Corresponding Author Nicholas M. Fisk, University of Queensland Centre for Clinical Research, RBWH Campus, Herston, Brisbane 4029, Queensland, Australia. Tel.: 61-7- 3346 5300; E-mail: email@example.com
Maternal and perinatal disease accounts for nearly 10% of the global burden of disease, with only modest progress towards achievement of the Millennium Development Goals. Despite a favourable new global health landscape in research and development (R&D) to produce new drugs for neglected diseases, R&D investment in maternal/perinatal health remains small and non-strategic. Investment in obstetric R&D by industry or the not-for-profit sector has lagged behind other specialties, with the number of registered pipeline drugs only 1–5% that for other major disease areas. Using a Delphi exercise with maternal/perinatal experts in global and translational research, we estimate that equitable pharmaceutical R&D and public sector research funding over the next 10–20 years could avert 1.1% and 1.9% of the global disease burden, respectively. In contrast, optimal uptake of existing research would prevent 3.0%, justifying the current focus on health service provision. Although R&D predominantly occurs in high-income countries, more than 98% of the estimated reduction in disease burden in this field would be in developing countries. We conclude that better pharmaceutical and public sector R&D would prevent around 1/3 and 2/3, respectively, of the disease burden addressable by optimal uptake of existing research. Strengthening R&D may be an important complementary strategy to health service provision to address global maternal and perinatal disease burden.
Point de vue: Réponse relative et absolue à la charge de morbidité mondiale de la santé maternelle et périnatale par l’investissement dans la Recherche et Développement
Les maladies maternelles et périnatales représentent près de 10% de la charge mondiale de morbidité, avec seulement de modestes progrès vers la réalisation des Objectifs pour le Développement du Millénaire. Malgré un nouveau paysage favorable de la santé mondiale dans la Recherche et Développement (R&D) pour la production de nouveaux médicaments pour les maladies négligées, les investissements en R&D dans la santé maternelle/périnatale restent faibles et non-stratégiques. L’investissement dans la R&D obstétricale par l’industrie ou par le secteur sans but lucratif est à la traîne d’autres spécialités, avec le nombre de médicaments en vue répertoriés, représentant seulement 1 à 5% celui d’autres domaines de maladies importantes. En utilisant un exercice Delphi avec des experts de la santé maternelle/périnatale dans la recherche mondiale et translationnelle, nous estimons qu’un financement équitable de la R&D pharmaceutique et pour la recherche dans le secteur public au cours des prochaines 10 à 20 ans pourrait éviter respectivement 1,1% et 1,9% de la charge mondiale de morbidité. En revanche, l’absorption optimale de la recherche existante permettrait d’en éviter 3,0%, ce qui justifie l’accent mis actuellement sur la prestation des services de santé. Bien que la R&D s’effectue principalement dans les pays à revenus élevés, plus de 98% de la réduction prévue de la charge de la maladie dans ce domaine serait dans les pays en développement. Nous concluons que l’amélioration de la R&D du secteur pharmaceutique et public permettrait d’éviter respectivement environ 1/3 et 2/3 de la charge de morbidité pouvant être allégée par l’absorption optimale de la recherche existante. Le renforcement de la R&D peut être une stratégie complémentaire importante pour la prestation des services de santé afin de répondre à la charge de morbidité maternelle et périnatale.
Punto de vista: Direccionalidad relativa y absoluta de la carga global de enfermedad en salud materna y perinatal por inversión en I+D
Las enfermedades maternas y perinatales son responsables de casi un 10% de la carga global de enfermedad, y solo se ha alcanzado un progreso modesto en la consecución de los Objetivos de Desarrollo del Milenio. A pesar del panorama favorable dentro de la investigación y el desarrollo en salud global (I+D) para producir nuevos medicamentos para las enfermedades huérfanas, la inversión en I+D en salud materna/perinatal continúa siendo poca y nada estratégica. La inversión de la industria o del sector sin ánimo de lucro en I+D obstétrico se ha quedado rezagada con respecto a otras especialidades, con solo un 1-5% de medicamentos en desarrollo registrados en comparación con aquellos para otras enfermedades. Mediante un ejercicio Delphi, con expertos en investigación global y transnacional en salud materna/peri-natal, hemos estimado que un I+D farmacéutico y una financiación de la investigación del sector público equitativos en los próximos 10-20 años podrían evitar respectivamente un 1.1% y 1.9% de la carga global de enfermedad. Por otro lado, una respuesta optimizada de la investigación actual prevendría un 3.0%, justificando el enfoque actual sobre la provisión de servicios sanitarios. Aunque el I+D ocurre principalmente en países desarrollados, mas del 98% de la reducción estimada de la carga de enfermedad en esta área estaría en países en vías de desarrollo. Concluimos que una mejoría en el sector de I+D, tanto farmacéutico como publico, prevendrían, respectivamente, entre 1/3 y 2/3 de la carga de enfermedad, con una respuesta optimizada de la investigación actual. El fortalecimiento del I+D podría ser una importante estrategia complementaria a la entrega de servicios sanitarios para remediar la carga de enfermedad materna y perinatal a nivel global.
Pregnancy results in a substantial burden of avoidable disease. The 2001 Global Burden of Disease (GBD) study estimated that maternal disorders accounted for 1.7% and perinatal disorders 5.9% of Disability Adjusted Life Years (DALYs) (Lopez et al. 2006a,b) – a figure that would be substantially higher if stillbirths were included (Lawn et al. 2009). Maternal and perinatal disease results in substantial economic burden, estimated at over US$15 bn a year (Gill et al. 2007).
The United Nations Secretary General identified Millennium Development Goal 5 (MDG-5), to reduce the maternal mortality ratio by three-quarters between 1990 and 2015, as the one with least progress (Ban 2010). The UN estimates that the maternal mortality ratio declined only 6% between 1990 and 2005. Although a more recent study was somewhat more optimistic, even it estimated the annual decline since 1980 as only 1.8%, well below the 5.5% required to achieve MDG-5 (Hogan et al. 2010). The situation is slightly better for childhood mortality, much of which is attributable to deaths soon after birth, but the 22% decline in deaths among children under five since 1990 remains far below that required to achieve MDG-4, to reduce deaths in the under-5 s by two-thirds between 1990 and 2015 (United Nations Secretariat 2007). Yet while the scale of maternal mortality worldwide is now recognised, by far the greatest impact of poor health in pregnancy is on the child, with eight times as many child deaths in the first month of life (4 million) as maternal deaths, and an almost comparable number of stillbirths (3.2 million) (Stanton et al. 2006).
Approaches to reduce maternal and perinatal disease burden
Some countries have made substantial progress in reducing maternal and perinatal mortality. Factors associated with success include economic growth, measures of female empowerment (reduced total fertility, female literacy) and access to effective care (skilled birth attendants, physicians per capita) (Hogan et al. 2010; Robinson & Wharrad 2001). Case studies also highlight the importance of rapid access to emergency care, improvements in quality of care and reductions in unsafe abortions (Campbell et al. 2005; Chowdhury et al. 2007; Li et al. 2007; Yadamsuren et al. 2010). A recent review argues the case for facility-based intra-partum care as a pivotal strategy to reduce maternal mortality (Campbell & Graham 2006).
These findings have stimulated calls (Beaglehole & Bonita 2008; Campbell & Graham 2006; Marchal et al. 2009) to strengthen health systems in developing countries (Partnership for Maternal Newborn and Child Health 2007). Measures to alleviate poverty, improve education and expand access to effective health services are common to efforts to reduce most causes of disability and premature death. However, one topic frequently discussed in relation to other causes of disease burden, advances in treatment, has attracted little attention in the debate about how to reduce maternal and perinatal deaths. We speculated that there might be unrealised scope to develop new forms of diagnosis and treatment, using pharmaceuticals or technology, that could reduce further the burden of disease associated with childbirth. However, to determine whether this is so, our viewpoint must ask three questions.
The first question is whether there is a scarcity of new products in development to address disorders of pregnancy. This answer is clear as extended below. Certainly, the major obstetric diseases of preterm labour, pre-eclampsia and foetal growth restriction are still treated ineffectively, often in an atmosphere of therapeutic nihilism(Callaway et al. 2009). The second question is whether, if such products became available, they would impact visibly on maternal and perinatal mortality. Finally, the yield from such R&D efforts must be interpreted in the context of and complementary to current concerns about strengthening health service provision to address bottlenecks to scaling up critical interventions such as emergency intrapartum care (United Nations 2010). In this paper, we present estimates to address the second and third questions.
How much R&D is taking place in maternal health?
Drought of drugs in development
The pharmaceutical industry is responsible for more than half of all health research globally and produces around 20–25 new drugs per year (Bartlett 2007), but little in maternal-foetal health. We recently interrogated an industry database of over 37 000 drugs under development and found no new class of obstetric drug licensed since 1981 in either Europe or the USA, with little chance of one in the next decade (Fisk & Atun 2008).
We now extend this analysis in Table 1 with previously unpublished data from the Pharmaprojects database (http://www.pharmaprojects.com/) to categorise drugs under development by specialty area. There were 1636 drugs under development for neurological indications as of November 2007, 885 for musculoskeletal indications, 379 for genitourinary indications, but only 17 for maternal health indications. Obstetrics has only 1–5% of the pipeline of other mainstream specialties, fewer than for single orphan diseases such as Crohn’s disease or amyotrophic lateral sclerosis. This drug ‘drought’ relegates progress in obstetric therapeutics to the weaker and less well-resourced efforts of publicly funded investigator-driven research.
Table 1. Comparison of drugs under active development for maternal health indications compared with other mainstream disease areas (sourced from Pharmaprojects http://www.pharmaprojects.com/ as of November 2007)
Alimentary and metabolic
Blood and clotting
Poverty of public sector funding
Lack of R&D investment in maternal and perinatal health by the pharmaceutical sector is matched by underinvestment in the public sector. In a recent systematic review of research expenditure and grant allocations from national and international governmental and philanthropic funders, we documented comparatively low levels of investment in maternal/perinatal R&D with little evidence of capacity building (Fisk & Atun 2009). In high-income countries, typically only 1–3% of health research expenditure was on maternal/perinatal health. In low-resource settings, government funders focussed on infectious disease but not maternal and perinatal health despite the high disease burden, with little invested by global philanthropy. Clearly, decades of low political priority given to women’s health, the challenging nature of clinical research in pregnancy and the dearth of research capacity consequent upon chronic under-investment have taken their toll.
Minimal industrial R&D in maternal health coupled with chronic under-funding of academic research not only contributes to there being fewer new drugs for commercial translation, but retards development of diagnostic methods, devices and non-drug therapies, hampering progress towards reducing maternal/perinatal disease burden. Pregnant women thus miss out on many of the fruits of scientific advances realised in other fields.
Delphi exercise to estimate potential impact of investment in R&D on maternal and perinatal mortality
The previous sections demonstrate the paucity of investment in products to tackle disorders of pregnancy. But does this matter? How might the discovery of effective new products impact on maternal and perinatal mortality and how does this compare with measures to strengthen health systems? Answering this question is problematic in the face of imperfect information. Our discussions with key obstetric/neonatal and health R&D experts highlighted considerable divergence of opinion. We therefore undertook a Delphi process with selected international experts to obtain a consensus view.
Selection of experts
We surveyed research-active opinion leaders in maternal health, selected to encompass expertise in the major obstetric diseases, obstetric therapeutics, intrapartum care and perinatal aspects of neonatology, in both high- and low/middle-income geographical settings. Because the questionnaire explored the scope for innovation and implementation, preference was given (i) in the developing world to experts with links to international research involving the developed world and (ii) in the developed world to those with links to or experience in the developing world. Twenty experts working (non-exclusively) in the following areas were selected: four obstetric epidemiologists; 19 clinical professors; seven researchers active in pre-eclampsia, five in postpartum haemorrhage, six in preterm labour, three in growth restriction and three in birth asphyxia.
The modified Delphi process followed the usual principles of anonymity, feedback and iteration (Linstone & Turoff 2002). The questionnaire asked experts to estimate the percentage reduction in DALYs (see below) in (i) high- and (ii) low/middle-income countries expected to result from (a) better pharmaceutical investment, (b) better government and charitable investment in research into obstetric conditions and (c) optimal uptake of existing research within the current socioeconomic environment (control question). ‘Better’ was defined as equitable with other mainstream disease areas with high disease burden. A cover letter explained the rationale to determine the effect of better treatments and interventions, and the endpoints, and acknowledged they could take 10–20 years to have an impact. The letter and survey forms were modified with expert feedback over several drafts before issue (available from authors on request).
In Round One, 20 experts sent the questionnaire in late 2007 were asked to rank their responses in open-ended percentages. After two reminders, responses were collated and the discontinuous variable expressed non-parametrically as median and inter-quartile range (Murphy et al. 1998). Experts that replied were resent the questionnaire in Round Two and invited to modify their responses in the light of the anonymised group responses. Results were re-collated, and the degree of consensus established.
We used data on DALYs from the 2001 GBD study (Lopez et al. 2006a,b) which employs a limited number of maternal and neonatal headings. We excluded categories for which there are established treatments, even if inadequately used because of health service failings (e.g. sepsis, obstructed labour), to select six conditions that could potentially be treated with new therapeutic interventions (two maternal, two neonatal and two for stillbirths). The GBD 2001 study applies a 3% discount rate to future years of life lost (YLL) and years of life lost due to disability (YLD), with uniform weights to years lost at all ages. Disability weighting (DW) was according to the condition involved, (minimal DW for haemorrhage -typically 0.000 but 0.093 with severe anaemia-, DWs for low birth weight, < 2.5 kg (LBW) and birth asphyxia/trauma of 0.106 and 0.372, respectively, and no DW for hypertensive disease of pregnancy).
Unlike published GBD estimates, we included stillbirths, as they now number almost as many as neonatal deaths. These were defined as death of a foetus weighing ≥1000 g in the last third of gestation or ≥27 weeks and based on published estimates from WHO data for 2000 applied by the GBD 2001 authors to 2001 birth rates (Jamison et al. 2006). Because cause is unassigned, stillbirths are classified antepartum or intrapartum. We did not adjust for stillbirth age, so that a late stillbirth was valued the same as a preterm neonatal death and aligned with GBD 2001. Our measure is thus denoted as DALYs (3,0,1), where 3 indicates the discount rate used (3% in this case), 0 that there is no age weighting applied and 1 designating inclusion of stillbirths, as compared with the stillbirth-free GBD 2001 estimate of DALYs(3.0).
Relative and absolute scope to address each condition
Seventy per cent (14/20) of those contacted completed both rounds. Feedback from Round 1 (individual ratings and range of all respondents) was followed by a reduction in outlying ratings. Table 2 shows that investment in all categories offers at least comparable and usually greater scope to reduce existing DALYs in low/middle- than high-income settings for investment, with the exception of drugs for hypertension. For all six conditions, the relative payoff from optimal uptake of existing interventions was estimated to be greater than from investment in pharmaceutical or public sector research in low/middle-income settings but less clear in high-income settings, where both forms of investment reduced percentage disease burden from hypertension and LBW more than optimal uptake of existing interventions. Pharmaceutical investment held the greatest relative potential for hypertension, haemorrhage and LBW, but little for stillbirth and asphyxia. Increased government/charitable research was predicted to reduce disease burden from each cause by 10–20% in low/middle-income, and 8–13% in high-income settings.
Table 2. Scope to reduce DALYs from investment in different clinical conditions. Relative and absolute scope to reduce DALYs (3,0,1) (thousands) by disease and investment category. L/M = low/medium
DALYs(3,0,1) In thousands
DALYs, disability adjusted life years; SB, stillbirth.
To calculate absolute avertable burden, relative addressability was multiplied by DALYs (3,0,1) within each disease category, looking separately at high- and low/middle-income countries. We persisted with the GBD 2001 use of two large geographical areas, because GBD stillbirth data are only available in these categories and because seeking responses by individual country would have been unwieldy for the experts.
Although relative reductions were broadly similar, the reducible disease burden in each neonatal and stillbirth category was considerably higher than from the maternal conditions of hypertension and haemorrhage (Table 2). The greatest scope was with antepartum stillbirths (6.6, 13.2 and 19.7 m DALYs for increased pharmaceutical investment, increased public sector investment and optimal uptake of existing research, respectively).
Huge quantum of avoidable disease burden
Figure 1 shows the increasing gradient of avertable maternal, neonatal and stillbirth DALYs moving from pharmaceutical innovation to optimal uptake, and that >98% of avertable disease burden was in low/middle-income countries. Increased pharmaceutical research was estimated to prevent 17.5 m DALYs, and increased government/charitable funding nearly twice this burden at 31.5 m DALYs. However, the greatest effect was with optimal uptake of existing research, 49.8 m DALYs. Thus, better pharmaceutical and public sector research was estimated to yield 35% and 63%, respectively, of the potential reduction from optimal uptake of existing research.
Applying maternal and neonatal categories to the GBD 2001 denominator [1536.8 m DALYs (3,0)], which excluded stillbirths, suggests that better pharmaceutical and public sector research investment would avert 0.6% and 0.8% of global disease burden [DALYs (3,0)], respectively. However, adding stillbirths to GBD 2001 resulted in the more realistic figure of 1636.2 m DALYs (3,0,1)(Jamison et al. 2006), from which it can be estimated that better pharmaceutical and public sector research would avert 1.1% and 1.9% of global disease burden, respectively. The corresponding reduction from optimal uptake of existing research was estimated at 3.0%.
Validity of estimates
Because of uncertainty in estimating the benefits from better R&D, this Delphi exercise relied on empiric estimates of relative scope to address conditions, using a process of iteration among experts, unbiased by range disclosure and refined in response to their pooled opinions. To ensure robustness, we selected leading global experts with both low/middle- and high-income country experience and sufficiently research-active to be able to estimate both the probability of therapeutic advance, and its translatability. This necessarily limited the number of respondents, but just over half this number is considered sufficient to yield robust consensus opinion (Murphy et al. 1998).
The categories available within GBD and WHO stillbirth data sets are limited by poor reporting in under-resourced countries. LBW is a crude composite of prematurity and growth restriction, but is simpler and more accurate to record than gestational age or weight-for-gestation. LBW can be because of any of the three major obstetric conditions (preterm labour, pre-eclampsia, growth restriction), precluding addressability by individual disease. Lack of information on stillbirth is a further limitation. Given the participants’ expertise, we chose not to be more prescriptive than in the GBD categories and not to influence their views on percentage addressability with raw data. Where relevant however, the questionnaire gave insight into the breakdown of DALYs, such as for LBW for which the ratio of YLL and YLD differs between high- and low/middle-income countries. Although a maternal condition like hypertension could also result in stillbirth or LBW, the six disease categories assessed are mutually exclusive in their contributions to disease burden, justifying pooling of DALYs as in Figure 1.
The surprisingly high number of avertable DALYs in part reflects inclusion of stillbirths from 0.66 of gestation, now justified by their vast numbers. Because the years of life lost from a stillbirth is 30.4 years, the same as a neonatal death, the effect of including them in DALYs (3,0,1) is substantial, increasing overall disease burden by 6.6% compared to stillbirth-free GBD 2001 DALYs (3.0) (Jamison et al. 2006).
Prioritise R&D or health service provision?
Studies on the scope to reduce the global burden of maternal/perinatal disease have so far focussed on health service provision in low-income settings rather than R&D on innovative pharmaceutical products. The case for improving service provision is overwhelming, given evidence that many women have inadequate access to even basic care. Yet, the findings reported above suggest that increased pharmaceutical and public sector investment in R&D could achieve one and two-thirds the effect, respectively, of optimal uptake of existing research. On the one hand, the estimated effect of the three interventions above is not additive, as they overlap. On the other, they each address different issues; hence, it is unwise to assume that optimal uptake of existing research alone negates any contribution from co-investment in commercial or public sector R&D.
The scope for benefit was higher with increased government or charitable investment than with pharmaceutical investment in maternal health R&D. This may reflect: (i) the risks of pharmaceutical R&D in pregnant women, (ii) obstetric disease being perceived as less easily addressed through drug development than in other specialties—perhaps because of the hitherto refractory nature of maternal diseases— or (iii) the scepticism of experts reared in an environment devoid of pharmaceutical R&D.
The Women Deliver conference in 2007 identified three pillars for saving maternal and neonatal lives in low/middle-income settings: comprehensive reproductive health services, skilled care during and immediately after childbirth and emergency care when life-threatening complications develop (Partnership for Maternal Newborn and Child Health 2007). Half the maternal and neonatal mortality in the poorer regions of Africa and Asia could be averted by scaling up the three pillars coverage to 95% (Adam et al. 2005). But $6.1 bn would be required to 2015 to provide full maternal care to 73% of women in the world’s poorest 75 countries (Global Health Opportunities 2006 2006) and an additional $5.1 bn for universal coverage with life-saving interventions to reduce under-five mortality in 42 of the world’s poorest countries (Bryce et al. 2005). By comparison, the quantum of investment needed in pharmaceutical and non-pharmaceutical maternal health R&D is unclear (Fisk & Atun 2009); this reflects both the time lag for R&D returns and imprecision in estimating the funding needed for equity with other mainstream disease areas.
Although maternal conditions were considered more easily addressed than stillbirths or neonatal conditions, 95% of the avertable burden was in perinatal DALYs, reflecting their far greater number. Thus, better maternal R&D investment has its greatest impact on the baby, as similarly does optimal uptake of current knowledge.
Pharmaceutical and not-for-profit R&D is typically conducted in high-income countries focussed on disease burden in such countries. It was thus surprising that more than 98% of the disease burden that could be addressed by increasing R&D in either sector was in low/middle-income settings, especially as the endpoints chosen deliberately omitted categories dominated by health service provision or infections in the developing world. We attribute this to the fact that in obstetrics, in contrast to other specialties, the same conditions that cause most disease burden in the developed world also do so in the developing world. Indeed, the three chief obstetric pathologies of pre-eclampsia, growth restriction and preterm labour are developed world research targets of major relevance to the developing world (Costello et al. 2007).
Turning around maternal health R&D
Under-funding of pharmaceutical R&D for maternal health is a symptom of market failure. In spite of the substantial disease burden, investment in maternal health R&D is analogous to that for a single orphan condition in high-income settings or for a neglected tropical disease in low/middle-income settings. While specific global initiatives have been successfully developed to overcome investment failures in both these scenarios, there have been none for maternal health. As with these analogies, boosting R&D to address problems experienced in pregnancy will require incentives for pharmaceutical investment and measures to convince governments, charities, philanthropic agencies and international donor agencies of what could be achieved (Fisk & Atun 2008, 2009; PLoS Editors 2008).
In conclusion, our findings indicate that, along with overall development, female empowerment and improved access to effective care, there is an argument to consider what might be achieved by greater investment in R&D for maternal care. Better pharmaceutical and public sector R&D may prevent around 1/3 and 2/3, respectively, of the disease burden that is potentially addressable by optimal uptake of existing research. These data in the context of the current air of futility around meeting maternal and perinatal MDGs suggest that greater investment in R&D may be an important complementary strategy to strengthening health service provision to address the global maternal and perinatal disease burden.
We acknowledge Professor Alan Lopez for advice and Roy Howell from Informa UK for access to the Pharmaprojects database.