Using multi-country household surveys to understand who provides reproductive and maternal health services in low- and middle-income countries: a critical appraisal of the Demographic and Health Surveys.

OBJECTIVE: The Demographic and Health Surveys (DHS) are a vital data resource for cross-country comparative analyses. This study is part of a set of analyses assessing the types of providers being used for reproductive and maternal health care across 57 countries. Here, we examine some of the challenges encountered using DHS data for this purpose, present the provider classification we used, and provide recommendations to enable more detailed and accurate cross-country comparisons of healthcare provision. METHODS: We used the most recent DHS surveys between 2000 and 2012; 57 countries had data on family planning and delivery care providers and 47 countries had data on antenatal care. Every possible response option across the 57 countries was listed and categorised. We then developed a classification to group provider response options according to two key dimensions: clinical nature and profit motive. RESULTS: We classified the different types of maternal and reproductive healthcare providers, and the individuals providing care. Documented challenges encountered during this process were limitations inherent in household survey data based on respondents' self-report; conflation of response options in the questionnaire or at the data processing stage; category errors of the place vs. professional for delivery; inability to determine whether care received at home is from the public or private sector; a large number of negligible response options; inconsistencies in coding and analysis of data sets; and the use of inconsistent headings. CONCLUSIONS: To improve clarity, we recommend addressing issues such as conflation of response options, data on public vs. private provider, inconsistent coding and obtaining metadata. More systematic and standardised collection of data would aid international comparisons of progress towards improved financial protection, and allow us to better characterise the incentives and commercial nature of different providers.

Measuring maternal mortality: an overview of opportunities and options for developing countries.

BACKGROUND: There is currently an unprecedented expressed need and demand for estimates of maternal mortality in developing countries. This has been stimulated in part by the creation of a Millennium Development Goal that will be judged partly on the basis of reductions in maternal mortality by 2015. METHODS: Since the launch of the Safe Motherhood Initiative in 1987, new opportunities for data capture have arisen and new methods have been developed, tested and used. This paper provides a pragmatic overview of these methods and the optimal measurement strategies for different developing country contexts. RESULTS: There are significant recent advances in the measurement of maternal mortality, yet also room for further improvement, particularly in assessing the magnitude and direction of biases and their implications for different data uses. Some of the innovations in measurement provide efficient mechanisms for gathering the requisite primary data at a reasonably low cost. No method, however, has zero costs. Investment is needed in measurement strategies for maternal mortality suited to the needs and resources of a country, and which also strengthen the technical capacity to generate and use credible estimates. CONCLUSION: Ownership of information is necessary for it to be acted upon: what you count is what you do. Difficulties with measurement must not be allowed to discourage efforts to reduce maternal mortality. Countries must be encouraged and enabled to count maternal deaths and act.

Postpartum return of ovarian activity in nonbreastfeeding women monitored by urinary assays.

Assays of first morning urine samples for pregnanediol-3 alpha-glucuronide (PdG), estradiol-17 beta-glucuronide (E2G), and LH were used to monitor endocrine function in 16 regularly cycling women and 22 postpartum nonbreastfeeding women. Twice weekly blood samples were also obtained from the postpartum group. Ovulation was inferred by a significant rise in LH and PdG, and reversal of the E2G to PdG ratio. Luteal phase PdG excretion was measured by the peak of smoothed PdG levels and the area under the smoothed luteal phase PdG curve. The lower limits of normal established in 16 cycling women were a peak luteal phase PdG of 4 micrograms/ml and an area under the PdG curve of 20 micrograms/ml. In the postpartum women, 32% of first cycles were anovulatory, and among ovulatory cycles, 73% had abnormally low luteal phase PdG excretion or short luteal phases. In second and subsequent cycles, 15% were anovulatory and 26% had luteal phase abnormalities. There was a progressive increase in luteal PdG excretion from the first to third cycles. The mean delay before first ovulation was 45.2 days, and no woman ovulated before 25 days after delivery. The correlations between blood and urinary hormone levels were 0.78 for PdG, 0.65 for E2G, and 0.55 for LH. We conclude that assays of daily early morning urine samples provide reliable information on ovulation and luteal phase adequacy, and that there is gradual recovery of pituitary ovarian function after parturition.

Maternal health and the measurement trap.

This paper focuses on the contribution of measurement-related factors to the neglect of maternal health in resource allocation for programmes and in public health research. As the recent interest in maternal health has now progressed beyond the need for information primarily for the purpose of advocacy, measurement-related factors have emerged as powerful constraints on programme action. Three outstanding needs for information can be identified: first, to establish the levels and trends of specific maternal health outcomes; secondly, to identify the characteristics and determinants of health outcomes; and thirdly, to monitor and evaluate the effectiveness of programmes designed to influence health outcomes. In order to meet these needs, the emphasis placed on operational research by the current major initiatives in maternal health must be complemented by an equivalent emphasis on methodological studies. The call for improved information by international and national agencies should be made in unison with the call for action. Inadequate information is a reality that has to be faced throughout the world, but particularly in developing countries. The quality, quantity and scope of health-related data are the elements of this inadequacy and may be discussed in terms of four factors: the indicators, the data sources, the measurement techniques, and the conceptual framework. In this paper, the neglect of maternal health and the lack of information are shown to be self-reinforcing and constitute a measurement trap sprung by these four factors. Dismantling this trap has revealed a weak conceptual framework to lie at the very centre. Maternal health has tended to be conceptualized as a discrete, negative state, characterized by physical rather than social or mental manifestations, and by a narrow time-perspective focusing on pregnancy, delivery and the puerperium. The need to broaden this perspective and to develop equally broad operational definitions represent important steps forward that must be taken.

Women's hidden transcripts about abortion in Brazil.

Two folk medical conditions, "delayed" (atrasada) and "suspended" (suspendida) menstruation, are described as perceived by poor Brazilian women in Northeast Brazil. Culturally prescribed methods to "regulate" these conditions and provoke menstrual bleeding are also described, including ingesting herbal remedies, patent drugs, and modern pharmaceuticals. The ingestion of such self-administered remedies is facilitated by the cognitive ambiguity, euphemisms, folklore, etc., which surround conception and gestation. The authors argue that the ethnomedical conditions of "delayed" and "suspended" menstruation and subsequent menstrual regulation are part of the "hidden reproductive transcript" of poor and powerless Brazilian women. Through popular culture, they voice their collective dissent to the official, public opinion about the illegality and immorality of induced abortion and the chronic lack of family planning services in Northeast Brazil. While many health professionals consider women's explanations of menstrual regulation as a "cover-up" for self-induced abortions, such popular justifications may represent either an unconscious or artful manipulation of hegemonic, anti-abortion ideology expressed in prudent, unobtrusive and veiled ways. The development of safer abortion alternatives should consider women's hidden reproductive transcripts.

Institutional maternal mortality in Mali.

OBJECTIVES: To measure the institutional maternal mortality ratio (MMR) in Mali and suggest ways to reduce it. METHODS: Routinely recorded data from 24 health institutions in three regions were reviewed for 1988 to 1992. RESULTS: The overall MMR in the institutions was 201 maternal deaths per 100,000 live births. Hemorrhage, toxemia and infections accounted for 80% of the 360 recorded maternal deaths, almost all of which were preventable. The main reasons why these conditions result in death lie in poor quality and maldistribution of health services, lack of transport and late use of allopathic services. CONCLUSIONS: Maternal mortality is still a major public health problem in Mali, even among the small proportion of women who reach health facilities. Substantial new initiatives are urgently needed to reduce this major cause of preventable adult female mortality.

Erratum to "Maternal, neonatal and child health interventions and services: moving from knowledge of what works to systems that deliver" [International Health 2 (2010) 87-98].

The Publisher regrets that an error occurred in the name of the 6th listed co-author for this paper. B. Matthias was listed in the original paper instead of M. Borchert; the correct listing can be seen above.

Menstrual dysfunction: a missed opportunity for improving reproductive health in developing countries.

The field of population has undergone a paradigm shift to a broader focus on reproductive health, which recognises women's self-perceived health needs. Investigations in various countries reveal that menstruation is a primary concern of women. Yet sparse attention has been paid to understanding or ameliorating women's menstrual complaints. We propose including the management of menstrual complaints as part of reproductive health programming. Next steps should include further quantitative and qualitative research to understand the prevalence, determinants and consequences of menstrual dysfunction; developing appropriate protocols and low-cost interventions for diagnosis and treatment of menstrual morbidity and training of health care workers in resource-scarce settings; and developing educational interventions to facilitate women's understanding of normal menstrual function and variability as well as of the types, causes and appropriate treatments for menstrual dysfunction.

The impact of eliminating sterility on population growth.

PIP: The issue of this discussion is whether and to what extent treatment of infertility will contribute to a significant increase in fertility, and whether treatment of sexually transmitted diseases and infertility is justified as a humane effort to alleviate suffering, death, and the extent of impact on men, women, fetuses, and infants. The Brunham model produces estimates of the difference in the total fertility rate (TFR) due to sterility of between 9.7 and 18.0. However estimates with more realistic assumptions about life expectancy, menarche, menopause, and union formation yield TFR differences between 2.1 and 2.5. Accounting for secondary sterility and a more realistic simulation yields a change in the population growth rate of 0.8% and 0.9% compared to 2.5% and 1.9% in the Brunham model. The conclusion is that population growth rates would not be very likely to increase if sterility is eliminated. The Brunham model is an important attempt to model the effects of different assumptions about rates of partner exchange on disease prevalence and on sterility levels. The Brunham conclusion is that population growth rates increase by 50% or more when gonorrhea has a prevalence of 20% and a 12% probability of sterility per 6-month duration of illness. Chlamydia-related secondary sterility would not have quite as large an effect. Many other models are possible because of the many other fertility inhibiting factors. The Brunham model does not account for any age variation in fertility and mortality rates, which is not a realistic view of human behavior. The Brunham model also uses a very high mortality pattern, comparable to TFRs of 8.5 to 9.5 for a noncontracepting Hutterite natural fertility population, with short breast feeding, and universal marriage. When fecundability, exposure to risk of intercourse, postpartum and lactational amenorrhea, and coital frequency are taken into account, the impact of sterility is reduced. The proposed model accounts for sterility and postpartum amenorrhea, union formation, and coital frequency with realistic ages of menarche and menopause.^ieng

Characteristics and determinants of postpartum ovarian function in women in the United States.

OBJECTIVE: Our aim was to characterize the influence of breast-feeding on the postpartum return of ovarian activity for the purpose of assessing the relevance of the lactational amenorrhea method of contraception for women in the United States. STUDY DESIGN: Twenty-two non-breast-feeding and 60 breast-feeding women from Baltimore collected daily urine specimens that were assayed to determine ovulation and luteal phase adequacy. Vaginal bleeding was ascertained weekly, and breast-feeding women recorded infant feeding daily. Proportion-hazards models were used to relate measures of breast-feeding to the occurrence of ovulation. RESULTS: Two thirds of women ovulated before their first vaginal bleeding, but 47% of those cycles had decreased luteal-phase pregnanediol excretion. Breast-feeding frequency and suckling duration were significant predictors of the risk of ovulation (p < 0.001). Supplementation with bottle feeding was associated with a reduction in breast-feeding. CONCLUSION: A high degree of protection from pregnancy can be achieved using breast-feeding frequency and suckling duration, even with supplemented breast-feeding.

PIP: Between 1983 and 1987 in Baltimore, Maryland, researchers followed 60 breast-feeding mothers (55% single and 82% black) and 22 non breast feeding mothers (3% single and 93% white) for 72 weeks to better understand resumption of postpartum ovarian activity and the impact breast feeding has on resumption of ovarian activity. Laboratory personnel conducted daily urinary assays to detect ovulation and to evaluate the luteal phase. Ovulatory first vaginal bleeding episodes increased with the time (45% in the first 12 weeks to 100% after 49 weeks). 69% of breast-feeding mothers experienced an ovulatory first bleeding episode, but 47% of these cycles had reduced excretion of pregnanediol during the luteal phase, indicating a reduced likelihood of conception before first bleedings. The women's descriptions of perceived characteristics of their bleeding episodes were linked to ovulatory or anovulatory status. Specifically, women who reported regular or heavy bleeding were more likely to have had a preceding ovulation than those who reported spotting or light bleeding (84% vs. 35%; odds ration = 9.8; p .001). Resumption of ovulation was delayed in breast-feeding mothers. All non breast-feeding mothers menstruated within the first 12 weeks postpartum compared with just 20% of breast- feeding mothers. First ovulation occurred on average 45 days after delivery among non breast-feeding mothers and 189 days among breast- feeding mothers. The multivariate analysis revealed that breast-feeding episode contributed greatly and independently to the delay in ovulation (relative risk = .66 and .91, respectively; p .001). The protected effected of breast feeding remained even when mothers supplemented with bottle feeds 92% pregnancy rate over 6 months for both exclusive and partial breast feeders). These results indicated that breast feeding of increased frequency and suckling duration provides considerable protection from pregnancy, even when mothers use supplements.

The Maternal and Neonatal Programme Effort Index (MNPI).

A new index is presented that measures the effort levels of national programmes to reduce maternal and neonatal mortality. These indices come from a questionnaire instrument composed of 14 major headings and 81 items. Forty-nine countries including most of the population in each geographical region are covered. Data were collected from 10 to 25 raters in each country, who rated the 81 items for both the current year and 3 years ago, using a 0-5 scale running from no adequacy to full adequacy. The raters were drawn from a variety of positions and backgrounds, and were identified by a consultant retained in each country for that purpose. On average, country programmes score at about half of the maximum score, but this varies considerably across the 14 components of effort, from very low scores for access to treatment by rural women, to high scores for neonatal care. Regional averages are not far apart for the overall score, although South Asia scores are especially low and East Asia's are especially high. To a considerable extent regions agree in the relative stress they give to each of the 14 components. Over the 3-year period, average scores rose by about 10%. When countries are divided into three groups by their maternal mortality levels, most of the 14 components distinguish the high from the medium mortality countries; and about half of the components distinguish the medium from the low mortality countries. This new Maternal and Neonatal Programme Effort Index (MNPI) appears to yield useful measures for various dimensions of programme effort, and it relates sensibly to the output measure of maternal mortality, at least as it is currently measured.

Birth intervals and childhood mortality in rural Bangladesh.

This study investigates the relationship between birth intervals and childhood mortality, using longitudinal data from rural Bangladesh known to be of exceptional accuracy and completeness. Results demonstrate significant but very distinctive effects of the previous and subsequent birth intervals on mortality, with the former concentrated in the neonatal period and the latter during early childhood. The impact of short birth intervals on mortality, however, is substantially less than that found in many previous studies of this issue, particularly for the previous birth interval. The findings are discussed in terms of the potential for family planning programs to contribute to improved child survival in settings such as Bangladesh.

Postpartum depression: prevalence and determinants in Lebanon.

This study assesses the prevalence and determinants of postpartum depression (PPD). 396 women delivering in Beirut and a rural area (Beka'a Valley) were interviewed 24 hours and 3-5 months after delivery. During the latter visit, they were screened using the Edinburgh postnatal depression scale. The overall prevalence of PPD was 21% but was significantly lower in Beirut than the Beka'a Valley (16% vs. 26%). Lack of social support and prenatal depression were significantly associated with PPD in both areas, whereas stressful life events, lifetime depression, vaginal delivery, little education, unemployment, and chronic health problems were significantly related to PPD in one of the areas. Prenatal depression and more than one chronic health problem increased significantly the risk of PPD. Caesarean section decreased the risk of PPD, particularly in Beirut but also in the Beka'a Valley. Caregivers should use pre- and postnatal assessments to identify and address women at risk of PPD.

Media interventions to increase cervical screening uptake in South Africa: an evaluation study of effectiveness.

Successful cervical cancer prevention depends on reaching, screening and treating women with pre-invasive disease. We aimed to evaluate the effectiveness of two media interventions-a photo-comic and a radio-drama-in increasing cervical screening uptake. A randomized controlled trial compared a photo-comic on cervical cancer screening with a placebo comic. One month after the comics were distributed a radio-drama paralleling the photo-comic was broadcast on the community radio station and a retrospective evaluation was carried out. The trial was set in Khayelitsha, a peri-urban squatter community near Cape Town, South Africa. A random sample consisted of 658 women between the ages of 35 and 65 years, from a stratified sample of census areas. The main outcome measure was self-reported cervical screening uptake 6 months after distribution of the comics. Seven percent (18 of 269) of women who received the intervention photo-comic reported cervical screening during the 6 months follow-up, compared with 6% (25 of 389) of controls (P = 0.89). Women who recalled hearing the radio-drama were more likely to report attending screening (nine of 53, 17%) than those who did not (19 of 429, 4%; P < 0.001). We conclude that the photo-comic was ineffective in increasing cervical screening uptake in this population. The radio-drama may have had more impact, but only a minority of women recalled being exposed to it. Future research must concentrate not only on achieving high level of exposure to health messages, but also on investigating the links between exposure and action.

Risk of ovulation during lactation.

60 breastfeeding mothers in Baltimore and 41 in Manila recorded their infant feeding patterns daily, and gave additional information at weekly interviews. Ovarian activity was monitored by assays for hormone metabolites in daily urine samples. On average, women in Baltimore breastfed less often but for longer at each feed than women in Manila, and the mean times until ovulation were 27 and 38 weeks post partum. 41% of first ovulations had luteal phase defects. Anovular first menses were common (45.1%) during the first 6 months post partum but the rate fell greatly thereafter. The risk of ovulation was reduced by a higher frequency of breastfeeds, longer duration of each feed, and less supplementary feeding. During the first 6 months post partum, amenorrhoeic women had low risks of ovulation (below 10%) with partial breastfeeding, and exclusive breastfeeding reduced the risk to 1-5% with either frequent short feeds or infrequent longer feeds. However, if the woman started menstruating before 6 months post partum, or if she continued breastfeeding beyond 6 months, the risk of ovulation rose, and contraception would be needed.

PIP: The factors predicting the return of ovulation during lactation were investigated in 60 breast feeding mothers from Baltimore, Maryland, and 41 mothers from Manila, the Philippines. The women in Manila breast fed more frequently (11.0 mean daily feeds) than those in Baltimore (5.8 feeds) throughout the first postpartum year. However, during the first 6 months, the average length of an individual breast feed was longer among US women. Although both groups gave few supplementary feeds during the first 20 weeks after delivery, more non-bottle feeds were given thereafter in Manila. Use of bottle feeds was similar in both groups. Women who first ovulated before 6 months gave significantly fewer mean daily breast feeds during this period (8.5) than those who first ovulated after this point (10.7). Women who first ovulated before 6 months also had a lower percentage breast feeds of total feeds (84.2%) than women who first ovulated later (88.2%). The mean duration of amenorrhea was 25.3 weeks in Baltimore and 31.7 weeks in Manila; the average times before first ovulation were 27.0 weeks and 38.0 weeks, respectively. Multivariate analyses showed that the length of suckling per feed, the daily number of breast feeds, and the proportion of breast feeds were all significantly and independently associated with a low risk of ovulation. Up to 6 months postpartum, exclusive breast feeding reduced the risk of ovulation by 98-99% during amenorrhea and by 94-97% after anovular menstruation. Only amenorrheic women practicing exclusive breast feeding during the first 6 postpartum months can achieve a pregnancy rate below 2%. These findings suggest that contraceptive use is indicated among women who resume menstruating before 6 postpartum months or continue breast feeding beyond 6 months.

Investigating induced abortion in developing countries: methods and problems.

Interest in abortion research is reemerging, partly as a result of political changes and partly due to evidence of the contribution of induced abortion to maternal mortality in developing countries. Information is lacking on all aspects of induced abortion, particularly methodological issues. This article reviews the methodological dilemmas encountered in previous studies, which provide useful lessons for future research on induced abortion and its complications, including related deaths. Adverse health outcomes of induced abortion are emphasized, because these are largely avoidable with access to safe abortion services. The main sources of information are examined, and their relevance for assessing rates of induced abortion, complications, and mortality is addressed. Two of the major topics are the problems of identifying cases of induced abortion, abortion complications, and related deaths, and the difficulties of selecting a valid and representative sample of women having the outcome of interest, with an appropriate comparison group. The article concludes with a discussion of approaches for improving the accuracy, completeness, and representativeness of information on induced abortion. Although the prospects for high-quality information seem daunting, it is essential that methodological advances accompany program efforts to alleviate this important public health problem.

PIP: Studies on induced abortion started dwindling after 1984 when the US government stopped financing abortion-related activities. Recently much interest has focused on induced abortion because it is a major cause of maternal mortality in developing countries. A summary of methodological studies is provided. The sources of data for such studies are somewhat deficient official records of legal abortions. In most developing countries abortion is illegal, and scrutiny of hospital records on complication (a 49% rate in a study in Latin America and 46% hospitalization) is a source. Induced abortion surveys of women in reproductive age and retrospective household surveys are other approaches. The World Fertility Survey, a population-based survey, missed 20-50% of spontaneous abortions and even more induced abortions. Expensive prospective studies requiring large samples did not always provide more accurate information either, e.g., in a Sao Paulo, Brazil, study 25% of 1801 women were lost to follow-up. Pregnant women often do not admit their condition. The measurement of complication rates from induced abortions requires examination of the often incomplete and inconsistent records of emergency, surgery, intensive care, pathology, and anesthesia wards and morgue registers. Some women never go to hospitals or are cared for by traditional healers. The measurement of induced abortion mortality could rely on data of vital registration systems (often shoddy), health service records, and community-based surveys. In a 1967 Latin American study, 33% of deaths were misclassified, and 53% were attributed to circulatory causes in an Egypt study. Abortion case identification is confounded by unintentional (16-83% of menstrual regulations are done on nonpregnant women) and intentional reporting errors (50% of those getting an abortion did not report it in Hungary in 1978 according to WHO data).