Intermittent preventive treatment of malaria in pregnancy: a new delivery system and its effect on maternal health and pregnancy outcomes in Uganda
A K Mbonye a, I C Bygbjerg b, P Magnussen c
Malaria is the leading cause of morbidity and mortality in Uganda. It mainly affects pregnant women and children age less than five years.1,2 Surveillance reports show that proportionate mortality ratio (PMR) due to malaria for all ages has increased progressively from 20.2% in 1988 to 32.1% in 2004.3 This increase has been attributed to high transmission of malaria in areas that were previously free of the disease, limited access to adequate treatment in the formal health-care facilities, increasing resistance to antimalaria drugs, and inadequate treatment of malaria at home where most people receive the first treatment.4 The resurgence of malaria necessitates intensification of treatment and preventive interventions and identification of new delivery approaches to increase access to effective interventions.
Previous studies have reported the involvement of community resource people, such as traditional birth attendants, drug-shop vendors, community reproductive-health workers, and adolescent peer mobilizers in malaria treatment and prevention. A study in Saradidi, Kenya, found volunteer village health workers effective in giving malaria treatment.5 In the Gambia, chemoprophylaxis given to pregnant women by traditional birth attendants reduced malaria-related morbidity and poor pregnancy outcomes.6,7 More recently, in south-western Uganda, a study showed that community health workers can classify pneumonia, interpret the severity of the disease, and provide treatment for fevers with antimalarials.8
In this study, we tested a new delivery system of intermittent preventive treatment (IPTp) in pregnant women at a community level in Mukono district, Uganda. The study assessed delivery of IPTp with sulfadoxine-pyrimethamine (SP) to pregnant women through community resource people; these approaches constituted the intervention arm while provision of IPTp at health units was the control arm. The primary outcome measures were access and adherence to IPTp, number of malaria episodes, prevalence of anaemia, and birth weight; while live births, abortions, still births, and maternal and child deaths were secondary endpoints.
The new delivery system of IPTp in Uganda was conceptualized because of the persistent low use of health-service-based interventions. Only 47% of women attend four antenatal care visits as recommended by the policy, and only 42% of births have skilled attendants.9,10 Low use of antenatal care limits access to malaria prevention in pregnancy and has been attributed to the high cost of services, long distances to health units, negative perceptions of SP and insecticide-treated nets, and low male involvement.11–13 Malaria-control programmes that rely only on antenatal care as a delivery system are likely to have poor coverage and adherence to IPTp.14–16
Materials and methods
Study area and population
The study was done in nine subcounties near the lakeshore region of Mukono district in central Uganda. The background characteristics of the study area have been published elsewhere;13,17 but briefly, the district has a rural population of 88% and is mainly inhabited by Baganda people, the largest tribe in Uganda. The area experiences high temperatures and heavy rainfall from March to May and from October to November. Most areas are endemic for malaria, whereas those on the shores of Lake Victoria are holoendemic and experience high transmission levels.18
The intervention aimed to train community resource people to distribute IPTp. In Uganda, IPTp consists of two therapeutic doses of SP (three tablets of 500 mg sulfadoxine + 25 mg pyrimethamine).3 Nine rural subcounties from the district that were homogenous in terms of geography and cultural attributes were selected. Within each subcounty, at least two parishes were randomly selected.
In total, 25 parishes were selected, four tested IPTp at health units (control) while 21 tested IPTp with the new approaches (intervention; Fig. 1). The intervention was not randomized because not all the parishes were uniform in regard to the availability of health units and people available to deliver the treatment. Some parishes had health units providing IPTp; others had both a health unit and people to deliver the treatment while others had only people who referred pregnant women to neighbouring parishes with health units for essential care. The inequitable distribution of health services in the study area and the current health-seeking behaviour influenced the design of the study. Despite this, the populations in the two study arms were comparable in terms of socioeconomic status and other background characteristics (Table 1).
We targeted all the pregnant women who lived in the study area. One of the key components of the intervention study was a community mobilization and sensitization campaign to ensure that all women received information on the intervention and where to get the SP. Because there was no register of people in this area, the number of pregnant women who accepted the first dose of SP served as the denominator for calculating the proportion of women who received the second dose of SP and other outcome variables.
Fifty-one people able to deliver treatment were identified. All these people were trained for one week. The following areas were covered: dangers of malaria in pregnancy; malaria prevention in pregnancy; the benefits of SP and its side-effects; taking blood samples for parasite count and haemoglobin analysis; taking the baby’s weight; and estimating gestational age. Recruitment of pregnant women lasted from May 2003 to February 2005; and those who consented to participate in the study were given SP as directly observed therapy during the second and third trimester as recommended by the policy in Uganda. Pregnant women received iron and folic acid supplementation, deworming, and information on proper nutrition. Those who presented in the first trimester were given an appointment to come back. Exclusion criteria were refusal to give consent to the study and a history of allergy to drugs containing sulphonamide.
Data were collected by the community resource people from pregnant women at recruitment, before receiving the second dose of SP and at 36 weeks gestation or at delivery on sociodemographic characteristics, obstetric history, reported malaria illnesses, use of insecticide-treated nets, doses of SP and dates when taken, pregnancy outcomes and deaths. These data were also collected at health units.
To have comparable groups, pregnant women who exclusively accessed IPTp (both the first and second dose of SP) with the health units constituted the control group, while pregnant women who exclusively accessed IPTp with the new approaches were the intervention group.
Table 1. Background characteristics of pregnant women who accessed IPTp in heath units and the new delivery system
Data were entered into EpiInfo version 6.0 (CDC, Atlanta, GA, United States of America), cleaned and transferred to Stata version 8.2 (Stata corp. College Station, TX, USA). Bivariate analyses were done to calculate access and adherence rates to IPTp. The comparison of the effect of the intervention between the new delivery system and health units on access, use of IPTp, maternal health and pregnancy outcomes were calculated with a two-sample proportion test. All estimates with P < 0.05 were considered significant. To assess the effect of SP delivery approaches on maternal-health outcomes, we constructed a logistic regression model while controlling for the effects of age, education, parity and occupation of respondents. The dependent variable was access to the first dose of SP at either the new approaches or health units, while the independent variables were reported number of malaria episodes, parasitaemia, haemoglobin levels at 36 weeks and birth weight. The model fit was assessed for robustness with the log-likelihood ratio test and model estimates were computed at P < 0.001.
The study aimed to detect differences in mean haemoglobin of 0.5 g/dl, mean birth weight of 150 g, and a difference of 12% in the proportion of women accessing IPTp between the health units and the new approaches. The minimum sample size required were 310 at each SP distribution outlet; with a study power of 80% and 5% level of significance. To achieve this, 25 parish clusters were selected. On average a parish has a population of approximately 2000 people; therefore the total population was 54 000. With a crude birth rate of 46.9/1000 per year, a total of 3596 pregnant women were expected over a 17-month study period. The number of parish clusters (21 for the new approaches and four for health units) were determined by the sample size required for measuring the primary outcome variables and the rate at which pregnant women would be recruited. For all the analyses, pregnant women accessing IPTp with the new approaches were grouped after calculations showed no significant differences among the different community approaches on all the outcome variables.
Self-reported malaria was defined as a febrile illness characterized by feeling hot with headache, joint pains and general weakness. The reference period for self-reported malaria episodes was two weeks before the assessment. Anaemia was defined as haemoglobin concentrations less than 11.0 g/dl.
The study was approved by the Danish National Committee for Biomedical Research Ethics and the Uganda National Council of Science and Technology. Verbal consent was obtained from all pregnant women who participated in the study.
A total of 2785 women participated in the study. The mean age was 23.9 years, range 14–46 years. There was no difference in mean age of women accessing IPTp with health units (control) and those receiving treatment through the new delivery system (intervention), 23.6 years versus 24.0 years (P > 0.12). Most women in both sets had attained primary education, while more women using IPTp with health units had attained secondary level education (23.2% versus 18.4%) a factor that influences health seeking behaviour (Table 1).
Effect of a new delivery system on access and compliance to IPTp
Of the women using the new delivery system, 1905 (92.4%) of 2062 received first dose of SP in the second trimester compared with 523 (76.1%) of 687 receiving treatment in health units, P < 0.0001. The mean gestational age for all women was 21.4 weeks (range 16–36 weeks). Women visiting health units had a mean gestational age of 23.1 weeks versus 20.8 weeks with the new delivery system P < 0.0001. Adherence to two doses of SP was 39.9% (281 of 704) with health units and 67.5% (1404 of 2081) with the new delivery system, P < 0.0001 (Table 2). Few women, 286 (17.0%) of 1685 accessed SP from both health units and the community approaches.
Effect of a new delivery system on maternal health and pregnancy outcomes
Most, 116 (62.7%) of 185 women who accessed IPTp in health units delivered their babies with the help of skilled attendance, while fewer, 460 (40.6%) of 1133, who accessed IPTp with the new delivery system gave birth with the help of skilled attendants. In both approaches, most deliveries (over 98%) were live births while there were few abortions and stillbirths; most births (over 95%) were singleton babies (Table 3).
Blood samples for haemoglobin analyses were collected from 761 (27.3%) of 2785 pregnant women at recruitment; 389 (23.1%) of 1684 women after the first dose of SP; and 228 (13.5%) of 1689 at 36 weeks or at delivery. At 36 weeks gestation or at delivery, mean haemoglobin was 10.8 g/dl (range 9.9–11.6 g/dl) among women visiting health units and 11.0 g/dl (10.7–11.5 g/dl) with the new delivery system (P > 0.6). The proportion of anaemia (haemoglobin < 11.0 g/dl) was lower at health units, 40.6%, than in patients treated with the new approaches, 56.6% (P < 0.02). Both delivery approaches were associated with a high mean birth weight, 3201 g (range 3030–3236 g) with health units and 3223 g (range 3186–3264 g) with the new delivery system. Women using the new delivery system registered a lower proportion of low birth weight, 57 (6.0%) of 950, than did those in health units, 12 (8.3%) of 144, (P < 0.03; Table 3).
The prevalence of reported malaria episodes decreased from 261 (39.1%) of 667 at recruitment to 13 (13.1%) of 99 at delivery among women who accessed IPTp at health units. Reported malaria episodes decreased from 906 (49.5%) of 1830 at recruitment to 160 (17.6%) of 909 at delivery among women who accessed IPTp with the new delivery system. Both at recruitment and delivery, the prevalence of reported malaria episodes was lower in women accessing treatment at health units than in those accessing treatment in the new delivery system.
At recruitment, 573 (24.5%) were positive for Plasmodium falciparum parasites: 124 (22.2%) of 559 with health units and 449 (25.2%) of 1782 with the new delivery system (P > 0.15). After the first dose of SP, 180 (15.9%) of 1132 were positive for P. falciparum parasites: 12 (10.6%) of 113 were recoded at health units and 168 (16.5%) of 1018 with the new delivery system. At 36 weeks or at delivery, 105 (16.1%) of 654 had parasitaemia: three (9.1%) of 33 in health units and 102 (16.4%) of 622 with the new delivery system (P < 0.04; Table 3).
A total of 1061 (115 with the health units and 946 with the new delivery system) outcomes of pregnancies were documented at the end of the study (Table 3). With the health units, 113 (98.3) were live births, 1 (0.9%) was an abortion and 1 (0.9%) was a stillbirth. With the new delivery system, 928 (98.1%) were live births, 5 (0.5%) were abortions while 13 (1.4%) were stillbirths (Table 3).
There were few registered deaths in this study. By the end of the intervention, 5 (2.7%) of 185 children born to women who accessed IPTp with the health units died while 21 (1.9%) of 1105 children born to women who used the new delivery system died. No woman died among those who accessed IPTp with the health units and 4 (0.4%) died among those who accessed IPTp with the new delivery system (Table 3).
Multivariate analyses were used to assess the effect of SP delivery approaches on outcomes in maternal health and pregnancy. The effect of age, education, parity, occupation of respondents and use of insecticide-treated nets were controlled. The differences observed in Table 3 were still significant for parasitaemia at 36 weeks (odds ratio: 2.3, 95% confidence interval, CI: 1.2–5.5, P < 0.03), self-reported malaria episodes at 36 weeks (OR: 1.6, 95% CI: 1.3–6.1, P < 0.04); and birth weight (OR: 1.8, 95% CI: 1.6–5.4, P < 0.03; data not shown).
The results of the present study show that women using the new delivery system accessed IPTp early and most them adhered to the two doses of SP. In health units and the new approaches, women experienced a reduction in malaria episodes, anaemia, parasitaemia and low birth weight. Although these results cannot be attributed to the intervention alone, after controlling for age, education, parity and occupation, significant differences existed for parasitaemia, reported malaria episodes and birth weight; indicating the importance of access and adherence to IPTp. There was easy access to community resource people, and their ability to make frequent home visits, especially community reproductive health workers and adolescent peer motivators, was an important factor in reaching pregnant women early with IPTp, which explains why early access to the first dose of IPTp (gestation age) was lower for women in the new approaches compared with those in health units (Table 2).
Adolescents and women in their first pregnancies were more likely to visit health units than access care with the new approaches (Table 2). This trend has also been observed with larger surveys in Uganda9,10 and is probably because such people have no experience with pregnancies and they visit health units for explanations and care in case of complications. Generally more women visiting health units were using insecticide-treated nets and were likely to get effective case management of malaria, information on benefits of insecticide-treated nets and other malaria prevention interventions. This probably explains why women at the health units had lower episodes of malaria at recruitment and subsequent measurements. However, early access and adherence to IPTp are important factors in mitigating the negative effects of malaria in pregnancy. This is probably why the new delivery system was associated with a lower proportion of low birth weight and slightly higher (though not significantly so) mean haemoglobin concentration and mean birth weight than was access through health units.
There were few abortions and still births, which is consistent with previous studies on malaria in pregnancy that have shown that in high transmission areas malaria leads to anaemia and low birth weight.6,19–21 Although measuring mortality was not the primary objective of the study, data on deaths among women who accessed IPTp and their children were collected. Fewer children born to mothers who accessed IPTp at health units died and no woman was reported dead at this delivery point. Relatively, more children and women died among the group that accessed IPTp with the new delivery system. This finding could be related to the different health seeking behaviours for effective case management of malaria and other life threatening conditions at health units among the different groups of women. For example, a higher proportion of women who accessed IPTp at health units also gave birth in them with skilled care (Table 3), thus reducing the risk of dying from pregnancy-related complications, especially during delivery or immediately thereafter. We recommend a study with adequate sample size to evaluate further the effect of the new delivery system on maternal and neonatal mortality.
The case definition for malaria used in this study could have led to an overestimate of malaria episodes. This is because the local term omusujja gw’ensiri is commonly used interchangeably with another local term omussujja, which is non-specific and refers to any febrile illness.13,22
A limitation with this study is that placenta malaria was not measured. This is a sensitive measure for malaria in pregnancy and could have given more insight into the effect of the new delivery system. Use of community resources in studies has its own limitations, such as how much laboratory and clinical work can be done. Community resource people are not professionals and may not provide accurate data.
Data on maternal health and pregnancy outcomes in a malaria endemic area are of public-health importance for two reasons. First, these data can be used for routine monitoring of malaria treatment and prevention interventions, and this could help policy-makers and programme managers make decisions regarding treatment and prevention policies. Second, local analyses and interpretation of data on pregnancy outcomes could be used by health workers to improve quality of care and as a source of motivation. We recommend a further study in which midwives are taught to analyse data and calculate mean haemoglobin levels, anaemia and low birth weight, and assess how this can influence them to encourage pregnant women to access and to adhere to malaria prevention interventions.
Scaling up provision of IPTp and maintaining this approach has policy implications. The resource people will have to be trained, facilitated and linked to the health units to get SP, basic supplies and effective supervision. The results of this study need to be disseminated to policy-makers and programme managers at the Ministry of Health and the district and sub-county managers to gain consensus on these results. ■
- The national health policy Kampala: MoH (Ministry of Health); 1999.
- A review of the situation of malaria prevention and control in Uganda. Kampala: MoH; 2005a.
- Management of uncomplicated malaria: a practical guide for health workers. 3rd edn. Kampala: Ministry of Health, 2005b.
- Implementing guidelines for the national home based management of fever strategy. Kampala: MoH; 2002.
- Spencer HC, Kaseje DC, Sempebwa EK, Huong AY, Roberts JM. Malaria chemoprophylaxis to pregnant women provided by volunteer village health workers in Saradidi, Kenya. II. Effect on parasitaemia and haemoglobin levels. Ann Trop Med Parasitol 1987; 81: 83-9.
- Menendez C, Todd J, Alonso PL, Lulat S, Francis N, Greenwood BM. Malaria chemoprophylaxis, infection of the placenta and birth weight in Gambian primigravidae. J Trop Med Hyg 1994; 97: 244-8.
- Greenwood BM, Greenwood AM, Snow RW, Byass P, Benett S, Hatib Njie AB. The effect of malaria chemoprophylaxis given by traditional birth attendants on the course and outcome of pregnancy. Trans R Trop Med Hyg 1989; (83): 589-94.
- Kallander K, Tomson G, Nsabagasani X, Nsungwa-Sabiti J, Senyonjo Y, Pariyo G, et al., et al. Can community health workers and caretakers recognise pneumonia in children? Experiences from Western Uganda. Trans R Trop Med Hyg 2006.
- The Uganda demographic and health Survey, 2001. Entebbe: Bureau of Statistics.
- The Uganda demographic and health survey-preliminary report, 2006. Kampala: Bureau of Statistics.
- Amooti-Kaguna B, Nuwaha F. Factors influencing choice of delivery sites in Rakai district of Uganda. Soc Sci Med 2000; 50: 203-13.
- Ndyomugyenyi R, Neema S, Magnussen P. The use of formal and informal services for antenatal care and malaria treatment in rural Uganda. Health Policy Plan 1998; 13: 94-102.
- Mbonye AK, Bygbjerg IC, Magnussen P. Intermittent preventive treatment of malaria in pregnancy: a community based delivery system and its effect on parasitaemia, anaemia and low birth weight. Int J Infect Dis 2007.
- Menendez C. Priority areas for current research on malaria during pregnancy. Ann Trop Med Parasitol 1999; 93: S71-4.
- Helitzer-Allen DL, McFarland DA, Wirima JJ, Macheso AP. Malaria chemoprophylaxis compliance in pregnant women: A cost-effectiveness analysis of alternative interventions. Soc Sci Med 1993; 36: 403-7.
- Schultz L, Steketee R, Chitsulo L, Macheso A, Nyasulu Y, Ettling M. Malaria and childbearing women in Malawi: knowledge, attitudes and practices. Trop Med Parasitol 1994; 45: 65-9.
- The Statistical Abstract. Entebbe: Bureau of Statistics; 2002.
- Uganda malaria control strategic plan, 2005/06-2009/10. Kampala: Ministry of Health.
- Brabin BJ. The risks and severity of malaria in pregnant women. Applied field research in Malaria reports No.1. Geneva: WHO; 1991.
- McGregor IA, Wilson ME, Bilewicz WZ. Malaria infection of the placenta in The Gambia, West Africa, its incidence and relationship to birth weight and placenta weight. Trans R Trop Med Hyg 1983; (93): 529-534.
- Garner P, Gulmezoglu AM. Prevention versus treatment for malaria in pregnant women (Cochrane review). In: The Cochrane Library, Issue 1, Oxford: Update Software, 2001.
- Kengeya-Kayondo JF, Seeley JA, Kajura-Bajenja E, Kabunga E, Mubiru E, Sembajja F, et al. Recognition, treatment seeking behaviour and perception of cause of malaria among rural women in Uganda. Acta Trop 1994;58:267-73. doi:10.1016/0001-706X(94)90020-510.1016/0001-706X(94)90020-5
- Department of Community Health, Ministry of Health, Box 7272, Kampala, Uganda.
- Institute of immunology, Microbiology and International Health, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark.
- DBL-Centres for Health Research and Development, Institute of Pathobiology, Faculty of Life Sciences, University of Copenhagen, Copenhagen, Denmark.