Bulletin of the World Health Organization

Health-care-associated infection in Africa: a systematic review

Sepideh Bagheri Nejad a, Benedetta Allegranzi a, Shamsuzzoha B Syed b, Benjamin Ellis c & Didier Pittet d

a. Clean Care is Safer Care, World Health Organization Patient Safety, Geneva, Switzerland.
b. African Partnerships for Patient Safety, World Health Organization Patient Safety, Geneva, Switzerland.
c. World Health Organization Patient Safety, Geneva, Switzerland.
d. Infection Control Programme and WHO Collaborating Centre on Patient Safety, University of Geneva Hospitals and Faculty of Medicine, 4 Rue Gabrielle Perret-Gentil, 1211 Geneva 14, Switzerland.

Correspondence to Didier Pittet (e-mail: didier.pittet@hcuge.ch).

(Submitted: 11 March 2011 – Revised version received: 05 July 2011 – Accepted: 08 July 2011 – Published online: 20 July 2011.)

Bulletin of the World Health Organization 2011;89:757-765. doi: 10.2471/BLT.11.088179

Introduction

Health-care-associated infection (HAI) is a major global safety concern for both patients and health-care professionals.13 HAI is defined as an infection occurring in a patient during the process of care in a hospital or other health-care facility that was not manifest or incubating at the time of admission. This includes infections acquired in the hospital and any other setting where patients receive health care and may appear even after discharge. HAI also includes occupational infections among facility staff.1 These infections, often caused by multiresistant pathogens, take a heavy toll on patients and their families by causing illness, prolonged hospital stay, potential disability, excess costs and sometimes death.46

The burden of HAI is already substantial in developed countries, where it affects from 5% to 15% of hospitalized patients in regular wards and as many as 50% or more of patients in intensive care units (ICUs).7,8 In developing countries, the magnitude of the problem remains underestimated or even unknown largely because HAI diagnosis is complex and surveillance activities to guide interventions require expertise and resources.6 Surveillance systems exist in some developed countries and provide regular reports on national trends of endemic HAI,9 such as the National Healthcare Safety Network of the United States of America or the German hospital infection surveillance system. This is not the case in most developing countries10 because of social and health-care system deficiencies that are aggravated by economic problems. Additionally, overcrowding and understaffing in hospitals result in inadequate infection control practices, and a lack of infection control policies, guidelines and trained professionals also adds to the extent of the problem.

This review provides a general overview of the endemic burden of HAI in Africa based on the information available in the scientific literature. It also identifies information gaps, examines differences in HAI epidemiology between developed and developing countries and highlights the possible role of the World Health Organization (WHO) in preventing HAI.

Methods

Search strategy and selection criteria

A literature search was performed from January 1995 to December 2009 with no language restriction to retrieve publications on the epidemiology of the most common HAIs in African countries: health-care-associated urinary tract infection (HA-UTI), surgical site infection (SSI), hospital-acquired pneumonia/ventilator-associated pneumonia and health-care-associated bloodstream infection. PubMed was searched using a combination of the following keywords, including “cross-infection” as the MeSH term: “nosocomial infection”, “hospital-acquired”, “incidence”, “prevalence” and “rate” together with the individual country names. The Cochrane Library was searched for any relevant review papers. Reference lists of retrieved articles were hand searched for additional studies.

A separate search was run in the WHO regional medical database for Africa, African Index Medicus, using a shorter list of essential keywords and with no time restriction. The abstract books of the following international conferences were also searched from 2004 to 2009: Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Annual Congress of the Society for Healthcare Epidemiology of America (SHEA), European Congress of Clinical Microbiology and Infectious Diseases (ECCMID), International Federation of Infection Control (IFIC), the International Congress on Infectious Diseases (ICID), and the first African Conference on Infection Prevention Control (IPCAN), held in 2009. For the purposes of this review, African countries are defined as those belonging to the WHO African Region, which comprises all African countries except for Egypt, the Libyan Arab Jamahiriya, Morocco, Somalia, Sudan and Tunisia.

All studies examining the epidemiology, microbiology or impact of HAI (on costs, hospital stay, attributable mortality, etc.) were selected and included in the review. Reports of outbreaks were excluded. Full texts of relevant English and French articles were obtained and scrutinized. Studies were classified according to the patient population (adults, neonates/children, mixed ages) and the type of infection, stratified into five categories: general HAI (studies covering at least the four most frequent types of HAI); health-care-associated urinary tract infection; health-care-associated bloodstream infection; surgical site infection; and hospital-acquired pneumonia/ventilator-assisted pneumonia. The most recent data were used for papers reporting HAI rates from different years.11 Prevalence, cumulative incidence and incidence densities were defined as previously described.12

The following predefined criteria were established to assess the quality of the studies: prospective design; use of standardized definitions (i.e. according to those of the NNIS/NHSN system of the US Centers for Disease Control and Prevention);13 detection of at least all four major infections for studies on HAI in general; and publication in a peer-reviewed journal.

Results

The PubMed search yielded 232 papers. Of these, 12 met the eligibility criteria and were included in the review.11,1424 Seven additional papers were identified through other searches, (Fig. 1)2531 resulting in a total of 19 papers from 10 countries. No review was retrieved through the Cochrane Library. No relevant abstracts were presented at the SHEA, ECCMID, and ICID conferences between 2004 and 2009, or at IPCAN 2009. Two abstracts from the United Republic of Tanzania were included in the 2004 ICAAC abstract book, but both referred to the same published study16 already retrieved and they were excluded. Five relevant abstracts were identified from IFIC conferences, but one did not report any HAI rate and two presented the results of the same study.

Fig. 1. Flow diagram for systematic review of the literature on the burden of health-care-associated infection (HAI) in Africa, 1995–2009
Fig. 1. Flow diagram for systematic review of the literature on the burden of health-care-associated infection (HAI) in Africa, 1995–2009

General study characteristics

The general characteristics of the papers are presented in Table 1. In general, hospital-wide studies were principally prevalence surveys, and those carried out in specific wards (surgical wards and ICUs) were mostly incidence studies (Table 2). HAI proportions were reported mainly as infection per 100 patients. The denominator for SSI was the number of operated patients, except for one study that used the number of operations. Two studies16,17 were classified as high quality.

Rates and isolated pathogens

Overall rate

Hospital-wide HAI prevalence varied between 2.5% and 14.8% in Algeria,19 Burkina Faso,15 Senegal14 and the United Republic of Tanzania.11 Overall HAI cumulative incidence in surgical wards ranged from 5.7% to 45.8% in studies conducted in Ethiopia30 and Nigeria.20 The latter reported an incidence as high as 45.8% and an incidence density equal to 26.8 infections per 1000 patient-days in paediatric surgical patients.20 In a study conducted in the surgical wards of two Ethiopian hospitals, the overall cumulative incidence of patients affected by HAI was 6.2% and 5.7%.30

Surgical site infection was the most common infection encountered in two studies investigating overall HAI incidence rates among surgical patients.20,30 Similarly, a study from Burkina Faso on HAI prevalence among surgical patients reported surgical site infection as being the most common type, followed by urinary tract infection and hospital-acquired pneumonia.23

Following an intervention to prevent HAI, an Algerian study reported a decrease in HAI prevalence from 9.0% in 2001 to 4.0% in 2005 (P < 0.001).11 In this study, surgical site infection was the most common type of HAI from 2002 to 2004, whereas urinary tract infections and hospital-acquired pneumonia were the most common HAIs in 2001 and 2005, respectively. In a study from Nigeria, the implementation of an infection control programme in a teaching hospital succeeded in reducing the rate of HAI from 5.8% in 2003 to 2.8% in 2006.32

Four studies on HAI in general included microbiological data. In one such study,11 the isolated pathogens, by order of decreasing frequency, were Pseudomonas aeruginosa, Escherichia coli, Klebsiella pneumoniae and Enterobacter spp. Another study15 reported Enterobacter cloacae as the most common pathogen followed by E. coli, Staphylococcus aureus and P. aeruginosa. A report on HAI cumulative incidence in surgical patients20 showed the following distribution: K. pneumoniae (38.7%); E. coli (22.7%); P. aeruginosa (16.8%) and S. aureus (10.7%). The fourth study, which included the surgical wards of two hospitals, showed different microbiology patterns in the two facilities.30

Ten studies (eight with a focus only on surgical site infection) and three conference abstracts reported an incidence of surgical site infection ranging from 2.5% to 30.9% following various types of surgical procedures.11,16,17,20,2529,31,3335 In six studies and one conference abstract,16,17,20,26,28,29,33 elective and emergency procedures were included. One study focused on elective caesarean section only,27 and no information on the type of surgery was available in three articles11,25,31 and two abstracts.34,35 In Nigeria, the cumulative incidence was 23.6 per 100 operations.29 When reported, the incidence of surgical site infection by wound classification ranged from 6.5% to 20.2% in clean wounds, 10.1% to 23.8% in clean-contaminated wounds, 13.3% to 51.9% in contaminated wounds and 44.1% to 83.3% in dirty wounds.17,20,25,28,29 Superficial, deep and organ/space surgical site infection accounted for 38.2% to 73%, 6.8% to 46.5%, and 10.4% to 20.5% of all surgical site infections, respectively.16,17,20,31 In an Algerian study, the cumulative incidence of surgical site infection decreased from 11.9% in 2001 to 2.5% in 2005 (P < 0.01) following an infection control intervention.11 In a survey conducted in the United Republic of Tanzania, surgical site infection was identified after discharge in 21% of patients, one third of whom were re-hospitalized because of such infection.16 In another Tanzanian study, 19.4% of patients developed surgical site infections after surgery, and in 36.4% of these patients the problem was identified during post-discharge follow-up.17 In a Ugandan study, the overall cumulative incidence of surgical site infection was 10% among surgical patients in general and 9.4% among women who underwent caesarean section.25 The authors report that in this last group the figure dropped dramatically with respect to former incidence estimates (some of them higher than 50%) after the introduction of a standardized protocol for surgical wound management.25 In a study conducted in Ethiopia, the cumulative incidence of surgical site infection was 21% based on clinical criteria and 38.7% based on bacteriological criteria in patients who had undergone abdominal surgery.26 In a study from Kenya, the cumulative incidence of surgical site infection after caesarean section was 19% overall and 33% among women who had been in labour for more than 12 hours (versus 15% among women whose labour had lasted fewer hours).27 In a study from the Central African Republic, three of 51 patients who developed surgical site infections were identified after discharge. Of note, only 25% of all patients who were asked to return for a follow-up visit on the 30th day after surgery actually attended.31

Five studies reported microbiology data on surgical site infection. In three,16,17,26 S. aureus and E. coli were the leading pathogens recovered from infected wounds. Other reported isolates included Klebsiella spp, Enterococcus spp, Pseudomonas spp. and other enterobacteriaceae. A Nigerian study29 reported E. coli as the most common pathogen (34.4%), followed by Klebsiella spp. (21.9%), Pseudomonas spp. (15.6%), Staphylococcus spp. (12.5%), Proteus spp. (9.4%), and E. coli and Proteus spp. (6.3%). In a study from the Central African Republic, S. aureus and Proteus mirabilis were the most common pathogens isolated from the infected surgical sites.31

The prevalence of urinary tract infection was 0.7% and 4.5% in two studies from Algeria and Senegal, respectively,11,15 while a retrospective study from Nigeria reported a frequency of 12.3%.18 The study from Algeria reported that the prevalence of UTI decreased from 3% to 0.7% in 2001 and 2005, respectively, following an infection control intervention.

The hospital-wide prevalence of hospital-acquired pneumonia was 1.7% and 2.9% in studies conducted in Algeria and in Senegal, respectively.11,15 In another study from Algeria, the cumulative incidence of hospital-acquired pneumonia in the neonatal ICU was 2.4%.21 No microbiology data were reported in these studies. In one Senegalese study conducted in an ICU, the proportion of ventilated patients affected by ventilator-associated pneumonia was 50%.24

Limited information was available on the impact of HAI in terms of cost, prolonged hospital stay and attributable mortality. In the study from the United Republic of Tanzania,17 the mean postoperative hospital stay was 5.4 days for uninfected patients compared with 13 days for those with surgical site infection. In the study from Burkina Faso, the hospital stay was 10 days longer on average in patients with HAI.23 A study from Ethiopia26 reported that a delay in hospital discharge was attributable to surgical site infection in 14.7% of patients. In another Ethiopian study,28 the mean postoperative stay was 19.6 days in patients with surgical site infection compared with 11.3 days in uninfected patients. In the same study, mortality was 4.9% overall, but 10.8% for patients with surgical site infection compared with 3.9% for patients without infection.

Discussion

The small number of papers retrieved is evidence that little information is available on the epidemiology of HAI in African countries. The review has shown that published studies were conducted in 10 African countries only. Of the 19 publications retrieved, several are from the same country. In addition, the scope of the studies is limited, since most were conducted in single hospitals or single wards. Of note, more studies may have been conducted, but not published for different reasons. Additional reports of outbreaks of HAI exist, but we focused our review on the endemic burden, which represents most HAI. Our review showed great variability in study design and in the reported prevalence and incidence of HAI. Standardized criteria and definitions for the diagnosis of HAI were not used in all studies and this may explain, at least partially, the variation in the rates of HAI. This was reflected in the overall quality of the studies (low for 17 of the 19 [88.9%] included reports) and makes any comparison with other studies difficult, particularly those from developed countries. In many cases, using standardized definitions implies the availability of reliable laboratory conditions usually lacking or poor in resource-limited settings. Patient charts and records may also be less accurate or even non-existent.

The overall prevalence of HAI ranged from 2.5% to 14.8%, up to twice as high as the average European prevalence (7.1%) reported by the European Centre for Disease Prevention and Control.36 These findings are consistent with HAI pooled prevalence and incidence data reported from a recently published systematic review37 on the burden of endemic HAI in developing countries (10.1%, 95% confidence interval, CI: 8.4–12.2, and 7.4%, 95% CI: 4.4–12.2, respectively). In this review, the pooled cumulative incidence and density of ICU-acquired HAI were 34.7% (95% CI: 23.6–47.7) and 47.9 per 1000 patient–days (95% CI: 36.7–59.1), respectively. This is much higher than the estimated density of 13.6 per 1000 patient–days in the United States. In a systematic review related to HAI in neonates, Zaidi and colleagues38 reported the HAI frequency to be three to 20 times higher in resource-limited countries compared with industrialized nations.

Some important aspects need to be considered when interpreting our findings. African settings able to conduct surveillance studies and publish data may have greater resources to implement infection prevention and control programmes than those who do not collect and publish data. Thus, the real burden of HAI is likely to be even greater in settings with weaker infrastructures and fewer resources. Most included studies (13/19) were conducted in university/teaching hospitals (Table 1) that usually function as referral hospitals and accept patients requiring more complex care. For these reasons, such hospitals generally report higher infection rates. No national studies were identified and only one multicentre study in two hospitals was retrieved, which makes the difficulties of conducting coordinated and regular HAI surveillance in Africa all too clear. For all these reasons and the lack of quality previously mentioned, this review does not provide a comprehensive picture of HAI in the African continent. Instead, it provides the best overview possible while highlighting the many existing gaps.

Most papers focusing on a particular type of infection studied surgical site infection. Apart from one study reporting a cumulative incidence of surgical site infection of 2.5%, the cumulative incidence of HAI following various types of surgical procedures ranged from 10.0% to 30.9%, a rate markedly higher than in high-income countries. As an example, the average cumulative incidence rate of surgical site infection was 2.6 per 100 surgical procedures in a nationwide study conducted in the United States39 and 3 per 100 surgical interventions in different European countries.40 Although limited data were available on the impact of HAI, surveys conducted in surgical wards clearly documented that patients affected by surgical site infection had an increased hospital stay.17,23,26,28

Despite some obstacles, there are encouraging signs that the importance of HAI has started to be recognized in Africa. An Algerian study11 documents how the introduction of a prevention programme at the facility level in 2001 reduced the overall hospital-wide prevalence of HAI over five consecutive years (2001–2005). In Uganda,25 the implementation of a standardized protocol for surgical wound management dramatically reduced surgical site infection after caesarean section.

Importantly, infection control improvement has been undertaken nationally in some countries, such as Senegal, where a national programme to reduce HAI (Programme national de lutte contre les infections nosocomiales [PRONALIN]), implemented in 2004, has become a catalyst for similar programmes in other countries in the region. Tangible results, achieved with little investment, have been the establishment of infection control committees in main hospitals, national training for more than 3500 health-care workers, two national prevalence surveys on HAI, and the development of new national policies for medical waste management and antibiotic use. A national hand hygiene programme with a focus on training and advocacy has been embedded within these activities. The Infection Prevention and Control Africa Network and the Réseau international pour la planification et l’amélioration de la qualité et de la sécurité dans les systèmes de santé en Afrique coordinate regional and international efforts.

In 2005, WHO launched the First Global Patient Safety Challenge “Clean Care is Safer Care” to create a global momentum and commitment to reduce HAI. The objectives are to raise awareness of the importance of HAI as a major patient safety issue, build country commitment to tackle the problem, and develop tools and guidance documents. Within the development process, the WHO guidelines on hand hygiene in health care and the multimodal hand hygiene improvement strategy underwent a pilot test phase to assess their feasibility and adaptability to local contexts and to the local resources available.41,42 Allegranzi et al.37 reported the successful implementation and adaptation of the strategy at the test site for Africa in Bamako, Mali. The intervention consisted of introducing a locally-produced alcohol-based hand rub, monitoring hand hygiene compliance, providing performance feedback, educating staff, posting reminders in the workplace and promoting an institutional safety climate. The results clearly demonstrate that multimodal hand hygiene promotion is feasible and effective in low-income settings.37 Other key measures for achieving basic infection control in health-care settings have been highlighted by WHO programmes: core components for infection prevention and control programmes,43 strategies for injection44 and blood transfusion safety,45 safe medical waste management,46 standards for sterilization and disinfection,47 water and sanitation,48 and occupational heath measures.49

A technical paper on patient safety detailing 12 key action areas, including the reduction of HAIs, was prepared by the WHO African Regional Office in collaboration with WHO Patient Safety and endorsed by all 46 African Member States.50 More recently, African Partnerships for Patient Safety was launched in response to this political commitment to improve patient safety, particularly by reducing HAI, across the region.51 The 12 key action areas form the basis of the new programme, which works through hospital partnerships between Africa and Europe with a focus on the exchange of knowledge and skills between front-line health-care professionals.52 Each partnership is working on HAI prevention through systematic situational analyses of patient safety and focused interventions. These activities are already being taken up by national systems.

Initiatives such as those described herein demonstrate that professionals and policy-makers consider HAI a very serious problem and that simple, low-cost interventions can be successfully implemented in Africa, despite the continent’s fragmented political and financial situation. These efforts need support and encouragement by WHO and other agencies and organizations. In light of the paucity of data highlighted by our review, efforts to reduce HAI should begin with surveillance activities aimed towards estimating the burden of morbidity and mortality associated with HAI.


Acknowledgements

We thank Michael Borg, Nizam Damani, Elizabeth Mathai, Shaheen Mehtar and Rosemary Sudan for their invaluable contributions to this study.

Funding:

This work was funded by WHO.

Competing interests:

None declared.

References

Share