Chapter 24. Using health inequality analyses in evidence-informed decision-making
Overview
WHO defines evidence-informed decision-making as “a systematic and transparent approach that applies structured and replicable methods to identify, appraise, and make use of evidence across decision-making processes, including for implementation” (1). In the context of health inequality monitoring, the results generated by inequality analysis constitute a form of scientific codified evidence obtained through systematic and replicable processes. This evidence, interpreted within the context of other scientific and tacit evidence (which includes opinions, expertise, lessons learnt, and operational insights from diverse stakeholders), is used to help identify high-priority areas for action and inform and evaluate actions to tackle health inequity.
Parts 1 and 2 of this book underscore the general importance of health inequality monitoring for advancing health equity, emphasizing continuous engagement with affected and target populations (see Chapter 4); audience- and purpose-driven reporting (see Chapter 7); and the role of health inequality monitoring in equity-oriented policy making (see Chapter 8). This chapter bridges a gap between data analysis and implementation, focusing on the integration of evidence derived from inequality analysis with other forms of evidence in decision-making processes.
The objective of this chapter is to describe approaches and considerations for using evidence about health inequalities to inform equity-oriented decision-making. It outlines a systematic approach to appraise the results of inequality analysis to identify high-priority areas for action. The chapter highlights considerations for cross-checking interpretations and consulting other forms of evidence as part of decision-making. It also demonstrates how the results of inequality analysis can be used as inputs in larger assessments and review processes for equity-oriented national health programming.
Identify high-priority areas for action
Inequality analyses can generate a plethora of results, especially if they involve multiple relevant indicators, dimensions of inequality and settings, assessed using disaggregated data and multiple summary measures of inequality. The use of a systematic approach to summarize this evidence can help to derive meaning from a large collection of findings and serve as an entry point to identify where actions are warranted. In some cases, this approach may lead to other questions that prompt further in-depth analyses and investigations, such as into the drivers of inequalities (see Chapter 25).
A scoring system to identify high-priority areas for action
The following exercise applies a systematic scoring system to the results of inequality analysis. It requires engagement with stakeholders (including affected populations) familiar with the results of inequality monitoring and the monitoring context. These stakeholders consider the inequality analysis results, with due acknowledgement of any limitations, such as those posed by sampling size and/or missing data for certain subpopulations. Stakeholders also consider these results in relation to contextual factors such as targets, health-care agendas and broader priorities relevant within the monitoring context. Typically, this process seeks to reach consensus among stakeholders.
For this prioritization exercise, inequalities across all health indicators by each dimension of inequality should be assessed. This entails reviewing the latest status, time trend and benchmarking – as well as multiple summary measures of health inequality. There may be a large amount of data to consider simultaneously. As a means of compiling the assessment, a table can be created, listing the health indicators as rows and the inequality dimensions (and overall averages) as columns. The inequality dimension columns can be divided further into absolute and relative inequality. Based on the results of monitoring, each cell of the table is assigned a score – for example, ranking from 1 to 3. Colour-coding may be applied:
1 ( blue) indicates no or low inequality by the selected inequality dimensions. This may mean that no further action is currently needed, although there may be issues of unmet need, forgone care or underreporting that need to be explored.
2 ( yellow) indicates some inequality, and action may be warranted.
3 ( red) indicates high inequality and the need for urgent action.
The overall average for each indicator can also be scored from 1 to 3. The average scores are calculated for each indicator (by row) and each dimension (by column), and then ranked to show the overall level of priority for action.
Using this approach, the criteria for scoring should be established at the discretion of the people carrying out the exercise (what is considered to be high or low inequality will vary). These criteria should be well thought out and stated clearly. The example above uses a scale of 1–3, but other approaches are also possible. Depending on the context and the preferences of the stakeholders, the results could be scored using a binary scale (e.g. action needed or action not needed). Alternatively, a multiple-value scale may be used to rank the level of urgency by two, three, four or more values. See Box 24.1 for an example of the application of the scoring system.
BOX 24.1. Example of a scoring system to identify high-priority areas for action
Table 24.1 illustrates how a three-point scale could be applied to assess absolute and relative inequality for three maternal and child health indicators. According to average inequality and national average scores, the most urgent priority for action was evident for births attended by skilled health personnel. The indicator of antenatal care (at least one visit) demonstrated the lowest urgency for action. For all three indicators, inequality related to education tended to be of moderate to high concern.
TABLE 24.1 Example of scoring system to identify high-priority areas for action across three indicators related to maternal and child health in a hypothetical country context
| Indicator | Economic status | Education | Sex | Place of residence | Subnational region | Average score | National average | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Absolute | Relative | Absolute | Relative | Absolute | Relative | Absolute | Relative | Absolute | Relative | |||
| Antenatal care: at least one visit | 1 | 1 | 2 | 2 | 1 | 1 | 1 | 1 | 1.3 | 1 | ||
| Births attended by skilled health personnel | 3 | 3 | 3 | 3 | 3 | 2 | 3 | 3 | 2.9 | 3 | ||
| Measles immunization coverage among children aged one year | 2 | 2 | 3 | 3 | 1 | 1 | 1 | 1 | 3 | 2 | 1.9 | 2 |
No or low level of concern
Moderate level of concern
High level of concern
Although this method lacks the ability to show nuances in the state of inequality, its simplicity is also an asset. Additional information, such as global and regional averages or trends over time, could be added to the table to provide extra context. The overarching purpose of priority-setting of both health indicators and dimensions of inequality is to help policy-makers interpret and apply the results of inequality monitoring. The conclusions derived from this exercise are not intended to be definitive, but they can provide input into wider discussions to determine where follow-up action is needed most. As discussed in Chapter 8, the introduction and implementation of policy is complex, depending on the availability of resources and infrastructure and political will.
A simplified variation of this approach involving thresholds and heatmaps may be undertaken by analysts as a preliminary step in exploring the findings. For a given set of results, analysts establish numerical thresholds that correspond to high, moderate and low inequality. Other thresholds may be established to identify where inequality has increased, decreased or remained constant over time. Referencing these thresholds, in a similar manner to Table 24.1, heatmaps (a type of visual that may be formatted similarly to a table, applying colour-coding that corresponds to data values; see Chapter 23) can be developed to visualize hotspots of potential concern. This approach may be a useful starting point to guide discussions and can be an input to the scoring system described above.
As one example, in the WHO State of inequality: HIV, tuberculosis and malaria report (2), difference values of 20 percentage points or higher between the richest and poorest wealth quintiles were deemed to constitute high inequality, values of 5–20 percentage points constituted moderate inequality, and values of 5 percentage points or less were considered low inequality. These general thresholds were agreed by stakeholders involved in the analysis to aid the initial interpretation of the broad set of results across countries. Individual results for countries were interpreted in light of other contextual information. In other iterations of health inequality monitoring, different thresholds may be applicable.
Thresholds used to define low, moderate or high inequality are largely contextual and depend on the indicator and national policy and programming factors.
Consult other forms of evidence
Identification of high-priority areas for action on its own does not fully explain the results or result in the identification of the solutions that could be applied. The findings of inequality monitoring should be contextualized alongside other scientific evidence about the situation of inequality, knowledge about the lived realities of the affected population, and knowledge about the broader monitoring context and WHO guidelines and recommendations. This allows for consideration of a larger breadth of complementary evidence beyond what is captured in the immediate inequality analysis.
Consulting other forms of evidence can shed light on confusing or incomplete findings, the importance of the findings, local factors, root causes of inequalities (or inequities), and possible inroads for remedial actions. Quantitative analyses may be useful to answer questions exploring causation and associations, patterns and trend. Qualitative studies may help to delve into questions of how and why inequalities are observed, including understanding the underlying context and lived experiences. Mixed methods is “research in which the investigator collects and analyses data, integrates the findings, and draws inferences using both qualitative and quantitative approaches or methods in a single study or programme of inquiry” (3). The WHO Handbook for conducting assessments of barriers to effective coverage with health services elaborates on how health inequality monitoring can feed into a mixed-methods assessment of the reasons behind differences in service coverage (4).
Engagement with existing literature about inequalities should begin at the planning stages of health inequality monitoring, and be done continuously as monitoring is conducted, including when assessing and applying the analysis results. A review of existing qualitative and quantitative and mixed-methods studies may entail conducting structured syntheses, such as modelling, literature reviews or evidence and gaps maps (see Chapter 3). Joint displays that show qualitative evidence alongside quantitative results may be particularly useful for integrating different types of information at the planning, implementation and presenting stages of monitoring (5).
Sometimes inequality analyses lead to new questions for which further study or assessments may be required. After considering the strengths and limitations of various evidence sources, possible explanations can be assessed and areas for further exploration can be identified. See Chapter 16 for more on the use of emerging and novel data sources, and Chapter 25 on further quantitative approaches and measures to explore inequalities and their drivers.
Situations in which evidence is insufficient (e.g. due to a lack of data or lack of analysis or synthesis of existing data) may lead to recommendations for extended monitoring activities. Some of the main conclusions in the WHO State of inequality: HIV, tuberculosis and malaria report (2), for example, pointed to the need for more and better data to address gaps in inequality monitoring for these diseases, and the need for regular inequality analysis and reporting to track changes over time (6).
Cross-check interpretations
The actions that are ultimately taken to address health inequalities should be based on a thorough assessment of available evidence to ensure the results are accurately interpreted and contextualized. The contents of this book endeavour to provide a foundational knowledge base of issues pertaining to health inequality monitoring, including critically evaluating the results and conclusions derived from the process. The issues highlighted below represent some general considerations that should be cross-checked when considering how to prioritize and use evidence to inform actions to improve population health. This is not a comprehensive overview of all possible issues that may be pertinent when interpreting the results.
An initial consideration is whether subgroups mask inequalities. As described in Chapter 17, decisions about the composition and number of subgroups for inequality monitoring have implications for the analysis and interpretation of inequalities. When subgroups are constructed in a way that includes heterogeneous populations, they may mask inequalities. For example, monitoring inequalities by subnational region where multiple provinces or districts are grouped together can mask the situation in single provinces or districts. An approach that defines rural areas broadly may not capture the situation in remote communities within the rural area or rural communities that are located in close proximity to large urban centres (see Chapter 5). Data triangulation from different sources can help explore how the categorization of inequality dimension may be masking the realities contained within a subgroup and prompt more nuanced analysis.
The limitations in the underlying health information system may factor into the conclusions of inequality monitoring. For example, underreporting for certain indicators may occur due to human resource shortages; information, communication and technology and computerization deficits; and inadequate coordination for information flows. Underdiagnosis of some diseases (e.g. tuberculosis, HIV, neglected tropical diseases, hypertension and some other noncommunicable diseases) tends to be higher in areas or subgroups experiencing disadvantage. Furthermore, the absence of reliable information on population counts makes it challenging to estimate denominator values (see Chapter 13). Thus, data about the number of people with a given disease or condition may not be able to capture the extent to which underdiagnosis and unmet need are happening for a given subpopulation or area.
Attention is warranted to ensure the population from which data were collected corresponds to the population where action is to be implemented. When considering evidence for action, the geographical unit for which the data are available should ideally match the scope at which programming occurs. For example, data that reflect the situation in a province would be more appropriate for use to inform provincial-level programming than data collected from a national sample. Often, however, this correlation is suboptimal. When conducting inequality analysis for uptake by subnational governments, attention to getting to the lowest geographical unit possible for which disaggregated findings are still statistically relevant is important.
Inequality monitoring as part of larger assessments and review processes
Inequality monitoring may be integrated as part of larger assessments and review processes in the health sector. For example, the results of inequality monitoring yield evidence and inputs into several aspects of the WHO Innov8 approach for reviewing national health programmes to leave no one behind (7). This approach provides a systematic and comprehensive eight-step method for integrating the results of health inequality monitoring into national health programmes. It involves a comprehensive review of national health programming and related evidence by a multidisciplinary team, as detailed in the technical handbook for the approach. The results of inequality analyses are used in the Innov8 first step of understanding the baseline of the programme through a diagnostic checklist; the third step of identifying who is being left out of the programme; and the eighth step of strengthening monitoring and evaluation practices, including through the incorporation and/or strengthening of health inequality monitoring on a routine basis.
The Innov8 approach has been adapted and applied across several countries and different contexts, and the approach continues to evolve (8). For example, it was used in Indonesia as part of a review of national neonatal and maternal health action plans, and efforts to expand their equity orientation, rights basis and gender responsiveness, and to address social determinants of health (9). The review process helped to bolster demand for health inequality monitoring and its use in planning. In Nepal, the Innov8 approach helped to identify populations not covered by the national Adolescent sexual and reproductive health programme (10).
WHO has been advancing mixed methods to unpack barriers to health services (4, 11). Health inequality monitoring is part of these mixed-methods approaches, which also draw on a desk review, key informant interviews and focus groups and apply data triangulation techniques across all sources. The assessments can be done as parallel convergent or explanatory sequential. For example, in 2017, the Nigerian Government decided to revise and update its policy on the health and development of adolescents and young people in Nigeria. The Government commissioned a national situation analysis to inform the update, and a barrier assessment (including health inequality monitoring) was conducted to complement the analysis (12). The findings of the assessment fed into the development of a new adolescent health policy and related strategic plan.
References
1. Evidence, policy, impact. WHO guide for evidence-informed decision-making. Geneva: World Health Organization; 2021 (https://iris.who.int/handle/10665/350994, accessed 12 June 2024).
2. State of inequality: HIV, tuberculosis and malaria. Geneva: World Health Organization; 2021 (https://iris.who.int/handle/10665/350198, accessed 5 June 2024).
3. Tashakkori A, Creswell JW. The new era of mixed methods. J Mixed Methods Res. 2007;1(1):3–7. doi:10.1177/2345678906293042.
4. Handbook for conducting assessments of barriers to effective coverage with health services: in support of equity-oriented reforms towards universal health coverage. Geneva: World Health Organization; 2024 (https://iris.who.int/handle/10665/377956, accessed 1 July 2024).
5. Guetterman T, Fetters M, Creswell J. Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Ann Fam Med. 2015;13(6):554–561. doi:10.1370/afm.1865.
6. Hosseinpoor AR, Bergen N, Kirkby K, Schlotheuber A, Fuertes CV, Feely SM, et al. Monitoring inequalities is a key part of the efforts to end AIDS, tuberculosis, and malaria. Lancet. 2022;399(10331):1208–1210. doi:10.1016/S0140-6736(21)02756-2.
7. Innov8 approach for reviewing national health programmes to leave no one behind: technical handbook. Geneva: World Health Organization; 2016 (https://iris.who.int/handle/10665/250442, accessed 12 June 2024).
8. Zamora G, Koller TS, Thomas R, Manandhar M, Lustigova E, Diop A, et al. Tools and approaches to operationalize the commitment to equity, gender and human rights: towards leaving no one behind in the Sustainable Development Goals. Glob Health Action. 2018;11(Suppl. 1):75–81. doi:10.1080/16549716.2018.1463657.
9. Koller TS, Saint V, Floranita R, Koemara Sakti GM, Pambudi I, Hermawan L, et al. Applying the Innov8 approach for reviewing national health programmes to leave no one behind: lessons learnt from Indonesia. Glob Health Action. 2018;11(Suppl. 1):1423744. doi:10.1080/16549716.2018.1423744.
10. Adolescent sexual and reproductive health programme to address equity, social determinants, gender and human rights in Nepal: report of the pilot project. New Delhi: World Health Organization Regional Office for South-East Asia; 2017 (https://iris.who.int/handle/10665/259549, accessed 5 June 2024).
11. Handbook for conducting an adolescent health services barriers assessment (AHSBA) with a focus on disadvantaged adolescents. Geneva: World Health Organization; 2019 (https://iris.who.int/handle/10665/310990, accessed 12 June 2024).
12. Assessment of barriers to accessing health services for disadvantaged adolescents in Nigeria. Brazzaville: World Health Organization Regional Office for Africa; 2019 (https://iris.who.int/handle/10665/324926, accessed 30 May 2024).