A review of health system infection control measures in developing countries: what can be learned to reduce maternal mortality

The importance of a strong health system as the essential route to achieving improvements in maternal health and reductions in maternal mortality is widely accepted [1]. Effective coverage of maternity services requires timely and affordable access, by all sectors of the population, to appropriate care of sufficient quality and safety to help assure positive health outcomes. Good access, safety and quality are the overriding aims of all health systems and such factors are crucial when considering the problem of infections resulting from childbirth. Improving and maintaining infection control as part of delivery care requires an efficiently functioning health system.

Labour and delivery are especially hazardous times of pregnancy. Apart from the risks of severe bleeding and obstructed labour, life threatening infections can be introduced into the mother and baby's organs and bloodstream. 'Maternal sepsis' is a general term which has been used to include various obstetric and genito-urinary tract infections introduced into the mother [2]. The World Health Organization ranks maternal sepsis as the sixth leading cause of disease burden for women aged 15-44 years, after depression, HIV/AIDs, tuberculosis, abortion and schizophrenia. As many as 5.2 million new cases of maternal sepsis are thought to occur annually and an estimated 62,000 maternal deaths will result from the condition [2]. Added to the burden of loss of women's lives caused by sepsis are the long term consequences of infertility and the association of maternal sepsis with over one million infection related neonatal deaths every year [3, 4].

A specific form of maternal sepsis is known as puerperal sepsis, an infection which is introduced during childbirth, but manifests in the post partum period within the first 42 days after delivery. It is of special importance because it is a serious, life threatening disease of the mother with infection of the womb and abdominal cavity, bloodstream infection, fever and pain [5]. In industrialized countries, puerperal sepsis is rare, causing 2.1% of maternal deaths. In Latin America and the Caribbean, its contribution to maternal mortality is 7.7%, ranking lower than hypertensive disorders, haemorrhage, obstructed labour and abortion. In Africa and Asia, it is the second commonest cause of maternal mortality after haemorrhage, causing 9.7% and 11.6% of deaths respectively [6]. Other infections resulting from childbirth cause a considerable burden of morbidity and include infections of the genital tract, Caesarean section wound infections and urinary tract infections, but are usually not life threatening.

In developing countries, many women still deliver at home, making prevention of infection at home and in the community important, especially if family members and traditional birth attendants are unaware of the need for infection prevention. The provision of delivery care by health professionals and in health facilities is expected, and indeed, likely to decrease infection rates because of use of clean practices, sterile gloves and instruments. Yet the tumultuous history of puerperal sepsis and its association with institutional delivery care and the birth attendant is well recorded. Infective organisms causing puerperal sepsis are often introduced when the birth attendant conducts invasive procedures such as vaginal examination, instrumental or caesarean delivery. When childbirth in hospitals became more common in Western countries in the early 20^th century, an increase in maternal mortality occurred, much of which was due to spread of infection between women in labour by the attending health professional and use of invasive obstetric procedures [7]. This occurred despite knowledge of how infections were spread which dated back to the mid 19^th century. Increasing concerns of hospital and healthcare associated infections are also currently recorded across many medical disciplines, even in high income, industrialised countries [8]. Given these experiences, the increasing use of health facilities for childbirth in developing countries [9] calls for an attitude of watchfulness. In India, for example, the national policy promotes institutional deliveries which have steadily increased in the last 15 years from 26% to 41% [10, 11]. Studies here have shown that sepsis could be responsible for as much as 40% of maternal deaths [12, 13]. In Mexico, 84% of deliveries occur in health facilities and rising Caesarean section rates were over 27% in the public sector and 70% in the private sector in 2005 [14]. Here, septic shock has been documented to account for as much as 5 to 10% of mortality [15]. There is no direct evidence of infection rates rising as a result of increasing institutional delivery rates. However, it is plausible that increasing utilisation of under resourced health facilities can result in stresses to the health system, overcrowding, poor environmental conditions, overworked health workers, shortages of drugs and supplies and sub standard clinical practices. These falling standards of care may include deteriorating infection control practices, resulting in an increased risk of institutionally acquired puerperal sepsis.

The epidemiology and aetiology of puerperal sepsis and other infections resulting from childbirth in developing countries are reviewed elsewhere [6, 16‐18]. Specific interventions necessary to prevent and treat infections are well known and include good hand hygiene, antisepsis, surgical sepsis and antibiotics. However, evidence on the more complex interventions relating to improvement of compliance, practice and behaviour is less well documented. Drawing from the broader infection control literature, this article reviews health system infection control measures pertinent to labour and delivery units in developing country health facilities.

A structured literature review was conducted between March and May 2009. The objective of the review was to inform the development of strategies to prevent infections transmitted during labour and delivery in developing country health facilities. We searched for literature which described infection control measures. Evidence of effectiveness was of interest, but we did not restrict our review to these studies as we wished to ascertain whether ideas were being tried out that required further testing. We anticipated that infection control measures were likely to be implemented within wider health system activities, so we did not confine our search only to maternity care.

The review was structured in so far as a list of terms was defined and used to search electronic data bases in a methodological manner, but was not intended to be a systematic review with pre-defined data extraction forms, plans for data synthesis, quality assessment or specific selection criteria [19]. The electronic bibliographic databases MEDLINE, EMBASE, CINAHL, POPLINE and the Cochrane library were searched to 2009 with no earlier date or language restriction. The databases were searched using the following terms alone and in combination: infection control AND [mater* OR neonat* OR health care OR health system OR quality care] AND [sepsis OR infection] AND [control OR prevent*]. No language restrictions were applied. Over 2000 articles were initially found. The titles and abstract of the articles were screened. First, articles on infection control interventions in community settings, antenatal care, abortion care, interventions directed specifically at the child or neonate and on specific infections (e.g. malaria, tuberculosis, HIV) were excluded. In a second step, we attempted to find articles from developing countries and related to infection during childbirth and safe or clean delivery. Almost no relevant articles were identified, so we included articles of general infection control measures and from developed countries. This yielded 116 articles.

The abstracts of these 116 resulting articles were scrutinised by two of the authors independently and in some cases, full texts were retrieved with 54 articles eventually found to be relevant to our objective. The selection of these 54 articles was not based on specific criteria, but on subjective decision making based on whether the article provided useful information on potential means to prevent infection transmission in labour and delivery units and whether interventions described were likely to overcome problems and challenges relevant to health systems in low resource settings.

Websites were searched based on the authors pre-existing knowledge of international agencies working in relevant areas, including United Nations agencies (UNICEF, UNFPA, World Health Organization); maternal health groups (American College of Nurse Midwives, Engender Health, International Confederation of Midwives, International Federation of Gynaecologists and Obstetricians, JHPIEGO, John Snow Inc, Partnership for Maternal, Newborn and Child Health); and quality, patient safety and surveillance organisations (Agency for Healthcare Research and Quality, Centres for Communicable Diseases, National Institute for Health and Clinical Excellence, Patient Safety Alliance, Quality Assurance Project). Web based citations from published papers were also searched.

The articles included had a global or overall developing country perspective. Reviews and guidelines were found (Table 1) and primary studies seeking to evaluate the effects of various infection interventions (Table 2). The developing countries included in the studies were in Asia (Nepal, India, Pakistan, Thailand), Africa (Egypt, Malawi, Mozambique, South Africa) and South America (Argentina and Columbia).

Characteristic problems related to infection control in developing countries include bad antibiotic prescribing practices, poorly functioning laboratory services, lack of surveillance data and sub-optimal design or construction of buildings and water and sanitation systems. Overcrowding of facilities and insufficient numbers of health workers are commonly noted. Increased bed numbers, nurse to patient ratios and bed space are known to have negative effects on infection transmission. Managers roles are not well specified, which contributes to the poor quality of services [20‐23]. The combination of limited resources and general health conditions such as malnutrition, anaemia, and underlying infectious disease pose added risks [24]. Given such challenges, establishment of good infection control practices are believed to require a broad spectrum of interventions which address the availability and use of appropriate technologies, clear procedural guidelines and functionality of the health system [25‐27].

Technological advances for preventing and treating puerperal sepsis in health facilities have been reviewed elsewhere and include supplies and equipment such as hand rubs and low cost disposable equipment, improved antibiotics and other drugs for treatment of severe infections, and microbiological diagnostic techniques [25, 27]. Alcohol based antiseptic products are more efficacious than soap and water in reducing bacterial counts, and are convenient to use especially in basic health facilities where supplies of running water may be limited [21, 28, 29]. Systematic reviews have not however, found an established link between use of hand hygiene products and reductions in nosocomial infections [30]. The application of antiseptic washes to the vaginal area during labour has received much current interest, but there is insufficient evidence of its effectiveness in preventing maternal infection [27, 31, 32].

The World Health Organization's Global Patient Safety Challenge was set up to highlight the need for multimodal approaches to prevent health care associated infections alongside technological innovations [26, 33]. Increasingly, multifaceted and multicomponent interventions, which draw from psychological, educational, organisational, administrative, technological and medical perspectives, are being evaluated [34]. Such interventions, which are implemented either alone or in combination, include guideline use, education and training, organisational change, surveillance and quality improvement.

The infection control measures described have been somewhat artificially categorised as it can be observed that each intervention, for example 'education' or 'surveillance' in fact comprises several other components such as performance feedback, use of guidelines or technological improvements. The least complex intervention found - introduction of new clinical guidelines - was not found to result in practice change [36]. To achieve optimal reductions in infection rates, there is some evidence that multimodal and multifaceted interventions are effective [40, 43]. Most current studies suffer from the limitations of quasi-experimental designs, the lack of controls and the multi-component nature of the interventions. Evidence on cost-effectiveness in infection control is lacking. There are few evaluations on infection control measures in labour and delivery, yet the principles of infection control remain the same across clinical areas and in different resource settings, so the findings of the studies are relevant to maternity care. The link between infection control, the quality and safety of services and health system factors is widely recognised [21, 29].

The findings of this review suggest a need to implement and evaluate complex and multifaceted approaches in obstetric units. The choice of the specific combination of components to be evaluated can be informed by what is known from the wider infection control literature, from existing information on ways to improve quality in maternity care and by tailoring strategies to address underlying problems of infection control [47]. Unnecessary or wasteful components need to be weeded out [21, 50, 51]. Measurement challenges remain and ways to standardise metrics across different settings, case mixes or facilities will allow better comparisons to be made between studies. Improvements of proxy indicators of morbidity, such as attributable length of stay have been called for [34] and may be especially relevant because puerperal sepsis is a comparatively rare event. For policy level decision making, the viability of allocating resources to infection control programmes would depend on demonstrating the merits between the costs of prevention and the costs of treatment [40]. The need for action in the implementation of infection control measures should not be seen as a competing priority with research support, especially where resources are limited. Studies of effectiveness should be designed to act as agents of change by also catalysing improvements in practice.

The interventions described in this review highlight the importance of behavioural change in the health workforce, yet such change is notoriously difficult to implement. Inadequate understanding of complex motivational factors may play a role [18]. Even if individuals are well motivated, working in a chaotic environment or a setting with poor infrastructure can be a barrier to change. Our knowledge of how specific characteristics or contexts affect behaviour is inadequate - for example, it is unclear why physicians and nursing assistants have been found to be poorer hand washers than nurses, or why compliance with infection control is more common during weekdays and in intensive care units [29]. Such uncertainties emphasise the need to study infection control interventions using a broad lens, combining knowledge from diverse perspectives such as psychology, education, organisational and management science and technology. Behavioural theories from models developed in psychology have been used to examine infection control practices such as hand washing in health care providers, concluding that it is the interdependence of various factors including environment, organisation and structure that matters, rather than individual behaviour [29]. Such understanding can help shape the design of appropriate interventions, so that instead of simplistic interventions targeting health workers to wash their hands, viable strategies are those that make changes which affect interactions between individuals, and how they function within their environment and their institutions. Recent expansion of patient safety initiatives in the developed world can be seen to draw from organisational, sociological and psychological theory [52, 53], and will provide future lessons for infection prevention in developing countries and for maternity care.

As one of the leading causes of maternal mortality in low and middle income countries, it is surprising how little attention has been paid to puerperal sepsis and infection control during childbirth. Globalization of health policy and consequent responses can have varying impacts. On the one hand, some targeted approaches which focus on single causes of maternal mortality and morbidity in developing countries appear to have created momentum and interest for specific conditions. For instance, considerable attention has recently been given to studying the use of misoprostol and other technologies to improve the management of post partum haemorrhage in several low and middle income nations [27]. Scaled up, multicentre research has provided unequivocal evidence of the efficacy of magnesium sulphate in preventing deaths from eclampsia [54]. Advocacy for the condition is now part of the Clinton Global Initiative Commitment and an International Call to Action has been developed [55]. A global campaign to end obstetric fistula is supported by an array of international organisations in which obstetric care to prevent obstructed labour is a core activity [56]. On the other hand, a high profile, World Health Organization supported Global Patient Safety Challenge focusing on infection control, [26, 33] has not resulted in actions to reduce infection risks in labour wards in developing countries.

An estimated 358,000 women die every year from the complications of childbirth and up to 15% of these are due to puerperal sepsis [57]. A simplistic extrapolation of the finding that infection rates can be reduced by 32% using optimal infection control measures [40], would suggest that the deaths of over 17,000 women could be prevented every year. Millions of women suffering from maternal sepsis and its long term consequences would also benefit. There should be no excuses for delaying targeted, global action to implement and evaluate infection control measures during labour and delivery for the prevention and reduction of puerperal sepsis and other related conditions.

This review has highlighted three overarching lessons related to infection control and maternal mortality reduction. Firstly, despite limited evidence on effective infection control measures during labour and delivery and from low resource settings, it appears that education, surveillance, organisational change and quality improvement interventions should be introduced, confirming the need for a health systems approach to reduce maternal mortality, especially in relation to sepsis. Second is the need to improve our understanding of organisational and behavioural change to effectively implement infection control measures. In doing so, we will need to be informed by diverse and multidisciplinary perspectives. Finally, globalized, targeted health policies or initiatives have the potential to bring attention to, and catalyse action for what is currently a neglected, but important cause of maternal death worldwide.