Behavioral Determinants of Routine Health Information System Data Use in Senegal: A Qualitative Inquiry Based on the Integrated Behavioral Model

"Understanding the behavioral determinants of RHIS data use is a crucial yet seldom addressed topic in the literature, particularly in the context of low- and middle-income countries..."

Routine health information system (RHIS) data are essential for planning and decision making at all levels of the health system yet tend to be underutilised. RHIS data use strengthening programmes often focus on providing technical solutions and infrastructural support while failing to address behavioural barriers, which can be difficult to identify and address. This article shares research applying the Integrated Behavior Model (IBM) to examine how attitudes toward RHIS data, perceived norms concerning RHIS data use, and the ability to use RHIS data influence the demand and use of RHIS data among stakeholders in Senegal.

IBM is an adaptable theoretical framework that combines health behaviour research theories, such as the Theory of Reasoned Action and the Theory of Planned Behavior. The IBM postulates that behavioural intention is the strongest determinant of human behaviour, which can be predicted by an:

• An individual's attitudes toward the behaviour: Attitudes arise from an assessment of the behaviour itself (experiential attitudes) and beliefs about the likelihood that performing the behaviour will have certain outcomes (instrumental attitudes).
• Perceived norms of the behaviour: Norms are divided into: (i) injunctive norms (whether others in one's social group approve or disapprove of the behaviour) and (ii) descriptive norms (how common the behaviour is within one's social group).
• One's perceived ability to perform the behaviour: Perceived ability is divided into: (i) perceived control (the extent to which one feels responsible for one's behaviour; and (ii) self-efficacy (one's confidence in being able to perform a behaviour despite the challenges that may arise).

Data for the qualitative study come from in-depth interviews conducted between January 2019 and November 2019 with 18 key informants, 9 of whom were high-level decision makers and 9 of whom were mid-level personnel within organisations active in malaria, tuberculosis, and HIV programmatic areas in Senegal. The study treats data demand and data use as modifiable health behaviours. The study includes the first 2 steps outlined by Marco Yzer for determining effective, targeted health communications: (i) clearly defining the health behaviour to be targeted, and (ii) gathering information directly from the population to be reached with the behaviour change effort.

The interviews revealed that RHIS data use is primarily determined by data demand (i.e., intention to use RHIS data) but also by competency to use RHIS data (i.e., knowledge and skills for RHIS data use) and situational constraints affecting RHIS data use processes. Data demand is affected by several factors, including attitudes about the quality, availability, and relevance of RHIS data for decision making, as well as perceived norms around RHIS data use and an individual's perceived ability to use RHIS data. The paper presents the emergent themes as supporting evidence for the different constructs in the IBM. For instance:

• Attitudes: Respondents felt using RHIS data was time consuming and unreliable and that, often, the level of granularity was insufficient for their needs. National-level respondents generally expressed low trust in the data use competencies of data producers and users in the lower levels of the health system/organisations, as well as their level of engagement and understanding of the value of data. Several respondents saw the bottom-up approach of reinforcing positive instrumental attitudes at lower levels of the system as important in promoting a culture of accountability, as well as in valuing and using routine data. Suggested opportunities to reinforce the attitudes included during periodic workforce training activities and during pre-service training.
• Social norms: Respondents from organisations within the Ministry of Health (MOH) often reported that the establishment of guidelines, standard operating procedures, and manuals improved the use of RHIS data by setting norms and guidelines around data use. Some stakeholders pointed out that the availability of bulletins and newsletters enabled the comparison of performance indicators across districts and set standard expectations. Multiple respondents pointed out that many data users valued the opinion of their peers from other parts of the country and thus strived to produce better-quality data and to use data more effectively.
• Perceived ability: Respondents felt they lacked control over the data production, data sharing, and data dissemination processes, which in turn affected their beliefs in their ability and comfort to use RHIS data consistently. Among civil society respondents, the lack of direct access to national RHIS data and lack of clarity on data sharing policies were reported to hinder their ability to review RHIS data and to assess their accuracy. This lack of control of data collection affected both their level of trust in the data and their sense of involvement in the national processes related to RHIS data. Consensus emerged that the perceived low levels of skills for data use, lack of necessary equipment, lack of office space, and inadequate wages contribute to the low motivation among health workers to use RHIS data. Respondents pointed to the need to strengthen the capacity of all data users through interventions such as training workshops, coaching, and supervision.

In short, respondents broadly acknowledged the importance of RHIS data and the role of their quality in guiding the decision-making process. These experiential attitudes appeared to be common knowledge within this population, suggesting they would not be important targets for behaviour change communication among data users working at strategic levels of their organisations. For this cadre of data user, it may be far more effective to target the instrumental attitudes rooted in the beliefs that RHIS data are not reliable for guiding decision making due to their poor quality and inconsistent availability. Other implications of the reserch include, for example:

• In light of findings highlighting the concept of trust as an overarching theme, potential interventions could include communication of success stories and lessons learned in matters related to RHIS data use together with the promotion of role modeling from health authorities who use RHIS data despite data quality concerns.
• Policymakers should explore solutions to some of the systemic problems and work to support a data use culture by, for example, articulating clear policies on database access and data sharing. Such policies may promote transparency and allay some of the concerns that civil society actors have regarding using RHIS data. Based on the perceived lack of involvement in data production and dissemination processes in Senegal found within this population, it would be important to include the perspectives of civil society actors in policy development.
• National stakeholders should redesign existing data use strengthening programmes with a renewed emphasis on the behavioural aspects of routine data demand and data use. The advantages of leveraging preexisting mechanisms (e.g., training workshops) extend beyond merely changing attitudes and engaging data users at the individual level. At the ecological level, they contribute to the establishment of norms around data use (both descriptive and injunctive) through personalised normative feedback, which can motivate people to perform the desired behaviour, facilitate teamwork, and enable actors to feel involved in the broader RHIS improvement efforts.

The researchers assert that additional studies could help quantify the relative importance of the specific IBM constructs and clarify the comparative influence of identified barriers and facilitators on RHIS data use. There is also a need to explore the behavioural determinants of RHIS data demand and use at facility and district levels for a more comprehensive understanding of priority areas for action at all levels of the Senegalese health system.

In conclusion, this study has sought to identify current best practices and to suggest modifications to existing practices that are socially and culturally compatible with the local context. This work may potentially inform future strategies by Senegal's MOH to continue to strengthen the District Health Information Software version 2 (DHIS2), Senegal's national RHIS. "Sustaining the gains resulting from RHIS data use strengthening programs ultimately depends on the ability to promote a culture where data are not only valued at the individual, organizational, and policy levels but also used consistently to drive action."