Data is vital for public health program decision making and intervention, for example in prevention and control of the epidemic of human immunodeficiency virus (HIV) and acquired immune deficiency syndrome (AIDS) [[1], [2], [3]]. To achieve the goal of ending the HIV/AIDS epidemic by 2030, the Joint United Nations Programme on HIV/AIDS (UNAIDS) recommends that member countries collect, analyse and disseminate high-quality HIV/AIDS data [4]. Since HIV/AIDS data is captured in national public health information systems (PHISs) [5,6], the quality of the PHIS data collection process is vital for acquisition of high-quality HIV/AIDS data.
The HIV/AIDS epidemic has remained a critical public health challenge in China [3,7,8]. By October 2019, about 958,000 persons were recorded living with HIV/AIDS (PLWHA) nationwide. The 2018 national HIV/AIDS epidemic estimation results indicate the actual number of PLWHA ranged from 1.1 to 1.4 million by the end of 2018 and keep growing in the near future [8,9]. The Chinese HIV/AIDS Comprehensive Response Information Management System (CRIMS) is a national repository of data for HIV/AIDS “project planning, budgeting, implementation, monitoring and evaluation” [6]. The CRIMS data collection process needs to be of high quality to meet the information needs of the decision-makers on HIV/AIDS prevention and control.
The CRIMS commenced officially in 2008 as a sub-system of the China Information System for Disease Control and Prevention, which is a large-scale web-based disease surveillance system [5,7]. A variety of electronic reporting forms have been developed for data collection and entry into the CRIMS by the Chinese Centre for Disease Control and Prevention (China CDC) [5,6,10]. The data sources for these forms primarily include case reporting and management, healthcare services for PLWHA, intervention services on high-risk groups, and national HIV/AIDS prevention and control program management [5,11]. The major data producers and collectors for the system are the county CDCs and hospitals that provide public health services and interventions to the target groups [11].
In the last decade, a data-driven performance assessment scheme has been established for assessing the data quality of the CRIMS [5,12]. Implementation of the scheme has led to an improvement in the quality of the reporting data in the system [12,13]. However, certain reporting data, e.g., case demographics, case follow-up, and intervention delivery, were still inaccurate, incomplete, missing, delayed, under-reported or leaking [[14], [15], [16], [17]]. The national data quality assessment of intervention in the population at high risk for HIV/AIDS between 2014 and 2018 suggested that 79.4 % (70.5 %–98.3 %) of the data recorded in the CRIMS and in the paper records were consistent. However, in 2018 four types of consistency rates dropped to 85.3 %, 91.0 %, 78.8 %, 70.5 %, respectively, all ranking lowest within the five-year span [18,19]. A literature review of the CRIMS data management studies in peer reviewed Chinese and English electronic databases showed that 61 % (37/61) of the studies focused on assessing the quality of data stored in the CRIMS [13]. The other studies focused on the development and management of the information systems or the influential factors on data collectors. Few studies identified or provided evidence on where, when and how data quality problems occur, the causes of poor data quality, or what strategies can be implemented to improve data quality. As data quality problems often occur in the data collection process [[18], [19], [20], [21]], there is an urgent need to understand the factors influencing the performance of the process so as to generate insight on data quality management for the CRIMS.
To date, the quality of the PHIS data collection process is an under-researched area [22]. Our previous study identified that several PHIS data quality assessment methods were focused on the data collection procedure, i.e., data recording, storage and audits, and the functions of the PHIS system that facilitate or hinder data collection [[22], [23], [24]]. Little attention had been given to the effect of the contextual factors (organizational, personnel or environmental) on the quality of the data collection process [22]. To address this knowledge gap, we have constructed a four-dimensional (4D) framework based on a systematic literature review of the topic from the international literature [25]. Unlike other data quality frameworks issued by public health institutions such as the United States CDC’s Guidelines for Evaluating Public Health Surveillance Systems and the CIHI Data Quality Framework [20,23,24], the 4D Framework is primarily focused on assessing the PHIS data collection process.
An expert elicitation study to validate the structure of the 4D Framework confirmed that it should cover four dimensions (data collection management, data collection environment, data collection personnel, and data collection system) [26]. These dimensions comprised 16 subcomponents and 116 indicators including 82 facilitators and 34 barriers (see Fig. 1). By providing guidance to practitioners to harness the facilitators and to address the barriers, the 4D Framework can be a promising quality improvement model to strengthen the management of the public health data collection process.
In this study we applied the 4D Framework to assess the quality of the CRIMS data collection process. We aimed to identify the gaps in the process and suggest improvement strategies for HIV/AIDS data collection in China.