Real-World Data and Paper-Based Disease Registries in the Small Island Developing State of Barbados During the COVID-19 Pandemic

Abstract

Significant data is being produced on the impact of COVID-19 on aspects of clinical care. However, less is known about the impact on real-world health data. The US Food and Drug Administration defines real-world data as “data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources,” including disease registries.¹ The methodology used by the Barbados National Registry (BNR)—active pursuit of first-hand clinical data using paper-based charts from multiple sources—makes it an ideal example of real-world data. Real-world data can overcome the barriers to clinical trials often present in small island developing states. This paper reviews the impact of the COVID-19 pandemic on the data of the BNR within the context of the real-world data cycle. Data collected retrospectively for 2016–2018, undergoing traceback during the pandemic, demonstrated a greater reliance on death certificate registration. A 38% reduction in the collection of new cases was noted in the postpandemic period compared to data collected in previous periods. The lack of access to source data delayed cancer registry reporting. We conclude that, given the challenges highlighted during the COVID-19 pandemic, more effort should be placed on providing timely access to real-world data for public health decision-making, particularly in small island developing states.

Introduction

The Barbados National Registry for Chronic Non-Communicable Diseases (BNR) reports incidence and mortality statistics for cancer in Barbados through a population-based cancer registry. Research suggests that quality of care can be achieved through 3 steps: (1) the ability to report on trends for a defined population, (2) determining the relationship of these trends to quality of care, and (3) measuring and monitoring this interplay to promote improvement and accountability.² Clinical trials are the reference standard for this type of data. However, clinical trials in developing countries, like those in the English-speaking Caribbean, face several barriers, including a lack of skilled staff, difficult patient recruitment, and ethical constraints.³ Real-world data and the evidence it provides can be an identifiable way of providing direction in health care decision-making.⁴ This paper examines the impact of the COVID-19 pandemic on the real-world data cycle of a population disease registry in the Caribbean by examining its impact on data acquisition, data processing, storage platforms, and the return value of cancer registry data for the Barbados National Registry.

Methods

Health care in Barbados is primarily free to citizens and is accessed as a social service through the lone public tertiary care hospital, Queen Elizabeth Hospital, and 9 community primary care health centers (polyclinics). The BNR collects data retrospectively from paper-based records obtained from the pathology, radiotherapy, and medical records departments of the hospital as well as from private pathology laboratories. Additionally, from 2019, data collected from encounters at polyclinics were available through the Health Information System of the Ministry of Health and Wellness. The data were collected and abstracted into the CanReg5 software, an open-source cancer registry database provided by the International Agency for Research on Cancer.⁵

The data collection and abstraction processes were treated as separate activities. This paper reviews the data collected, fully abstracted, cleaned, and analyzed by the cancer registry prior to March 2020 (pre-COVID-19; registration years 2008 and 2013–2014) vs after March 2020 (post-COVID-19; fully collected but undergoing traceback or cleaning; registration years 2015–2018). These data were reviewed to determine the impact of the COVID-19 pandemic.

Zhang et al outline the steps to produce data for analysis from the raw data acquired at encounters or through the provision of health services.³ The data collection process of the BNR was aligned to this real-world data cycle, as illustrated in Figure 1.

Figure 1.

The Barbados National Registry Process in the Framework of Real-World Data

Comparing 2020 against 2021, the impact of the COVID-19 pandemic was assessed by examining its impact on the data at the 4 stages of the real-world data cycle: acquisition, processing, storage, and return value. This was achieved by reviewing impacts at the data sources on the processing steps (data collection and abstraction), as well as on storage (using the CanReg5 database) and on the reporting and data requests of the registry as the main return-value activities.

Results

The University of the West Indies, which manages the BNR, closed and halted any research and data collection activities between March 24, 2020, and June 8, 2020, which meant the team could not access the paper-based records. The results of these actions on the data are outlined in Table 1.

Table 1.

Summary of Impacts on the Data Collection, Sources, and Processing of Barbados National Registry Data

	Registration years 2008, 2013–2014 (pre-COVID-19, pre-March 2020)	Registration years 2015–2018 (post-COVID-19, post-March 2020)
Death certification only cases	152	262
“Clinical only” cases	82	251
	Data Collection Year 2020	Data Collection Year 2021
New database entries	3,806	2,372
New database entries, February–June	22% of total cases	10% of total cases
Abstractions	947	1,285

While the cancer registry was able to maintain an average of 2 sources per case from 2016–2018, the orientation of the sources changed, with more cases registered as death-certificate only (DCO) as seen in Table 1. In addition, there were fewer cases from the private laboratory and the private physicians, whose paper-based notes were inaccessible, and more clinical-only cases, where data was accessible from the polyclinic health information system during the pandemic.

Despite the ability to access already-collected data via existing cloud-based servers, the lack of access to source data delayed reporting for the cancer registry. Analysis and quality control could be performed remotely. However, traceback and error resolution at the source were impacted, delaying the report by 1 year and 7 months for the 2015 report and 1 year and 5 months for the subsequent report (2016–2018) when compared against deadlines set pre-COVID.

Data is often requested for reporting by the Ministry of Health and Wellness, researchers, and doctors for health presentations to the public. This is one of the registry's key methods of returning value. There were 21 data requests in 2020 and 14 in 2021, down from 29 in 2018 and 2019.

Discussion

In order to realize the benefits of real-world data provided via disease registries in small island developing states, the data must overcome one of its greatest challenges: suboptimal quality.⁶ The COVID-19 pandemic demonstrated the impact that paper-based data collection can have on the registry's ability to provide quality real-world data. More effort should be placed on providing timely access to data for analysis and health decision-making. While surveillance can provide the data needed for action, paper-based active pursuit for the collection of raw data can be negatively impacted by any health system shock that reduces access to data at the source. As the expectation of more pandemics and climate-related barriers to data are on the horizon, direct access to encounter data through health information systems can reduce this impact.

Recommendations

1. To achieve the goal of accurate, timely information based on real-world data, developing states need to be supported in the development of resilient health information systems integrated across public and private sectors.

2. Countries need to foster cultures that prioritize data collection and dissemination as critical elements of the response in a health crisis.

3. Given limited resources in small island developing states, individual registries should consider leveraging regional resources through organizations such as universities.

Conclusion

In countries with limited resources, leveraging real-world data is an efficient way of obtaining information necessary to inform policy and clinical and public health decision-making. However, persistence with paper-based systems hinders the potential usefulness of the massive amounts of data collected data by healthcare workers and others in the field. In addition to funds spent on clinical treatment and management, financial investment is also needed to support health information systems that are capable of producing timely, accurate data to make decisions with clarity of vision and purpose.