The International Conference on Population and Development (ICPD) in 1994 gave us the Programme of Action as a conceptual map for where we wanted to go in the next 20 years; but as a roadmap, it was much vaguer. How would we know if we were covering any distance on that road?

International donors, national policy makers and programme managers would want to measure progress and achievements in sexual and reproductive health. But the ICPD definition was too comprehensive to measure, explain or communicate easily to the common person, let alone to the millions of women and men whose well-being was the target. Indeed, how does one know if one has good reproductive health?

For health and social science researchers, moving into this new frontier over the past decade has brought exhilaration—in stretching our knowledge base and skills—and also consternation over the absence of clear indicators for reproductive health. The major stakeholders have achieved only limited consensus on a short list of key indicators, so new ones have had to be established and tested.

Indicators were needed to identify points of progress on a range of health and development issues, such as having a trained health attendant at delivery, using a condom to avoid pregnancy or to prevent sexually transmitted infections (STI), or noting increases in women's empowerment and autonomy. Indicators also had to work with different populations, such as adolescents, males and core STI transmitters.

Horizons opened up
Perhaps the most beneficial outcome of ICPD from an evaluation perspective was that it forced many researchers to broaden their knowledge. We needed to understand the specifics not only of reproductive physiology, but also gender dynamics, infectious disease, social context and health service factors related to sexual and reproductive health. Researchers trained in the social sciences and demography have had to learn to converse with epidemiologists who study the causes and prevention of disease. Both have struggled to improve population-level monitoring of health outcomes.

ICPD has required everyone to think holistically about individual health because its issues cut across areas of specialty. We now see how early sexual debut can raise the risk of HIV/AIDS infection. We appreciate how iron, folic acid and other micronutrients can help pregnant women resist infection. We notice how community awareness of the symptoms of pregnancy-related complications might prompt emergency transport for mothers to maternity clinics. We see that postabortion care and voluntary HIV testing centres should include family planning counselling and services.

But with all the increases in our technical knowledge, health and social science researchers remain acutely challenged to evaluate the successes and failures in reproductive health. Tracking change requires adequate data, measurements and estimates, which are not always available.

Assessing the assessors
Some gains are worthy of note. For example, in the Demographic and Health Survey and Reproductive Health Survey programmes for female, male and health facility respondents, the health content has expanded and is better focused.¹ Biomarkers for HIV infection, anaemia and nutritional status, as well as geographic or spatial data, are increasingly collected in population-based household surveys.²

We also have seen the use of sophisticated measures such as Disability-Adjusted Life Years (DALYs) to assess the burden of disease from poor reproductive health. Methods and tools to estimate outcomes for entire populations have also improved, whether for maternal mortality, HIV prevalence or abortion.³

Do these indicators manage to inform us whether sexual and reproductive health has improved since 1994? Not necessarily, even though our capacity to make assessments has been strengthened.

Narrowing our choices
By embracing the full-faceted concept of reproductive health, we face a large and complex group of indicators. We therefore have the responsibility of rendering them into a manageable and meaningful set suitable for national monitoring systems. The World Health Organization's short list of 17 indicators⁴ and Population Action International's composite reproductive risk index⁵, based on ten key indicators, currently lead the field. Both are heavily focused on health and do not include measures of gender equality in other development areas, such as education, employment, income or legal rights.

One could argue that improving reproductive health is conceptually akin to economic development. They each involve the well-being of individuals and societies, and both are subject to many systems of influences. Thus both need many kinds of indicators to monitor progress.

Where have we failed?

• We remain wedded to large-scale national surveys, neglecting to develop the capacity of routine health information systems that could benefit community-level users much earlier, even before the data are aggregated for national officials.
• We closely monitor individual reproductive health outcomes, but neglect to adequately monitor the quality and use of programme services. We tend to assume these services do influence outcomes significantly, but we do not devote enough resources or patience to evaluating them at any level—national, community or programme.
• We have inundated the field with too many indicators, many measured in several different ways for the same topic. We have no common data system for ensuring their calculation at one point in time, let alone at multiple points. In the rush to identify performance measures, we have paid less attention to our ability to understand causal pathways and links, so that we have chosen indicators that are unconnected to each other.
• We do not agree on a core set of indicators, one that has empirical coherence and can be communicated to the policy makers and financial authorities who allocate resources. Meanwhile, other global health initiatives—for HIV/AIDS, vaccine development, malaria, tuberculosis, micronutrients and cervical cancer—tap the same fixed resources and have the advantage of focused interventions.

Reading the signs
Progress in reproductive health should be closely tied to progress toward the Millennium Development Goals (MDGs)—especially in terms of poverty reduction—as it influences the achievement of most, if not all, of the eight goals. If the ICPD had defined eight goals for improving reproductive health, we would not be struggling with the monitoring and evaluation challenge raised by the Cairo definition. However, a challenge is also an opportunity: reproductive health monitoring can and should be integrated into MDG monitoring.

The Gates Foundation and others are dedicating significant resources to tracking MDG progress and the contributions of global health funding. We must ensure that key reproductive health indicators are included in the set of indicators for each relevant MDG, and emphasise the importance of reproductive health investments in reducing health inequities. We should also support the needed data systems and the acquisition of rigorous research evidence. This work may then allow the true developmental agenda of ICPD to triumph.

Amy Tsui is Director of the Bill and Melinda Gates Institute for Population and Reproductive Health, and Professor of Population and Family Health Sciences, Johns Hopkins Bloomberg School of Public Health.

> Next section: Report Card Database

1. Demographic and Health Surveys,
   http://www.measuredhs.com/ aboutsurveys/dhs_surveys.cfm (accessed 14 May 2004 ); and Center for Disease Control and Prevention,
   http://www.cdc.gov/reproductivehealth/ logistics/
   global_rhs.htm#About%20DRH%20Surveys (accessed 14 May 2004 ).
2. J.T. Boerma and others, “Measurement of Biomarkers in Surveys in Developing Countries: Opportunities and Problems”, Population and Development Review 27 (2001): 303-314; and T. Ricketts, “Geographic Information Systems and Public Health”, Annual Review of Public Health 24 (2003): 1-6.
3. W. Graham and others, “Estimating Maternal Mortality Using the Sisterhood Method”, Studies in Family Planning 20 (1989): 125-35; The Alan Guttmacher Institute, Sharing Responsibility: Women, Society and Abortion Worldwide (New York, NY, AGI, 1999); and B. Zaba and others, “Adjusting HIV Prevalence Data Collected in Antenatal Clinics to Obtain Estimates for the General Female Population”, AIDS 14 (2000): 2741-2750.
4. World Health Organization, Global Monitoring and Evaluation Database, http://www.who.int/reproductive-health/global_monitoring/database.html (accessed 14 May 2004 ).
5. Population Action International, Data Sources and Methodology, http://www.populationaction.org/resources/publications/
   worldofdifference/rr2_methodology.htm (accessed 14 May 2004 ).