permit or constrain the application of DCTS-tools’ in the results section.

No new (sub-)themes were identified (i.e., data saturation was reached) after the analysis of 11 interviews.

Since we relied on referrals and snowball sampling for the recruitment of PHPs, it is impossible to calculate the response rate of our study. However, all individuals to whom we eventually sent the online informed consent form and questionnaire participated in this study.

Four main themes related to PHPs’ perspectives and needs regarding the application of DCTS-tools for CT of COVID-19 emerged from the data: ‘Distinct characteristics of CT with DCTS-tools’, ‘Anticipated benefits and challenges of DCTS-tools for CT of COVID-19’, ‘Circumstances in CT for COVID-19 that permit or constrain the application of DCTS-tools’, and ‘PHPs’ needs regarding the application and development of DCTS-tools’. Main- and sub-themes are discussed in more detail in the following sections.

In this study, we asked Dutch PHPs if, why, and how cases and contact persons could support the execution of CT for COVID-19 by PHPs through digital contact tracing support tools (DCTS-tools). Four main themes related to PHPs’ perspectives and needs were identified: ‘Distinct characteristics of CT with DCTS-tools’, ‘Anticipated benefits and challenges of CT with DCTS-tools’, ‘Circumstances in CT for COVID-19 that permit or constrain the application of DCTS-tools’, and ‘PHPs’ needs regarding the application and development of DCTS-tools’. PHPs seem to have an overall positive attitude towards the application of DCTS-tools, in all stages of the CT-process. It was anticipated that DCTS-tools can have important benefits for CT, namely: more efficient CT, more opportunities for cases and contact persons to participate in CT in a manner that best suits them, and enhanced quality of data collection and administration. However, several challenges were also anticipated, namely: adequate execution of CT strongly depends on the willingness and skills of cases and contact persons, and limited support and guidance for cases and contact persons in the CT-process. PHPs appear slightly more positive about DCTS-tool 1 in the contact identification stage, and DCTS-tool 3 in the contact monitoring stage, as interviewees were more worried about the anticipated challenges of DCTS-tool 2 in the contact notification stage. Notably, PHPs’ attitudes seem conditional on the circumstances under which CT is performed, and the fulfillment of their needs in the development and implementation of DCTS-tools.

Our findings are largely in line with results from a previous study with a similar set-up and similar aims that we conducted among Dutch PHPs before the COVID-19 pandemic, in the context of mumps, shigella, and scabies [18]. However, in the previous study, we found that Dutch PHPs were less inclined to give cases and contact persons more autonomy and responsibility in CT in the context of communicable diseases that were perceived as relatively severe for individual and public health, as COVID-19 usually was at the time that the interviews for this study were conducted.

The current study indicates that this observed discrepancy is, at least to some degree, related to the scarcity of human resources at Dutch PHS to fully facilitate CT, especially during pandemic peaks, and the unprecedented magnitude of the COVID-19 crisis. For example, PHPs noted that DCTS-tools could provide opportunities to continue to identify, notify, and monitor contact persons, when these aspects of CT can no longer be facilitated by PHPs. In addition, interviewees noted that, compared to other communicable diseases, cases and contact persons are more aware and knowledgeable about CT for COVID-19, and about what is expected of them in that regard. Our results indicate that this makes it ‘easier’ for PHPs to rely on cases and contact persons in the execution of CT. This may potentially also be true for CT with DCTS-tools during future outbreaks of communicable diseases, and under non-pandemic circumstances, if general knowledge and awareness about CT is retained among the public to some extent beyond the COVID-19 pandemic. We emphasize, however, that this suggestion requires further in-depth exploration.

Since DCTS-tools rely on actions from cases and contact persons, PHPs’ perspectives and needs regarding DCTS-tools strongly depend on their trust in the willingness and ability of cases and contact persons to participate in CT through DCTS-tools, in a timely and complete manner. As of now, however, insights into this topic are scarce. Studies in the field of patient referral and expedited partner therapy for sexually transmitted diseases have found that notification rates and treatment outcomes can be improved by giving cases and contact persons more responsibilities and autonomy in the identification and notification of contact persons [21]. However, the identification and notification of contact persons from sexual networks is markedly different to the identification and notification of contact persons from (close) contact networks (e.g., in terms of exposure-risk assessment guidelines and stigma associated with sexual contact that cases and contact persons may experience). One study conducted during the COVID-19 pandemic found that contact persons could effectively be identified through self-led online interviews, indicating that cases may indeed have the potential to contribute to CT for COVID-19 in this regard [22]. However, this study was performed in an experimental setting (not in the context of CT) with healthy adults, which may not accurately reflect the circumstances of CT in practice. For example, previous research indicates that participation in CT generally suffers from mistrust in public institutions and privacy concerns, which may be especially challenging issues in the context of DCTS-tools that rely on relatively autonomous actions from cases and contact persons [23]. In addition, whilst digital tools may generally make CT easier and more accessible for some individuals, they may pose additional barriers for individuals who are less digitally skilled or literate [24]. Therefore, we conclude that further research is necessary to identify benefits, challenges, and needs regarding DCTS-tools from a ‘citizen’s perspective’. To this purpose, we are currently conducting a study (with a similar design and purpose to this study) among Dutch citizens. In addition, we plan to conduct small pilot studies, in which we will assess the conditions under which cases are willing and able to identify and notify their contact persons in a timely and complete manner.

One key strength of this study is that we conducted interviews with PHPs who were directly involved in the organization, coordination, and execution of CT during the COVID-19 pandemic, around the second pandemic peak in the Netherlands. Therefore, this study provides unique and valuable insights in the perspectives and needs of PHPs regarding the involvement of cases and contact persons in CT through DCTS-tools in a pandemic context. As such, we believe that this study is particularly relevant for public health practitioners and for the development of preparedness and response plans for (future) outbreaks of communicable diseases that transmit via (close) physical contact between individuals.

Another strength is that we interviewed a relatively diverse sample of PHPs. Therefore, we believe that we captured a broad variety of perspectives and needs from Dutch PHPs. Nevertheless, since we used a combination of purposive and convenience sampling, it should be kept in mind that we may have recruited PHPs with an above average interest or motivation to participate in the study, which may have (positively) biased our findings regarding the application of DCTS-tools. We are currently performing a quantitative follow-up study to assess PHPs’ attitudes towards DCTS-tools more thoroughly, in a larger population of PHPs.

An important limitation to our study is that our findings should primarily be interpreted in a pandemic context, where PHS-capacity for CT is limited, and where widespread circulation of a pathogen has already occurred. The limited availability of medical interventions (e.g., vaccines) to control the outbreak and the specific epidemiological characteristics of COVID-19, including its severity and infectiousness, at the time that this study was performed (the Alpha-variant becoming dominant in the Netherlands) should similarly be considered in the interpretation of our findings. Furthermore, since this study was conducted in the Netherlands, our findings may be more relevant to countries with similar characteristics, including, for example, the organization of the overall CT-system and the CT-protocols in place, and the degree of digitalization of the population (in 2018, 98% of Dutch households had access to broadband internet and 84% of the Dutch population had internet access through their mobile devices outside home or work) [25]. It may be that PHPs’ perspectives and needs regarding DCTS-tools differ in the context of a relatively new and/or isolated outbreak, when outbreak control and knowledge generation more critically rely on CT, when CT is performed for pathogens with different epidemiological characteristics (e.g., the currently dominant Omicron-variant), and/or in less digitalized countries. Nevertheless, in a similar study that we conducted before the COVID-19 pandemic, we found very similar anticipated benefits, barriers, and needs of Dutch PHPs, but in a different (non-pandemic) context and for different communicable diseases [18]. Therefore, we believe that our findings are, at least to some extent, still relevant to other contexts, and that DCTS-tools may also be of added value to CT in general (e.g., from an efficiency and/or public engagement perspective).

Internationally, CT for COVID-19 (and other communicable diseases) may increasingly depend on (autonomous) actions from cases and contact persons when resources for CT are scarce [8, 26]. DCTS-tools to support the execution of CT by PHPs, such as those discussed in this study, may be particularly useful under such circumstances [27]. We were, however, unable to find detailed descriptions or evaluations of DCTS-tools that allow cases and contact persons to contribute to the traditional execution of CT by PHPs in scientific literature. As such, we believe our study provides particularly useful insights for the development and implementation of DCTS-tools, as well as for readily existing DCTS-tools, that rely on actions from cases and contact persons in CT.

Based on our results, it appears crucial, from a public health practice perspective, that DCTS-tools are integrated in CT-protocols in a manner that allows PHPs to maintain some oversight and control over the CT-process. Therefore, in line with previous research, we suggest that DCTS-tools should be used in support of - rather than instead of - traditional CT, so that PHPs can remain personally involved to ‘guide’ the CT-process, or to motivate and support cases or contact persons, if necessary [12, 13, 28]. For example, in the contact identification stage, cases could be asked to start to collect their personal and their contact persons’ data using DCTS-tool 1, immediately after they receive a positive test result. When cases are then contacted by a PHP for CT (as usual), the collected data can be automatically transferred to PHS, after which a case and a PHP can jointly review the data. For cases who do not use DCTS-tool 1, the identification of contact persons can be carried out as usual. In the contact notification stage, cases and PHPs could jointly decide which contact persons are notified by whom and/or cases could quickly notify contact persons in advance of - rather than instead of – a PHP, using DCTS-tool 2. Similarly, in the contact monitoring stage, PHPs and contact persons could jointly decide if DCTS-tool 3 should be used or not. We emphasize, however, that the precise integration of DCTS-tools in CT-protocols requires further (context-specific) tailoring, and careful consideration of relevant standards for data security and privacy protection (e.g., the General Data Protection Regulation). Finally, for the purposes of this study, we treated DCTS-tools 1, 2, and 3 as separate entities. However, we would like to note that it may be beneficial for practical reasons to combine the DCTS-tools into one or two mobile and/or web-based applications (e.g., so that cases and contact persons do not have to download/use multiple applications).

Dutch PHPs seem positive towards giving cases and contact persons more autonomy and responsibility in CT for COVID-19 through DCTS-tools. However, PHPs’ attitudes appear to be conditional on the circumstances under which CT is performed and the fulfillment of PHPs’ needs in the development and application of DCTS-tools. In future research, we will investigate the perspectives and needs of cases and contact persons on DCTS-tools for COVID-19, and conduct pilot studies to investigate the application of DCTS-tools in practice.