Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Integrated community case management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms in children under five years in Eastern Uganda

Integrated community case management of malaria and pneumonia increases prompt and appropriate... Background: Efforts to improve access to treatment for common illnesses in children less than five years initially targeted malaria alone under the home management of malaria strategy. However under this strategy, children with other illnesses were often wrongly treated with anti-malarials. Integrated community case management of common childhood illnesses is now recommended but its effect on promptness of appropriate pneumonia treatment is unclear. Objectives: To determine the effect of integrated malaria and pneumonia management on receiving prompt and appropriate antibiotics for pneumonia symptoms and treatment outcomes as well as determine associated factors. Methods: A follow-up study was nested within a cluster-randomized trial that compared under-five mortality in areas where community health workers (CHWs) treated children with malaria and pneumonia (intervention areas) and where they treated children with malaria only (control areas). Children treated by CHWs were enrolled on the day of seeking treatment from CHWs (609 intervention, 667 control) and demographic, illness, and treatment seeking information was collected. Further information on illness and treatment outcomes was collected on day four. The primary outcome was prompt and appropriate antibiotics for pneumonia symptoms and the secondary outcome was treatment outcomes on day four. Results: Children in the intervention areas were more likely to receive prompt and appropriate antibiotics for pneumonia symptoms compared to children in the control areas (RR = 3.51, 95%CI = 1.75-7.03). Children in the intervention areas were also less likely to have temperature ≥37.5°C on day four (RR = 0.29, 95%CI = 0.11-0.78). The decrease in fast breathing between day one and four was greater in the intervention (9.2%) compared to the control areas (4.2%, p-value = 0.01). Conclusions: Integrated community management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms and improves treatment outcomes. Trial registration: ISRCTN: ISRCTN52966230 Keywords: CHW, ICCM, Health System Research, Prompt treatment, Appropriate treatment, Treatment outcomes, Malaria, Pneumonia, Children, CMDs * Correspondence: nakayaga2001@yahoo.com Department of Public Health Sciences, Global Health (IHCAR), Karolinska Institutet, SE 17177, Stockholm, Sweden Clinical Epidemiology Unit, Makerere University College of Health Sciences, P.O. Box 7072, Kampala, Uganda Full list of author information is available at the end of the article © 2013 Kalyango et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Kalyango et al. Malaria Journal 2013, 12:340 Page 2 of 13 http://www.malariajournal.com/content/12/1/340 Background setting, health care was provided by several drug shops Although global mortality in children less than five years and private clinics in addition to the CHWs, government reduced from 12 million deaths in 1990 to 6.9 million and non-governmental organization (NGO) health units. deaths in 2011 [1], the number of deaths is still high and Although private health facilities are frequently used in many countries are at risk of not achieving Millennium Uganda, they are often manned by unqualified personnel Development Goal four [2]. This is more so in Africa [15] and may provide poor quality of health care for and South Asia where about 80% of the deaths are con- childhood illnesses [16]. It was, therefore, not clear centrated and the reduction in mortality is lower than whether integrated malaria and pneumonia care from elsewhere. Sub-Saharan Africa has a mortality reduction CHWs would increase prompt and appropriate treat- rate of 2.4% per annum which is insufficient to achieve ment for pneumonia symptoms given the complex the desired two-thirds reduction in mortality from 1990 public-private mix of health care. The primary aim of to 2015 [2]. The main causes of mortality in these re- this study was to determine the effect of integrated mal- gions include neonatal conditions, malaria, pneumonia, aria and pneumonia treatment on receiving prompt and and diarrhoea, with malnutrition being an important appropriate antibiotics for pneumonia symptoms. In contributing factor. These conditions could be addressed addition, factors associated with receiving prompt and with low cost interventions but children do not access appropriate antibiotics for pneumonia symptoms were the effective interventions promptly [1]. Many more lives determined and treatment outcomes in the intervention could be saved through interventions that improve and control areas were compared. access to effective prevention and treatment for com- mon childhood illnesses. In order to improve access, Methods community-based treatment of illnesses has been Study design recommended [3]. A follow-up study was nested within a cluster-randomized The home management of malaria (HMM) was one trial that compared community-based management of such strategy designed to provide anti-malarials for malaria alone to the integrated community management treatment of fevers at the community level. It provided of malaria and non-severe pneumonia. A cluster design prompt treatment for malaria [4] but children with non- was considered appropriate because of the behavioural malarial febrile illnesses like pneumonia which present nature of the intervention. Therefore groups of villages with symptoms similar to malaria [5] could have been (parishes) formed the clusters that were randomized to inappropriately treated with anti-malarials [6]. The the intervention or control rather than individuals. Al- World Health Organization (WHO) and UNICEF have though the current study is not typically a randomized now recommended the provision of integrated manage- controlled trial, but was nested in an ongoing cluster- ment of common childhood illnesses at the community randomized trial, the consort checklist [17] has been used level [7]. Following this recommendation, Uganda to guide the report as much as possible. adopted the integrated community case management of childhood illnesses (ICCM) under which community health workers (CHWs) will provide malaria, pneumo- Study area nia, and diarrhoea treatment for children less than five The study was conducted in eastern Uganda in an area years [8]. It is hoped that the ICCM strategy will build covered by the Iganga-Mayuge Health and Demographic on the achievements of the HMM intervention by also Surveillance Site (HDSS). Iganga-Mayuge HDSS is pre- improving access to treatment for pneumonia and diar- dominantly rural with a few peri-urban areas mainly in rhoea as well as improving treatment outcomes. the trading centres and Iganga Municipality. It has a Preliminary studies on integrated malaria and pneu- total population of about 70,000 people and about monia treatment in Ghana, Zambia and Uganda have 11,000 of these are children under five years. The main assessed effects on mortality [9], treatment-seeking, drug occupation of the people is subsistence farming. Malaria use [10,11], timely treatment of pneumonia, and treat- is endemic in the area with transmission peaks occurring ment outcomes where malaria diagnosis is done with the in the rainy seasons in the months of March and aid of rapid diagnostic tests (RDTs) [12]. However, there September; and respiratory tract illnesses are common. is still need for more evidence to inform the implemen- The HDSS is served by one government hospital, tation of the ICCM strategy, including more data on nine public health centres, three non-governmental child health outcomes [13,14] in various contexts. organization hospitals and 122 drug shops and private Uganda has a complex public-private mix of health care clinics. There were also 131 CHWs (locally known as provided by government, private-not-for-profit (PNFP), community medicine distributors (CMDs)) who had and private-for-profit (PFP) facilities including private been providing health care since 2009 under the cluster clinics, drug shops and a few hospitals. In the study randomized trial on which the current study is based. Kalyango et al. Malaria Journal 2013, 12:340 Page 3 of 13 http://www.malariajournal.com/content/12/1/340 The intervention were trained first in the community management of The cluster-randomized trial involved treatment of chil- malaria and then in the integrated community manage- dren aged 4–59 months by CHWs. CHWs have provided ment of malaria and pneumonia. This training was treatment for malaria in Uganda since 2002 under the conducted by Ministry of Health officials together with home-based management of fever strategy. They initially the study investigators. Health workers in the public, used a combination of chloroquine and sulphadoxine- non-governmental organization and private facilities also pyrimethamine until the malaria policy change in 2004 received a two-day training in integrated community that led to use of artemisinin-based combination therapy management of malaria and pneumonia. They were ori- (ACT) as first-line treatment for malaria [18]. The im- ented on the algorithms that were to be used by the plementation of the policy was slow and use of ACT by CHWs. In addition, they were trained on investigating CHWs had not yet been widely rolled out when the and documenting adverse events, and supervision and ICCM strategy was adopted in 2010. Therefore, there training of CHWs. were no CHWs in the study area treating children with ACT at the time of commencement of the cluster- Patient management and follow-up randomized trial in 2009. The CHWs were selected as part The CHWs in the control arm treated children with of the cluster-randomized trial and trained either in the anti-malarials, but referred children with danger signs or management of malaria alone (which constituted the con- those with pneumonia symptoms, regardless of severity, trol arm) or integrated management of malaria and pneu- to nearby health facilities. The CHWs in the interven- monia (which constituted the intervention arm). The tion arm treated children with anti-malarials and/or an- selection of CHWs has been described elsewhere [19]. tibiotics as appropriate and referred children with danger signs to nearby health facilities. No pre-referral Randomization medicines were given to children that were referred with Randomization was done by a statistician that was inde- danger signs. The diagnoses of malaria and pneumonia pendent of the study using stratified block randomization. were made using symptoms according to the Integrated Iganga-Mayuge HDSS has 65 villages which make up 26 Management of Childhood Illness (IMCI) criteria [20]. parishes that were divided into eight urban and 18 rural The details of the treatment algorithm have been de- clusters (parishes). The clusters from the rural area were scribed elsewhere [11,19] but briefly, a diagnosis of “mal- further grouped into three strata based on the population aria” was made if the child had fever or the caregiver size of children less than five years: i) 190–320, ii) 321– reported history of fever in the previous 24 hours. The 390, and iii) 391 and above, resulting in six clusters in each CHWs did not use rapid diagnostic tests (RDTs) because of these strata. The clusters from the urban area were they commenced their roles under the cluster- grouped into two strata based on population sizes of iv) randomized trial in 2009 before WHO recommended 280–430, and v) 431 and above. Random numbers were parasite-based diagnosis for malaria in 2010 [21]. “Pneu- generated in blocks of six for the rural clusters and in monia” was diagnosed following the presence of cough blocks of four for the urban clusters. From each of the or difficult breathing and fast breathing (≥50 breaths per three strata in the rural area three clusters were random- minute in children aged four to 12 months and ≥40 ized to the intervention arm and three to the control arm. breaths per minute in children 12 to 59 months). In each of the two strata in the urban area, two clusters Breathing rates were counted using wrist watches with were randomized to the intervention arm and two clusters minutes and seconds hands. Children were classified as to the control arm. Health facilities in the setting are having severe disease if they presented with any of the equally distributed, therefore, geography and distance to four general danger signs: convulsions, repeated formal health care was not used in the randomization vomiting, lethargy/unconsciousness or failure to feed, or process. other danger signs: chest in-drawings, noisy breathing, severe dehydration or pallor. The CHWs were required Training to follow up the children they treated and refer those All CHWs received three days’ training in the manage- that did not get well to nearby health facilities. Although ment of malaria. This training was done before diarrhoea is one of the illnesses targeted by the ICCM randomization. After randomization, the CHWs in the strategy, the CHWs in this cluster-randomized trial did intervention areas received a further three days’ training not treat it. This was because the trial commenced be- in the integrated management of malaria and pneumo- fore the adoption of the ICCM strategy and was in- nia. In addition, the CHWs in both arms received formed mainly by studies that had shown necessity to monthly refresher training. The details of the CHW integrate malaria and pneumonia treatment due to training have been described elsewhere [19]. Prior to the symptom overlap. The study reported in this paper was training of CHWs, the district health team members conducted over a period of five weeks that started in Kalyango et al. Malaria Journal 2013, 12:340 Page 4 of 13 http://www.malariajournal.com/content/12/1/340 mid-October to November 2011, two years after CHWs included in the analyses for prompt and appropriate an- commenced their treatment roles under the cluster ran- tibiotics for pneumonia symptoms. domized trial. Sample size Study supplies The sample size was estimated using the formula for CHWs in the intervention arm were supplied with sample size for comparing two proportions with adjust- artemether-lumefantrine (AL) and amoxicillin while ments for clustering [24]. The assumptions used were: CHWs in the control arm were supplied with AL only. 90% power, 5% level of significance that was two-sided, The drugs were refilled at the monthly meetings of coefficient of variation of the proportions between clus- CHWs with the cluster-randomized trial staff and formal ters within each group of 0.4, proportions of children re- health workers. The formulation of AL used was pre- ceiving prompt treatment for pneumonia of 13% in the packaged dispersible Coartem® (Novartis Pharma AG). control arm and 68% in the intervention arm from a This was supplied in packs of six tablets (20 mg previous study [12], 20% of children having pneumonia artemether, 120 mg lumefantrine) for children aged symptoms, a minimum sample size of 12 children per 4–35 months or 12 tablets for children aged ≥ 36 months. cluster and a drop-out rate of 10%. The coefficient of Dispersible amoxicillin tablets (125 mg) (Medophar, variation was assumed from literature suggesting that India or IDA Foundation, Netherlands) were pre- values of coefficients of variation are usually ≤0.25 and packaged into three age-based doses: 6 tablets for chil- seldom exceed 0.5 [24], therefore 0.4 was chosen to fall dren aged 4–11 months (one tablet twice daily for three within this range of acceptable values. Based on these days), 12 tablets for children 12–35 months (two tablets assumptions about 111 children with pneumonia symp- twice daily for three days), and 18 tablets for children toms were needed in each arm, translating to about 610 36–59 months (three tablets twice daily for three days). children treated by CHWs in each arm. The dosages used in the cluster-randomized trial were chosen after some studies showed that three-day dosing Definition of pneumonia symptoms was equally as effective as five-day dosing [22,23]. The Pneumonia symptoms were defined as caregiver reports amoxicillin tablets were procured in bulk and pre- of cough accompanied by fast and/or difficult breathing packaged by a local pharmaceutical industry (Kampala similar to the definition used in MICS and DHS surveys Pharmaceutical Industries) into the age specific doses. [25]. This definition of pneumonia symptoms based on The drugs were procured from the manufacturers caregiver reports of symptoms was used instead of the through local pharmaceutical distributors and distrib- more specific IMCI definition of pneumonia that the uted through the district system. The CHWs in the CHWs use, because the CHWs in the control arm did intervention arm were supplied with watches that they not assess and classify pneumonia symptoms and there- used for respiratory rate counting. Neither the CHWs in fore had no record of which children had pneumonia the intervention arm nor the ones in the control arm symptoms. In addition, the children could not be classi- had thermometers. fied by the field team of the current study because they were identified after they had received treatment from Supervision the CHWs and the symptoms and signs were likely to CHWs in both arms received monthly support super- have changed by the time they were enrolled into the vision from health workers based at the nearest health study. The definition has been used elsewhere to define facility. Details of the supervision are reported elsewhere presumed pneumonia [26]. [19]. Study outcomes Study participants The primary outcome of the study was the proportion of All children aged 4–59 months in the intervention and children receiving prompt and appropriate antibiotics control areas that were treated by CHWs for any illness for pneumonia symptoms. Prompt treatment was de- were eligible for inclusion, and they were consecutively fined as receiving the first dose of treatment on the day enrolled into the study. The children were identified of presentation of symptoms or the next day. If the child from the CHWs’ registers and traced to their homes on received an appropriate antibiotic for pneumonia the day they sought care from the CHW (day 1) or the promptly then it was classified as prompt and appropri- next day for those that sought treatment late in the ate. The following medicines were considered appropri- evening or at night (this was still within the first twenty ate for pneumonia as recommended in the national or four hours of seeking care). All enrolled children were CHWs’ treatment guidelines or the British National included in the analysis for the treatment outcomes Formulary which is widely used in Uganda: amoxicillin, while only children with pneumonia symptoms were erythromycin, azithromycin, ampicillin, ampicillin plus Kalyango et al. Malaria Journal 2013, 12:340 Page 5 of 13 http://www.malariajournal.com/content/12/1/340 cloxacillin, gentamicin, benzyl penicillin which could be Data management and analysis switched to oral amoxicillin, procaine penicillin fortified The data were double entered into FoxPro and exported (PPF), which could be switched to amoxicillin, chloram- to STATA 10 (STATA Corp, College Station, TX, USA) phenicol, cotrimoxazole, cefuroxime, and amoxycillin for analysis. The baseline characteristics of the children plus clavulanate [27-29]. The assessment of appropriate and caregivers were summarized using descriptive statis- antibiotics did not take into account appropriate doses, tics. Proportions of the study outcomes (prompt and ap- frequency, and duration mainly because most of the chil- propriate antibiotics and treatment outcomes) were dren that received antibiotics from sources other than the compared in the intervention and control arms and CHWs would have just started these treatments by the crude relative risks were estimated using Mantel- day four evaluation. It would therefore not be possible to Haenszel methods that considered the strata used in the determine if they took the medicines appropriately and for randomization. In order to simultaneously account for the correct length of time. Promptness of treatment was the effect of the intervention and covariates on the pri- assessed based on data collected on day one while appro- mary outcome, individual level analysis was done using priateness of antibiotics was based on day four data. generalized estimating equations based on Poisson re- The secondary outcome was treatment outcome which gression with robust variance estimation and an ex- included: having temperature ≥ 37.5°C or high respiratory changeable correlation matrix that accounted for rate on day four; received anti-malarials from other health clustering [30]. The individual level analysis was done providers after CHW-treatment or antibiotics additional after establishing that the results obtained using individ- to those provided by the CHWs in the intervention arm ual level analysis did not differ much from those or additional to those provided by other health providers obtained using cluster level analysis since there were few to whom children with pneumonia symptoms in the con- clusters in each arm (less than the 15 recommended for trol arm were referred; hospitalization; or death [12]. appropriate use of individual level analysis). Analysis was Treatment outcomes were assessed on day four. Children done on intention-to-treat basis. In order to select co- that had any of the treatment outcomes were classified as variates for inclusion into the multivariate model, uni- having treatment failure. In addition, self-reported treat- variate analysis was done to determine associations ment failure, defined as caregiver report of non-resolution between the various factors on which data were col- of illness symptoms, was assessed on day four. lected and receiving prompt and appropriate antibiotics using generalized estimating equations with Poisson re- Data collection gression. The factors with p-values less than 0.2 at uni- The data were collected by a team of 30 experienced variate analysis were considered for inclusion into the data collectors that were independent of the cluster- multivariate model. These factors were assessed for stat- randomized trial. They were trained on the current istical interaction and confounding with the interven- study for one week. They were supervised by two med- tion. The differences in proportions of children with fast ical officers and the first author (JNK). The data were breathing between day one and four were compared in collected using a pre-tested questionnaire (divided into the intervention and control arms using z-tests for com- two parts) that was presented in both English and parison of proportions. Lusoga (the main local language spoken in the area) allowing the caregiver to choose between the two lan- Ethical considerations guages. Part one of the questionnaires was administered The study was approved by Makerere University School on day one since receiving treatment from the CHW of Public Health Higher Degrees Research and Ethics and collected data on: socio-demographic characteristics Committee (reference IRB00005876) and the Uganda of the caregivers and children, children’s presenting National Council of Science and Technology (reference symptoms, time between onset of symptoms and seeking HS 898). Permission to conduct the study in the area care, and treatment received from the CHW. Part two of was also obtained from the administration of Iganga- the questionnaire was administered on day four to col- Mayuge Health and Demographic Surveillance Site and lect data on: medicines obtained from other sources, the local administration of the villages where the chil- counselling reported by the caregiver, adherence to med- dren were enrolled from. The caregivers of the children icines, hospitalization during current illness episode, and were informed about the study and then written in- resolution of symptoms as reported by the caregiver. formed consent was obtained from them to have their Temperature and breathing rates of the children were children enrolled into the study. The caregivers were measured on day one and day four. Axillary tempera- given one of the two copies of informed consent docu- tures were measured using digital thermometers. The ments that were signed by them and the person breathing rates were counted by the field assistants with conducting the informed consent. Access to the data the aid of wrist watches. was restricted to the investigators. Kalyango et al. Malaria Journal 2013, 12:340 Page 6 of 13 http://www.malariajournal.com/content/12/1/340 Results caregivers in the control arm were living in rural areas A total of 1,276 children were enrolled into the study (90% control, 82% intervention). (667 from the control arm, 609 from the intervention arm) as shown in the study profile (Figure 1). The children Illness characteristics of the children were enrolled from all clusters of the cluster-randomized Almost all children in both arms (97%) presented with a trial but in varying numbers (17–138 children per cluster) history of fever. However, only 15% in the intervention depending on the number of children that were treated by and 14% in the control arm had temperature ≥ 37.5°C at the CHWs during the study period. Children were sam- the time of day one data collection. Slightly more than pled from 98 of the 131 CHWs (75%) in the area. 80% of the children in both arms had cough and about 29% in both arms reported difficult breathing. Fast Demographic characteristics of the children and breathing in children was reported by 30% of the care- caregivers givers in both arms. However, based on the respiratory The demographic characteristics of the children and rates counted by the field assistants, a higher proportion caregivers were comparable in the intervention and con- of children in the intervention arm (20%) had fast trol arms (Table 1). About half of the children in both breathing on the day one visit compared to the control arms were female with a mean age of about 30 months. arm (11%). The proportions of children with reported The mean age of the caregivers was about 31 years in diarrhoea, convulsions and vomiting were comparable in both arms and most of them were female, married, with the two arms (Table 2). primary level education. Slightly higher proportions of Figure 1 Study profile. Kalyango et al. Malaria Journal 2013, 12:340 Page 7 of 13 http://www.malariajournal.com/content/12/1/340 Table 1 Demographic characteristics of children treated prompt and appropriate antibiotics (RR = 3.47, 95%CI = by CHWs and their caregivers in Iganga-Mayuge 2.41-4.99) (Table 3). demographic surveillance site The association between the intervention (integrated Characteristic Intervention arm Control arm malaria and pneumonia management) and receiving N = 609 N = 667 prompt and appropriate antibiotics was significant also at adjusted analysis (RR = 3.51, 95% CI = 1.75-7.03). The Children’s characteristics adjusted analysis included factors that were statistically Female sex, n (%) 329 (54.0) 336 (50.4) significant or were retained as confounders, and these Mean age, months (SD) 30.4 (15.9) 31.5 (16.4) included: caregiver having post-primary education (RR = Caregivers’ characteristics 1.56, 95% CI = 1.04-2.36), child presenting with no diar- Female sex, n (%) 562 (92.3) 625 (93.7) rhoea (RR = 1.33, 95% CI = 1.01-1.76) and living in urban Age in years, mean (SD) 30.9 (9.4) 30.5 (8.7) residence (RR = 1.56, 95% CI = 1.20-2.00) (Table 4). Education level, n (%)* None 59 (9.7) 62 (10.4) Treatment outcomes Primary 381 (62.7) 450 (67.7) There was no difference in overall treatment failure Post primary 168 (27.6) 146 (22.0) among children in the intervention and control arms Religion, n (%)*# (RR = 0.88, CI = 0.67-1.18). However, children in the Catholic 30 (4.9) 65 (9.8) control arm were more likely to have temperatures above 37.5°C on day four compared to children in the Protestant 244 (40.1) 223 (33.5) intervention arm (4% versus 1%). In addition, there was Muslim 310 (51.0) 335 (50.4) a higher difference between proportions of children with Other 24 (4.0) 42 (6.3) fast breathing on day one and four in the intervention Marital status, n (%)** (9.2%) compared to control areas (4.2%, p = 0.01). There Married 559 (92.1) 592 (88.8) were no deaths reported. Perceived treatment failure was Single 23 (3.8) 27 (4.1) significantly lower in the intervention arm compared to the control arm (RR = 0.58, 95%CI = 0.46-0.73) (Table 5). Divorced 19 (3.1) 37 (5.6) Among children with pneumonia symptoms, children Widowed 6 (1.0) 11 (1.7) in the control arm were more likely to receive additional Rural residence, n (%) 499 (81.9) 600 (90.0) anti-malarials (3% control, 0% intervention). * missing data (2 control, 1 intervention). ** missing data (2 intervention). # significantly different. Additional medicines received by children Prompt and appropriate treatment for pneumonia The use of additional medicines was infrequent. Among symptoms the children that received additional medicines after re- In the intervention arm, 63% of the children with pneu- ceiving care from the CHW, paracetamol was the most monia symptoms received antibiotics while in the con- frequently used medicine in children from both the trol arm 51% were referred to health facilities. However, intervention (76%) and control (74%) arms followed by 45% of children with pneumonia symptoms in the inter- cotrimoxazole (14% intervention, 16% control) and vention arm and 11% in the control arm received cough syrups (10% intervention, 6% control) (Table 6). Table 2 Illness characteristics of 1276 children treated by CHWs in Iganga-Mayuge HDSS Characteristic Intervention Control RR (95% CI) History of fever, n (%) 589 (96.7) 647 (97.0) 1.00 (0.98-1.02) Temperature ≥ 37.5 C, n (%) 90 (14.8) 94 (14.1) 0.98 (0.71-1.34) Cough, n (%) 500 (82.1) 541 (81.1) 1.00 (0.94-1.06) Fast breathing by history, n (%) 182 (29.9) 199 (29.8) 0.99 (0.84-1.18) Fast breathing based on respiratory rate counted by field assistant, n (%) 122 (20.0) 76 (11.4) 1.44 (1.09-1.91) Difficult breathing by history, n (%) 174 (28.6) 194 (29.1) 0.94 (0.79-1.13) Diarrhoea by history, n (%) 193 (31.7) 218 (32.7) 0.90 (0.76-1.07) Convulsions by history, n (%) 40 (6.6) 40 (6.0) 0.95 (0.62-1.45) Vomiting by history, n (%) 232 (38.2) 262 (39.5) 0.89 (0.76-1.04) Kalyango et al. Malaria Journal 2013, 12:340 Page 8 of 13 http://www.malariajournal.com/content/12/1/340 Table 3 Prompt and appropriate antibiotics among children with pneumonia symptoms Outcome indicator Intervention Control RR (95% CI) n (%) n (%) CHW gave antibiotic/referred* 158 (63.0) 142 (50.9) 1.69 (1.23-2.32) CHW gave antibiotic/referred** 205 (81.7) 142 (50.9) 2.01 (1.49-2.71) Appropriate antibiotics from any source 166 (66.1) 69 (24.7) 2.41 (1.91-3.04) Prompt and appropriate antibiotics 114(45.4) 30 (10.8) 3.47 (2.41-4.99) * CHWs in intervention arm gave antibiotics and those in control arm referred to health facility. ** CHWs in intervention arm gave antibiotics or referred and those in control arm referred to health facility. Discussion This study has demonstrated that integrated commu- This study has demonstrated that integrated community nity case management of malaria and pneumonia in- case management of malaria and pneumonia increases creases prompt and appropriate antibiotics for self- prompt and appropriate treatment for self-reported reported pneumonia symptoms in a rural area with a pneumonia symptoms. In addition, children in the con- large number of private clinics and drug shops but trol arm were more likely to have a high temperature on which may be manned by unqualified health providers. day four of treatment seeking. There was a greater re- These findings are similar to those reported in a study in duction in fast breathing in children in the intervention Zambia where prompt and appropriate treatment was compared to the control arm. higher among children treated by CHWs that could treat Table 4 Factors associated with prompt and appropriate treatment among children with self-reported pneumonia symptoms in Iganga-Mayuge DSS Characteristic(N)+ Prompt & appropriate, n(%) Unadjusted RR (95%CI) P-value Adjusted RR* (95%CI) P-value Cluster arm† Intervention (251) 114(45.4) 3.60 (1.92-6.73) <0.001 3.51 (1.75-7.03) <0.001 Control (279) 30(10.8) 1.00 1.00 Primary caregiver Male (35) 15(42.9) 1.45(0.85-2.50) 0.18 Female (495) 129(26.1) 1.00 Education, caregiver† ≤Primary (415) 96 (23.1) 1.00 0.001 1.00 0.03 Post primary (114) 48 (42.1) 1.73 (1.25-2.40) 1.56 (1.04-2.36) No occupation/housewife/farmer (434) 97(22.4) 1.00 0.013 Other (96) 47(49.0) 1.85 (1.14-3.01) Child had diarrhoea† Yes (193) 44(22.8) 1.00 0.006 1.00 0.04 No (337) 100(29.7) 1.46 (1.12-1.90) 1.33 (1.01-1.76) Child had convulsions Yes (43) 10(23.3) 1.00 0.09 No (486) 133(27.4) 1.31 (0.96-1.78) Child was vomiting Yes (219) 69(31.5) 1.20 0.11 No (310) 75(24.2) 1.00 (0.96-1.52) Residence† Urban (79) 43(54.4) 2.08 (1.25-3.45) 0.005 1.56 (1.20-2.00) 0.001 Rural (451) 101(22.4) 1.00 1.00 +N represents the number in each category. * Adjusted analysis includes all variables with †. Kalyango et al. Malaria Journal 2013, 12:340 Page 9 of 13 http://www.malariajournal.com/content/12/1/340 Table 5 Treatment outcomes of children treated by CHWs in Iganga-Mayuge DSS Characteristic Intervention Control RR (95% CI) All children treated by CHWs N = 609 N = 667 Overall treatment failure 90 (14.8%) 101 (15.1%) 0.88 (0.67-1.18) Received additional* antibiotics 7 (1.2%) 13 (2.0%) 0.77 (0.32-1.86) Received additional** anti-malarials 4 (0.7%) 10 (1.5%) 0.44 (0.15-1.28) Hospitalization 10 (1.6%) 9 (1.4%) 1.47 (0.64-3.38) Temperature ≥ 37.5°C on day 4 7 (1.2%) 23 (3.5%) 0.29 (0.11-0.78) Had fast breathing on day 4 66 (10.8%) 48 (7.2%) 1.29 (0.87-1.92) Difference in fast breathing, day 1& 4† 9.2% 4.2% p-value = 0.01 Self reported treatment failure 83 (13.7%) 170 (25.5%) 0.58 (0.46-0.74) Children with malaria symptoms Characteristic Intervention Control RR (95% CI) n = 589 n = 647 Overall treatment failure in malaria 84 (14.3%) 99 (15.3%) 0.85 (0.64-1.13) Received additional* antibiotics 6 (1.0%) 13 (2.0%) 0.65 (0.26-1.65) Received additional** anti-malarials 3 (0.5%) 10 (1.6%) 0.35 (0.11-1.11) Hospitalization 9 (1.5%) 9 (1.4%) 1.36 (0.59-3.16) Temperature ≥ 37.5°C on day 4 7 (1.2%) 22 (3.4%) 0.30 (0.11-0.81) Had fast breathing on day 4 62 (10.5%) 48 (7.4%) 1.22 (0.82-1.82) Difference in fast breathing, day 1& 4,† 9.5% 4.0% p-value = 0.006 Self reported treatment failure 78 (13.3%) 168 (26.0%) 0.56 (0.44-0.71) Children with self reported pneumonia symptoms Characteristic Intervention Control RR (95% CI) n = 251 n = 279 Overall treatment failure in “pneumonia” 43 (17.1%) 56 (20.1%) 0.77 (0.52-1.13) Received additional* antibiotics 2 (0.8%) 10 (3.6%) 0.30 (0.08-1.21) Received additional** anti-malarials 0 (0) 8 (2.9%) - Hospitalization 4 (1.6%) 3 (1.1%) 2.42 (0.60-9.84) Temperature ≥ 37.5°C on day 4 5 (2.0%) 11 (3.9%) 0.41 (0.10-1.75) Had fast breathing on day 4 34 (13.6%) 25 (9.0%) 1.30 (0.76-2.21) Difference in fast breathing, day 1& 4 † 18.3% 3.6% p-value < 0.001 Self reported treatment failure 22 (8.8) 74 (26.5) 0.35 (0.22-0.56) * Additional antibiotics received after CHW treatment in intervention arm or after treatment from other source where children were referred in control arm. The CHWs in the control arm did not have antibiotics and children with pneumonia symptoms are expected to receive antibiotics from other health providers. ** Additional anti-malarials received after treatment by CHWs. † difference between percentage of children with fast breathing on day one and four, p-value comparing difference in intervention and control. both malaria and pneumonia [12]. This similarity is participation in the study at the time of seeking care. In found despite the differences in study designs and con- addition, the current study was done in a setting with texts of the Zambian and the current study. In the several drug shops and clinics which are likely to be add- Zambian study, RDTs were employed for malaria diag- itional sources of antibiotics. As expected in such a set- nosis which was likely to have improved the accuracy ting, the difference between the intervention and control of the illness classification. The Zambian study was arms was lower compared to that in the Zambian study also in more controlled conditions where children were though not significantly. recruited as they sought care from the CHWs compared The proportion of children that received prompt and to the current study which is better classified as an ef- appropriate antibiotics was higher among children fectiveness study because children were recruited after whose caregivers had received post primary education. they sought care and were therefore not aware of their This may be due to better recognition of symptoms Kalyango et al. Malaria Journal 2013, 12:340 Page 10 of 13 http://www.malariajournal.com/content/12/1/340 Table 6 Additional medicines received by children after caregivers with primary or no education [32] that were treatment by CHW mostly living in the rural areas. Caregivers in urban Medicine Intervention Control areas have better knowledge of childhood illness and also have better health-seeking behaviour [33]. Dispar- (n = 42) (n = 96) ities in access to health care have been observed between Paracetamol 32 (76.2) 71 (74.0) rural and urban areas [34]. Cotrimoxazole* 6 (14.3) 15 (15.6) There was no difference in overall treatment failure Cough syrups 4 (9.5) 6 (6.3) among all children treated by CHWs in the intervention Amoxicillin* 1 (2.4) 5 (5.2) and control areas, similar to a study in Zambia [12]. Chlorphenramine 6 (14.3) 5 (5.2) However, contrary to the Zambian study where the pro- Metronidazole 2 (4.8) 4 (4.2) portion of children with persistent fever, fast or difficult breathing at follow up did not differ, the current study Quinine† 0 (0) 3 (3.0) found a higher proportion of febrile children in the con- Ampicillin* 0 (0) 2 (2.0) trol arm on day four despite both groups having similar Artemether-lumefantrine† 3 (7.1) 2 (2.0) proportions of children with high temperature on day Benzylpenicillin* 0 (0) 2 (2.0) one. In addition, although the children in the interven- Chloramphenicol* 0 (0) 2 (2.0) tion arm had a higher proportion of children with fast Chloroquine† 1 (2.4) 2 (2.0) breathing on day one and day four (although not signifi- cant on day four), there was a greater reduction in the Combination flue medicines 0 (0) 2 (2.0) proportions of children with fast breathing in the inter- Dexamethasone 1 (2.4) 2 (2.0) vention arm. The differences seen in the two studies Unknown antimalarial syrup† 0 (0) 2 (2.0) could have been a result of the way the outcomes were Ampiclox* 0 (0) 1 (1.0) presented. In the Zambian study persistent fever and fast Anti-emetic 1 (2.4) 1 (1.0) or difficult breathing were grouped together which could Cetrizine 0 (0) 1 (1.0) have masked outcome-specific differences. The initial higher fast breathing in the intervention arm could have Erythromycin* 0 (0) 1 (1.0) resulted from more caregivers in the control arm whose Ibuprofen 1 (2.4) 1 (1.0) children had pneumonia symptoms bypassing the CHWs Mebendazole 0 (0) 1 (1.0) and taking their children directly to other health pro- ORS 1 (2.4) 1 (1.0) viders. Similar differences in proportions of children Vitamins 1 (2.4) 0 (0) with respiratory symptoms were also noted in the study * considered as additional antibiotics. in Zambia. These findings suggest improved treatment † considered as additional anti-malarials. outcomes when children can access effective treatments promptly. Many of the children in the control arm that got antibiotics received cotrimoxazole against which among the more educated caregivers or better afford- high levels of resistance of pneumonia-causing bacteria ability of treatment resulting in more prompt care seek- have been reported [35,36]. This finding supports the ing. The study in Zambia found a tendency to more need to have the effective medicines for common child- prompt and appropriate treatment among caregivers hood illnesses at the first level of care because the chil- with primary or secondary education compared to those dren need them. This may prevent the delays that did not have any education although it was not sta- experienced in receiving necessary treatment. tistically significant [12]. Differences in receiving prompt There was no difference in overall treatment failure and appropriate antibiotics may point to inequalities that among children with pneumonia symptoms contrary to may exist in accessing health care in persons of different the Zambian study [12] where children classified as hav- social standing. Low education has been associated with ing pneumonia in the intervention arm were less likely low income and low knowledge on aspects that affect to have treatment failure compared to their counterparts use of health services [31]. in the control arm. The difference in findings could be a The children in urban areas were more likely to re- result of how pneumonia was classified in the two stud- ceive prompt and appropriate antibiotics. This may be ies. The current study used caregiver reports of pneumo- due to easier access to treatment in the urban areas. It nia symptoms whereas in the Zambian study the may also be related to caregivers’ knowledge of disease classification of pneumonia was based on the CHWs symptoms. There were higher proportions of caregivers classification; which may have led to lower misclassifica- with post primary education in the urban areas and tion of children as having pneumonia when they did not these may have better health knowledge than the [37]. This would make it easier to detect differences in Kalyango et al. Malaria Journal 2013, 12:340 Page 11 of 13 http://www.malariajournal.com/content/12/1/340 treatment failure in children treated by CHWs who the intervention and control arms compared to the re- treated and those that did not treat pneumonia. How- sults obtained using self reported pneumonia symptoms. ever, reduction in fast breathing (based on respiratory This suggests that the use of self reported pneumonia rates measured by the data collectors) between day one symptoms in the current study could have led to under- and four among children with pneumonia symptoms in rather than over-estimation of the effect of integrated the intervention arm of the current study (18%) com- malaria and pneumonia management because self- pared to reduction in fast breathing in the control arm reported pneumonia symptoms may be more sensitive (3.6%) is even higher than that seen among all children but less specific than the respiratory rate assessed by treated by CHWs (9.2% versus 4.2% respectively). data collectors on day one. While this study assessed only symptoms and treat- The analysis presented was done at the individual level ment of malaria and pneumonia, current WHO recom- whereas it would have been better to do cluster level mendations are to integrate also diarrhoea treatment analysis since there were fewer than 15 clusters in each into integrated community case management. The find- arm. However with analysis at the cluster level it would ings from this study support this recommendation to in- have been difficult to simultaneously control for other corporate ORS and zinc since a considerable proportion variables [30]. The results obtained from analysis at the of children had diarrhoea (about 30%) but only two chil- cluster and individual levels at unadjusted analysis were dren received ORS and likely no child zinc. compared and found similar. Adjusted analysis based on residuals for the primary outcome was done and found Methodological issues that it was still significant. The analysis done at the indi- Self-reported symptoms were used to classify children vidual level with adjustments for the cluster into whether they had possible pneumonia and thus randomization to show effects of other variables has, children may have been misclassified. The IMCI guide- therefore, been presented. In addition, there were imbal- lines, which CHWs use to make classifications, define ances in the number of children treated in the clusters pneumonia as cough or difficult breathing and fast which could have resulted in loss of power. However, breathing [20]. The classification of fast breathing is the effect may have been minimal since a difference in based on age-specific thresholds of breath counts which the main outcome variable has been detected between could not be ascertained based on the caregiver reports. the intervention and control areas. Nevertheless, other It was not possible to take the baseline assessments of associations may have been undetectable. respiratory rates in the children because they were lo- There were no buffer zones between the intervention cated on day one of treatment-seeking and for some of and control areas which may have led to dilution of the the children the illness may have changed by the time intervention effects since some caregivers in the control they were seen. In addition, the CHWs’ records could areas could have accessed antibiotics from the interven- not be used for classification of pneumonia because in tion areas. However, a three-fold difference between the the control arm, the CHWs do not assess and classify intervention and control areas is still demonstrated on pneumonia symptoms. They had no record of which receiving prompt and appropriate antibiotics. children had presented with pneumonia symptoms. Use Treatment failure may have been underestimated be- of caregivers’ reports to classify pneumonia symptoms cause danger signs of illness were not assessed on day has been used by Multiple Indicator Cluster Surveys four. This could have underestimated treatment effects (MICS) and Demographic and Health surveys (DHS) as in the integrated care arm. well as other studies to identify children with possible Another limitation of this study is that it did not in- pneumonia in the community [25,26], but it is likely to clude all children treated by the CHWs in the cluster- over-estimate the proportion of children with true pneu- randomized trial but only assessed those that were treated over a period of five weeks. There may have been monia. Therefore, the results of appropriate treatment obtained may reflect over-treatment with antibiotics variations in outcomes of children treated over the [37]. The desirable treatment rate may therefore not be whole year. 100%, until a more specific indicator has been found [37]. However, this measure though unspecific, does not Conclusion introduce differential bias between the study arms, thus Implementation of integrated community case manage- allowing for relative comparisons. A further classifica- ment is an effective approach in improving access to tion of pneumonia was done based on the respiratory prompt and appropriate antibiotics for pneumonia rate assessments of children on day one by the field as- symptoms and may improve treatment outcomes. This sistants. The results showed an even bigger difference in may reduce mortality in children less than five years the proportion of children that received prompt and ap- since pneumonia is one of the major killers of children propriate treatment for pneumonia symptoms between globally. Kalyango et al. Malaria Journal 2013, 12:340 Page 12 of 13 http://www.malariajournal.com/content/12/1/340 Abbreviations 5. Källander K, Nsungwa-Sabiiti J, Peterson S: Symptom overlap for malaria ACT: Artemisinin-based combination therapy; AL: Artemether-lumefantrine; and pneumonia—policy implications for home management strategies. CHWs: Community health workers; CMDs: Community medicine distributors; Acta Trop 2004, 90:211–214. HDSS: Health and demographic surveillance site; HMM: Home management 6. Källander K, Hildenwall H, Waiswa P, Galiwango E, Peterson S, Pariyo G: of malaria; ICCM: Integrated Community Case Management of Childhood Delayed care seeking for fatal pneumonia in children aged under five years Illnesses; IMCI: Integrated Management of Childhood Illness; NGO: Non in Uganda: a case-series study. Bull World Health Organ 2008, 86:332–338. Governmental Organization; ORS: Oral rehydration salts; PFP: Private for Profit; 7. Pneumonia: The forgotten killer of children. http://www.unicef.org/mdg/ PNFP: Private not for Profit; PPF: Procaine Penicillin Fortified; RDTs: Rapid mortalitymultimedia/Pneumonia_The_Forgotten_Killer_of_Children.pdf diagnostic tests; RR: Relative risk; SD: Standard deviation; UNICEF: United (Accessed November 18, 2012). Nations Children’s Fund; WHO: World Health Organization. 8. WHO: Child and Adolescent Health Program: Quarterly Report. Geneva: World Health Organization; 2011. Competing interests 9. Chinbuah MA, Kager PA, Abbey M, Gyapong M, Awini E, Nonvignon J, The authors declare that they have no competing interests. Adjuik M, Aikins M, Pagnoni F, Gyapong JO: Impact of community management of fever (using antimalarials with or without antibiotics) on childhood mortality: a cluster-randomized controlled trial in Ghana. Am J Authors’ contributions Trop Med Hyg 2012, 87:11–20. JNK, TA, SP, CK, and ER took part in: designing the study, development of 10. Seidenberg PD, Hamer DH, Iyer H, Pilingana P, Siazeele K, Hamainza B, study tools, data analysis and manuscript writing. KM took part in MacLeod WB, Yeboah-Antwi K: Impact of integrated community case development of study tools. JNK and KM participated in data collection. All management on health-seeking behavior in rural Zambia. Am J Trop Med authors read and approved the final manuscript. Hyg 2012, 87:105–110. 11. Kalyango JN, Lindstrand A, Rutebemberwa E, Ssali S, Kadobera D, Karamagi Acknowledgements C, Peterson S, Alfven T: Increased use of community medicine distributors We are very grateful to the CHWs and Iganga-Mayuge Health and and rational use of drugs in children less than five years of age in Demographic Surveillance Site administration and staff, most notably Mr Uganda caused by integrated community case management of fever. Edward Galiwango, Mrs Judith Kaija, Mr Michael Kibuuka, Mr Daniel Am J Trop Med Hyg 2012, 87:36–45. Kadobera, Mr Zach Ojiek, and the field team who made this study possible. 12. Yeboah-Antwi K, Pilingana P, Macleod WB, Semrau K, Siazeele K, Kalesha P, We also thank Dr Sheila Tusubira and Amanda Wanyana for their Hamainza B, Seidenberg P, Mazimba A, Sabin L, Kamholz K, Thea DM, contribution to data collection; and the data entrants for their hard work on Hamer DH: Community case management of fever due to malaria and this research. We thank the anonymous reviewer for the valuable comments pneumonia in children under five in Zambia: A cluster randomized and suggestions. controlled trial. PLoS Med 2010, 7:e1000340. 13. Hamer DH, Marsh DR, Peterson S, Pagnoni F: Integrated community case Financial support management: next steps in addressing the implementation research This study received financial support from Swedish Institute for Development agenda. Am J Trop Med Hyg 2012, 87:151–153. Agency (SIDA) and UNICEF/UNDP/World Bank/WHO Special Program for 14. Marsh DR, Hamer DH, Pagnoni F, Peterson S: Evidence for the Research and Training in Tropical Diseases. implementation, effects, and impact of the integrated community case management strategy to treat childhood infection. Am J Trop Med Hyg Disclaimer 2012, 87:2–5. The findings and conclusions in this article are those of the authors and do 15. Uganda national communication strategy for promoting rational use of not necessarily represent the views of SIDA or UNICEF/UNDP/World Bank/ medicines. http://apps.who.int/medicinedocs/documents/s16511e/s16511e. WHO Special Programme for Research and Training in Tropical Diseases. pdf (Accessed April20, 2013). 16. Awor P, Wamani H, Bwire G, Jagoe G, Peterson S: Private sector drug shops in Author details integrated community case management of malaria, pneumonia, and Department of Public Health Sciences, Global Health (IHCAR), Karolinska diarrhea in children in Uganda. Am J Trop Med Hyg 2012, 87:92–96. Institutet, SE 17177, Stockholm, Sweden. Clinical Epidemiology Unit, 17. Campbell MK, Piaggio G, Elbourne DR, Altman DG: Consort 2010 Makerere University College of Health Sciences, P.O. Box 7072, Kampala, statement: extension to cluster randomised trials. BMJ 2012, 345:e5661. Uganda. Department of Pharmacy, Makerere University College of Health 18. Malaria case management: home based management of fever. http:// Sciences, P.O. Box 7072, Kampala, Uganda. Department of Paediatrics, Sach’s www.health.go.ug/mcp/mt.html (Accessed September 10. 2010). Children’s Hospital, Södersjukhuset, Stockholm, Sweden. International 19. Kalyango JN, Rutebemberwa E, Alfven T, Ssali S, Peterson S, Karamagi C: Maternal and Child Health, Department of Women and Children’s Health, Performance of community health workers under integrated community Uppsala University, Uppsala, Sweden. African Field Epidemiology Network case management of childhood illnesses in eastern Uganda. Malar J (AFENET), P.O Box 12874, Kampala, Uganda. Department of Paediatrics and 2012, 11:282. Child Health, Makerere University College of Health Sciences, P.O. Box 7072, 20. Integrated Management of Childhood Illness chart booklet. http://whqlibdoc. Kampala, Uganda. Department of Health Policy, Planning and Management, who.int/publications/2008/9789241597289_eng.pdf (Accessed November School of Public Health, Makerere University College of Health Sciences, P.O. 20, 2011). Box 7072, Kampala, Uganda. 21. Guidelines for the treatment of malaria. http://whqlibdoc.who.int/ publications/2010/9789241547925_eng.pdf (Accessed September 15, 2012). Received: 6 June 2013 Accepted: 18 September 2013 22. ISCAP study group: Three day versus five day treatment with amoxicillin Published: 22 September 2013 for non-severe pneumonia in young children: a multicentre randomised controlled trial. BMJ 2004, 328:791. References 23. Pakistan Multicentre Amoxycillin Short Course Therapy (MASCOT) 1. UNICEF: In Commiting to child survival: A promise renewed. Edited by pneumonia study group: Clinical efficacy of 3 days versus 5 days of oral Anthony D, Mullerbeck E. New York: UNICEF; 2012:44. amoxicillin for treatment of childhood pneumonia: a multicentre 2. WHO: Health related Millennium Development Goals. Geneva: World Health double-blind trial. Lancet 2002, 360:835–841. Organization; 2011. 24. Hayes RJ, Bennett S: Simple sample size calculation for cluster- 3. CORE Group, Save the Children, BASICS and MCHIP: Community case randomized trials. Int J Epidemiol 1999, 28:319–326. management essentials: Treating common childhood illnesses in the 25. Pneumonia. www.childinfo.org/pneumonia.html (Accessed September 5, 2013). community. A guide for program managers. Washington DC: CORE group; 2010. 26. Diaz T, George A, Rao S, Bangura P, Baimba J, McMahon S, Kabano A: 4. Staedke SG, Mwebaza N, Kamya MR, Clark TD, Dorsey G, Rosenthal PJ, Healthcare seeking for diarrhoea, malaria and pneumonia among Whitty CJM: Home management of malaria with artemether-lumefantrine children in four poor rural districts in Sierra Leone in the context of free compared with standard care in urban Ugandan children: a randomised health care: results of a cross-sectional survey. BMC Publ Health 2013, controlled trial. Lancet 2009, 373:1623–1631. 13:157. Kalyango et al. Malaria Journal 2013, 12:340 Page 13 of 13 http://www.malariajournal.com/content/12/1/340 27. Department of Health Policy Planning and Management: Community Medicine Distributor training guide on Integrated Management of Malaria and Pneumonia. Kampala: School of Public Health Makerere University; 2010:24. 28. Joint Formulary Committee: British National Formulary. 53rd edition. London: British Medical Association and Royal Pharmaceutical Society of Great Britain; 2007. 29. Ministry of Health (Uganda): Uganda Clinical Guidelines. 2nd edition. Kampala: Ministry of Health, Uganda; 2010. 30. Hayes RJ, Moulton LH: Cluster randomized trials. 1st edition. Boca Raton (Florida): CRC Press; 2009. 31. Victora CG, Wagstaff A, Schellenberg JA, Gwatkin D, Claeson M, Habicht J-P: Applying an equity lens to child health and mortality: more of the same is not enough. Lancet 2003, 362:233–241. 32. Njama D, Dorsey G, Guwatudde D, Kigonya K, Greenhouse B, Musisi S, Kamya MR: Urban malaria: primary caregivers’ knowledge, attitudes, practices and predictors of malaria incidence in a cohort of Ugandan children. Trop Med Int Health 2003, 8:685–692. 33. Uzochukwu BSC, Onwujekwe EO, Onoka CA, Ughasoro MD: Rural–urban differences in maternal responses to childhood fever in south East Nigeria. PLoS One 2008, 3:e1788. 34. Kahabuka C, Kvale G, Hinderaker SG: Care-seeking and management of common childhood illnesses in Tanzania - Results from the 2010 demographic and health survey. PLoS One 2013, 8:e58789. 35. Hoa N, Trung N, Larsson M, Eriksson B, Phuc H, Chuc N, Lundborg C: Decreased Streptococcus pneumoniae susceptibility to oral antibiotics among children in rural Vietnam: a community study. BMC Infect Dis 2010, 10:85. 36. Joloba ML, Bajaksouzian S, Palavecino E, Whalen C, Jacobs MR: High prevalence of carriage of antibiotic-resistant Streptococcus pneumoniae in children in Kampala Uganda. Int J Antimicrob Agents 2001, 17:395–400. 37. Campbell H, el Arifeen S, Hazir T, O’Kelly J, Bryce J, Rudan I, Qazi SA: Measuring coverage in MNCH: challenges in monitoring the proportion of young children with pneumonia who receive antibiotic treatment. PLoS Med 2013, 10:e1001421. doi:10.1186/1475-2875-12-340 Cite this article as: Kalyango et al.: Integrated community case management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms in children under five years in Eastern Uganda. Malaria Journal 2013 12:340. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Malaria Journal Springer Journals

Integrated community case management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms in children under five years in Eastern Uganda

Loading next page...
 
/lp/springer-journals/integrated-community-case-management-of-malaria-and-pneumonia-zJ6V4t9PAJ

References (86)

Publisher
Springer Journals
Copyright
Copyright © 2013 by Kalyango et al.; licensee BioMed Central Ltd.
Subject
Biomedicine; Parasitology; Infectious Diseases; Tropical Medicine
eISSN
1475-2875
DOI
10.1186/1475-2875-12-340
pmid
24053172
Publisher site
See Article on Publisher Site

Abstract

Background: Efforts to improve access to treatment for common illnesses in children less than five years initially targeted malaria alone under the home management of malaria strategy. However under this strategy, children with other illnesses were often wrongly treated with anti-malarials. Integrated community case management of common childhood illnesses is now recommended but its effect on promptness of appropriate pneumonia treatment is unclear. Objectives: To determine the effect of integrated malaria and pneumonia management on receiving prompt and appropriate antibiotics for pneumonia symptoms and treatment outcomes as well as determine associated factors. Methods: A follow-up study was nested within a cluster-randomized trial that compared under-five mortality in areas where community health workers (CHWs) treated children with malaria and pneumonia (intervention areas) and where they treated children with malaria only (control areas). Children treated by CHWs were enrolled on the day of seeking treatment from CHWs (609 intervention, 667 control) and demographic, illness, and treatment seeking information was collected. Further information on illness and treatment outcomes was collected on day four. The primary outcome was prompt and appropriate antibiotics for pneumonia symptoms and the secondary outcome was treatment outcomes on day four. Results: Children in the intervention areas were more likely to receive prompt and appropriate antibiotics for pneumonia symptoms compared to children in the control areas (RR = 3.51, 95%CI = 1.75-7.03). Children in the intervention areas were also less likely to have temperature ≥37.5°C on day four (RR = 0.29, 95%CI = 0.11-0.78). The decrease in fast breathing between day one and four was greater in the intervention (9.2%) compared to the control areas (4.2%, p-value = 0.01). Conclusions: Integrated community management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms and improves treatment outcomes. Trial registration: ISRCTN: ISRCTN52966230 Keywords: CHW, ICCM, Health System Research, Prompt treatment, Appropriate treatment, Treatment outcomes, Malaria, Pneumonia, Children, CMDs * Correspondence: nakayaga2001@yahoo.com Department of Public Health Sciences, Global Health (IHCAR), Karolinska Institutet, SE 17177, Stockholm, Sweden Clinical Epidemiology Unit, Makerere University College of Health Sciences, P.O. Box 7072, Kampala, Uganda Full list of author information is available at the end of the article © 2013 Kalyango et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Kalyango et al. Malaria Journal 2013, 12:340 Page 2 of 13 http://www.malariajournal.com/content/12/1/340 Background setting, health care was provided by several drug shops Although global mortality in children less than five years and private clinics in addition to the CHWs, government reduced from 12 million deaths in 1990 to 6.9 million and non-governmental organization (NGO) health units. deaths in 2011 [1], the number of deaths is still high and Although private health facilities are frequently used in many countries are at risk of not achieving Millennium Uganda, they are often manned by unqualified personnel Development Goal four [2]. This is more so in Africa [15] and may provide poor quality of health care for and South Asia where about 80% of the deaths are con- childhood illnesses [16]. It was, therefore, not clear centrated and the reduction in mortality is lower than whether integrated malaria and pneumonia care from elsewhere. Sub-Saharan Africa has a mortality reduction CHWs would increase prompt and appropriate treat- rate of 2.4% per annum which is insufficient to achieve ment for pneumonia symptoms given the complex the desired two-thirds reduction in mortality from 1990 public-private mix of health care. The primary aim of to 2015 [2]. The main causes of mortality in these re- this study was to determine the effect of integrated mal- gions include neonatal conditions, malaria, pneumonia, aria and pneumonia treatment on receiving prompt and and diarrhoea, with malnutrition being an important appropriate antibiotics for pneumonia symptoms. In contributing factor. These conditions could be addressed addition, factors associated with receiving prompt and with low cost interventions but children do not access appropriate antibiotics for pneumonia symptoms were the effective interventions promptly [1]. Many more lives determined and treatment outcomes in the intervention could be saved through interventions that improve and control areas were compared. access to effective prevention and treatment for com- mon childhood illnesses. In order to improve access, Methods community-based treatment of illnesses has been Study design recommended [3]. A follow-up study was nested within a cluster-randomized The home management of malaria (HMM) was one trial that compared community-based management of such strategy designed to provide anti-malarials for malaria alone to the integrated community management treatment of fevers at the community level. It provided of malaria and non-severe pneumonia. A cluster design prompt treatment for malaria [4] but children with non- was considered appropriate because of the behavioural malarial febrile illnesses like pneumonia which present nature of the intervention. Therefore groups of villages with symptoms similar to malaria [5] could have been (parishes) formed the clusters that were randomized to inappropriately treated with anti-malarials [6]. The the intervention or control rather than individuals. Al- World Health Organization (WHO) and UNICEF have though the current study is not typically a randomized now recommended the provision of integrated manage- controlled trial, but was nested in an ongoing cluster- ment of common childhood illnesses at the community randomized trial, the consort checklist [17] has been used level [7]. Following this recommendation, Uganda to guide the report as much as possible. adopted the integrated community case management of childhood illnesses (ICCM) under which community health workers (CHWs) will provide malaria, pneumo- Study area nia, and diarrhoea treatment for children less than five The study was conducted in eastern Uganda in an area years [8]. It is hoped that the ICCM strategy will build covered by the Iganga-Mayuge Health and Demographic on the achievements of the HMM intervention by also Surveillance Site (HDSS). Iganga-Mayuge HDSS is pre- improving access to treatment for pneumonia and diar- dominantly rural with a few peri-urban areas mainly in rhoea as well as improving treatment outcomes. the trading centres and Iganga Municipality. It has a Preliminary studies on integrated malaria and pneu- total population of about 70,000 people and about monia treatment in Ghana, Zambia and Uganda have 11,000 of these are children under five years. The main assessed effects on mortality [9], treatment-seeking, drug occupation of the people is subsistence farming. Malaria use [10,11], timely treatment of pneumonia, and treat- is endemic in the area with transmission peaks occurring ment outcomes where malaria diagnosis is done with the in the rainy seasons in the months of March and aid of rapid diagnostic tests (RDTs) [12]. However, there September; and respiratory tract illnesses are common. is still need for more evidence to inform the implemen- The HDSS is served by one government hospital, tation of the ICCM strategy, including more data on nine public health centres, three non-governmental child health outcomes [13,14] in various contexts. organization hospitals and 122 drug shops and private Uganda has a complex public-private mix of health care clinics. There were also 131 CHWs (locally known as provided by government, private-not-for-profit (PNFP), community medicine distributors (CMDs)) who had and private-for-profit (PFP) facilities including private been providing health care since 2009 under the cluster clinics, drug shops and a few hospitals. In the study randomized trial on which the current study is based. Kalyango et al. Malaria Journal 2013, 12:340 Page 3 of 13 http://www.malariajournal.com/content/12/1/340 The intervention were trained first in the community management of The cluster-randomized trial involved treatment of chil- malaria and then in the integrated community manage- dren aged 4–59 months by CHWs. CHWs have provided ment of malaria and pneumonia. This training was treatment for malaria in Uganda since 2002 under the conducted by Ministry of Health officials together with home-based management of fever strategy. They initially the study investigators. Health workers in the public, used a combination of chloroquine and sulphadoxine- non-governmental organization and private facilities also pyrimethamine until the malaria policy change in 2004 received a two-day training in integrated community that led to use of artemisinin-based combination therapy management of malaria and pneumonia. They were ori- (ACT) as first-line treatment for malaria [18]. The im- ented on the algorithms that were to be used by the plementation of the policy was slow and use of ACT by CHWs. In addition, they were trained on investigating CHWs had not yet been widely rolled out when the and documenting adverse events, and supervision and ICCM strategy was adopted in 2010. Therefore, there training of CHWs. were no CHWs in the study area treating children with ACT at the time of commencement of the cluster- Patient management and follow-up randomized trial in 2009. The CHWs were selected as part The CHWs in the control arm treated children with of the cluster-randomized trial and trained either in the anti-malarials, but referred children with danger signs or management of malaria alone (which constituted the con- those with pneumonia symptoms, regardless of severity, trol arm) or integrated management of malaria and pneu- to nearby health facilities. The CHWs in the interven- monia (which constituted the intervention arm). The tion arm treated children with anti-malarials and/or an- selection of CHWs has been described elsewhere [19]. tibiotics as appropriate and referred children with danger signs to nearby health facilities. No pre-referral Randomization medicines were given to children that were referred with Randomization was done by a statistician that was inde- danger signs. The diagnoses of malaria and pneumonia pendent of the study using stratified block randomization. were made using symptoms according to the Integrated Iganga-Mayuge HDSS has 65 villages which make up 26 Management of Childhood Illness (IMCI) criteria [20]. parishes that were divided into eight urban and 18 rural The details of the treatment algorithm have been de- clusters (parishes). The clusters from the rural area were scribed elsewhere [11,19] but briefly, a diagnosis of “mal- further grouped into three strata based on the population aria” was made if the child had fever or the caregiver size of children less than five years: i) 190–320, ii) 321– reported history of fever in the previous 24 hours. The 390, and iii) 391 and above, resulting in six clusters in each CHWs did not use rapid diagnostic tests (RDTs) because of these strata. The clusters from the urban area were they commenced their roles under the cluster- grouped into two strata based on population sizes of iv) randomized trial in 2009 before WHO recommended 280–430, and v) 431 and above. Random numbers were parasite-based diagnosis for malaria in 2010 [21]. “Pneu- generated in blocks of six for the rural clusters and in monia” was diagnosed following the presence of cough blocks of four for the urban clusters. From each of the or difficult breathing and fast breathing (≥50 breaths per three strata in the rural area three clusters were random- minute in children aged four to 12 months and ≥40 ized to the intervention arm and three to the control arm. breaths per minute in children 12 to 59 months). In each of the two strata in the urban area, two clusters Breathing rates were counted using wrist watches with were randomized to the intervention arm and two clusters minutes and seconds hands. Children were classified as to the control arm. Health facilities in the setting are having severe disease if they presented with any of the equally distributed, therefore, geography and distance to four general danger signs: convulsions, repeated formal health care was not used in the randomization vomiting, lethargy/unconsciousness or failure to feed, or process. other danger signs: chest in-drawings, noisy breathing, severe dehydration or pallor. The CHWs were required Training to follow up the children they treated and refer those All CHWs received three days’ training in the manage- that did not get well to nearby health facilities. Although ment of malaria. This training was done before diarrhoea is one of the illnesses targeted by the ICCM randomization. After randomization, the CHWs in the strategy, the CHWs in this cluster-randomized trial did intervention areas received a further three days’ training not treat it. This was because the trial commenced be- in the integrated management of malaria and pneumo- fore the adoption of the ICCM strategy and was in- nia. In addition, the CHWs in both arms received formed mainly by studies that had shown necessity to monthly refresher training. The details of the CHW integrate malaria and pneumonia treatment due to training have been described elsewhere [19]. Prior to the symptom overlap. The study reported in this paper was training of CHWs, the district health team members conducted over a period of five weeks that started in Kalyango et al. Malaria Journal 2013, 12:340 Page 4 of 13 http://www.malariajournal.com/content/12/1/340 mid-October to November 2011, two years after CHWs included in the analyses for prompt and appropriate an- commenced their treatment roles under the cluster ran- tibiotics for pneumonia symptoms. domized trial. Sample size Study supplies The sample size was estimated using the formula for CHWs in the intervention arm were supplied with sample size for comparing two proportions with adjust- artemether-lumefantrine (AL) and amoxicillin while ments for clustering [24]. The assumptions used were: CHWs in the control arm were supplied with AL only. 90% power, 5% level of significance that was two-sided, The drugs were refilled at the monthly meetings of coefficient of variation of the proportions between clus- CHWs with the cluster-randomized trial staff and formal ters within each group of 0.4, proportions of children re- health workers. The formulation of AL used was pre- ceiving prompt treatment for pneumonia of 13% in the packaged dispersible Coartem® (Novartis Pharma AG). control arm and 68% in the intervention arm from a This was supplied in packs of six tablets (20 mg previous study [12], 20% of children having pneumonia artemether, 120 mg lumefantrine) for children aged symptoms, a minimum sample size of 12 children per 4–35 months or 12 tablets for children aged ≥ 36 months. cluster and a drop-out rate of 10%. The coefficient of Dispersible amoxicillin tablets (125 mg) (Medophar, variation was assumed from literature suggesting that India or IDA Foundation, Netherlands) were pre- values of coefficients of variation are usually ≤0.25 and packaged into three age-based doses: 6 tablets for chil- seldom exceed 0.5 [24], therefore 0.4 was chosen to fall dren aged 4–11 months (one tablet twice daily for three within this range of acceptable values. Based on these days), 12 tablets for children 12–35 months (two tablets assumptions about 111 children with pneumonia symp- twice daily for three days), and 18 tablets for children toms were needed in each arm, translating to about 610 36–59 months (three tablets twice daily for three days). children treated by CHWs in each arm. The dosages used in the cluster-randomized trial were chosen after some studies showed that three-day dosing Definition of pneumonia symptoms was equally as effective as five-day dosing [22,23]. The Pneumonia symptoms were defined as caregiver reports amoxicillin tablets were procured in bulk and pre- of cough accompanied by fast and/or difficult breathing packaged by a local pharmaceutical industry (Kampala similar to the definition used in MICS and DHS surveys Pharmaceutical Industries) into the age specific doses. [25]. This definition of pneumonia symptoms based on The drugs were procured from the manufacturers caregiver reports of symptoms was used instead of the through local pharmaceutical distributors and distrib- more specific IMCI definition of pneumonia that the uted through the district system. The CHWs in the CHWs use, because the CHWs in the control arm did intervention arm were supplied with watches that they not assess and classify pneumonia symptoms and there- used for respiratory rate counting. Neither the CHWs in fore had no record of which children had pneumonia the intervention arm nor the ones in the control arm symptoms. In addition, the children could not be classi- had thermometers. fied by the field team of the current study because they were identified after they had received treatment from Supervision the CHWs and the symptoms and signs were likely to CHWs in both arms received monthly support super- have changed by the time they were enrolled into the vision from health workers based at the nearest health study. The definition has been used elsewhere to define facility. Details of the supervision are reported elsewhere presumed pneumonia [26]. [19]. Study outcomes Study participants The primary outcome of the study was the proportion of All children aged 4–59 months in the intervention and children receiving prompt and appropriate antibiotics control areas that were treated by CHWs for any illness for pneumonia symptoms. Prompt treatment was de- were eligible for inclusion, and they were consecutively fined as receiving the first dose of treatment on the day enrolled into the study. The children were identified of presentation of symptoms or the next day. If the child from the CHWs’ registers and traced to their homes on received an appropriate antibiotic for pneumonia the day they sought care from the CHW (day 1) or the promptly then it was classified as prompt and appropri- next day for those that sought treatment late in the ate. The following medicines were considered appropri- evening or at night (this was still within the first twenty ate for pneumonia as recommended in the national or four hours of seeking care). All enrolled children were CHWs’ treatment guidelines or the British National included in the analysis for the treatment outcomes Formulary which is widely used in Uganda: amoxicillin, while only children with pneumonia symptoms were erythromycin, azithromycin, ampicillin, ampicillin plus Kalyango et al. Malaria Journal 2013, 12:340 Page 5 of 13 http://www.malariajournal.com/content/12/1/340 cloxacillin, gentamicin, benzyl penicillin which could be Data management and analysis switched to oral amoxicillin, procaine penicillin fortified The data were double entered into FoxPro and exported (PPF), which could be switched to amoxicillin, chloram- to STATA 10 (STATA Corp, College Station, TX, USA) phenicol, cotrimoxazole, cefuroxime, and amoxycillin for analysis. The baseline characteristics of the children plus clavulanate [27-29]. The assessment of appropriate and caregivers were summarized using descriptive statis- antibiotics did not take into account appropriate doses, tics. Proportions of the study outcomes (prompt and ap- frequency, and duration mainly because most of the chil- propriate antibiotics and treatment outcomes) were dren that received antibiotics from sources other than the compared in the intervention and control arms and CHWs would have just started these treatments by the crude relative risks were estimated using Mantel- day four evaluation. It would therefore not be possible to Haenszel methods that considered the strata used in the determine if they took the medicines appropriately and for randomization. In order to simultaneously account for the correct length of time. Promptness of treatment was the effect of the intervention and covariates on the pri- assessed based on data collected on day one while appro- mary outcome, individual level analysis was done using priateness of antibiotics was based on day four data. generalized estimating equations based on Poisson re- The secondary outcome was treatment outcome which gression with robust variance estimation and an ex- included: having temperature ≥ 37.5°C or high respiratory changeable correlation matrix that accounted for rate on day four; received anti-malarials from other health clustering [30]. The individual level analysis was done providers after CHW-treatment or antibiotics additional after establishing that the results obtained using individ- to those provided by the CHWs in the intervention arm ual level analysis did not differ much from those or additional to those provided by other health providers obtained using cluster level analysis since there were few to whom children with pneumonia symptoms in the con- clusters in each arm (less than the 15 recommended for trol arm were referred; hospitalization; or death [12]. appropriate use of individual level analysis). Analysis was Treatment outcomes were assessed on day four. Children done on intention-to-treat basis. In order to select co- that had any of the treatment outcomes were classified as variates for inclusion into the multivariate model, uni- having treatment failure. In addition, self-reported treat- variate analysis was done to determine associations ment failure, defined as caregiver report of non-resolution between the various factors on which data were col- of illness symptoms, was assessed on day four. lected and receiving prompt and appropriate antibiotics using generalized estimating equations with Poisson re- Data collection gression. The factors with p-values less than 0.2 at uni- The data were collected by a team of 30 experienced variate analysis were considered for inclusion into the data collectors that were independent of the cluster- multivariate model. These factors were assessed for stat- randomized trial. They were trained on the current istical interaction and confounding with the interven- study for one week. They were supervised by two med- tion. The differences in proportions of children with fast ical officers and the first author (JNK). The data were breathing between day one and four were compared in collected using a pre-tested questionnaire (divided into the intervention and control arms using z-tests for com- two parts) that was presented in both English and parison of proportions. Lusoga (the main local language spoken in the area) allowing the caregiver to choose between the two lan- Ethical considerations guages. Part one of the questionnaires was administered The study was approved by Makerere University School on day one since receiving treatment from the CHW of Public Health Higher Degrees Research and Ethics and collected data on: socio-demographic characteristics Committee (reference IRB00005876) and the Uganda of the caregivers and children, children’s presenting National Council of Science and Technology (reference symptoms, time between onset of symptoms and seeking HS 898). Permission to conduct the study in the area care, and treatment received from the CHW. Part two of was also obtained from the administration of Iganga- the questionnaire was administered on day four to col- Mayuge Health and Demographic Surveillance Site and lect data on: medicines obtained from other sources, the local administration of the villages where the chil- counselling reported by the caregiver, adherence to med- dren were enrolled from. The caregivers of the children icines, hospitalization during current illness episode, and were informed about the study and then written in- resolution of symptoms as reported by the caregiver. formed consent was obtained from them to have their Temperature and breathing rates of the children were children enrolled into the study. The caregivers were measured on day one and day four. Axillary tempera- given one of the two copies of informed consent docu- tures were measured using digital thermometers. The ments that were signed by them and the person breathing rates were counted by the field assistants with conducting the informed consent. Access to the data the aid of wrist watches. was restricted to the investigators. Kalyango et al. Malaria Journal 2013, 12:340 Page 6 of 13 http://www.malariajournal.com/content/12/1/340 Results caregivers in the control arm were living in rural areas A total of 1,276 children were enrolled into the study (90% control, 82% intervention). (667 from the control arm, 609 from the intervention arm) as shown in the study profile (Figure 1). The children Illness characteristics of the children were enrolled from all clusters of the cluster-randomized Almost all children in both arms (97%) presented with a trial but in varying numbers (17–138 children per cluster) history of fever. However, only 15% in the intervention depending on the number of children that were treated by and 14% in the control arm had temperature ≥ 37.5°C at the CHWs during the study period. Children were sam- the time of day one data collection. Slightly more than pled from 98 of the 131 CHWs (75%) in the area. 80% of the children in both arms had cough and about 29% in both arms reported difficult breathing. Fast Demographic characteristics of the children and breathing in children was reported by 30% of the care- caregivers givers in both arms. However, based on the respiratory The demographic characteristics of the children and rates counted by the field assistants, a higher proportion caregivers were comparable in the intervention and con- of children in the intervention arm (20%) had fast trol arms (Table 1). About half of the children in both breathing on the day one visit compared to the control arms were female with a mean age of about 30 months. arm (11%). The proportions of children with reported The mean age of the caregivers was about 31 years in diarrhoea, convulsions and vomiting were comparable in both arms and most of them were female, married, with the two arms (Table 2). primary level education. Slightly higher proportions of Figure 1 Study profile. Kalyango et al. Malaria Journal 2013, 12:340 Page 7 of 13 http://www.malariajournal.com/content/12/1/340 Table 1 Demographic characteristics of children treated prompt and appropriate antibiotics (RR = 3.47, 95%CI = by CHWs and their caregivers in Iganga-Mayuge 2.41-4.99) (Table 3). demographic surveillance site The association between the intervention (integrated Characteristic Intervention arm Control arm malaria and pneumonia management) and receiving N = 609 N = 667 prompt and appropriate antibiotics was significant also at adjusted analysis (RR = 3.51, 95% CI = 1.75-7.03). The Children’s characteristics adjusted analysis included factors that were statistically Female sex, n (%) 329 (54.0) 336 (50.4) significant or were retained as confounders, and these Mean age, months (SD) 30.4 (15.9) 31.5 (16.4) included: caregiver having post-primary education (RR = Caregivers’ characteristics 1.56, 95% CI = 1.04-2.36), child presenting with no diar- Female sex, n (%) 562 (92.3) 625 (93.7) rhoea (RR = 1.33, 95% CI = 1.01-1.76) and living in urban Age in years, mean (SD) 30.9 (9.4) 30.5 (8.7) residence (RR = 1.56, 95% CI = 1.20-2.00) (Table 4). Education level, n (%)* None 59 (9.7) 62 (10.4) Treatment outcomes Primary 381 (62.7) 450 (67.7) There was no difference in overall treatment failure Post primary 168 (27.6) 146 (22.0) among children in the intervention and control arms Religion, n (%)*# (RR = 0.88, CI = 0.67-1.18). However, children in the Catholic 30 (4.9) 65 (9.8) control arm were more likely to have temperatures above 37.5°C on day four compared to children in the Protestant 244 (40.1) 223 (33.5) intervention arm (4% versus 1%). In addition, there was Muslim 310 (51.0) 335 (50.4) a higher difference between proportions of children with Other 24 (4.0) 42 (6.3) fast breathing on day one and four in the intervention Marital status, n (%)** (9.2%) compared to control areas (4.2%, p = 0.01). There Married 559 (92.1) 592 (88.8) were no deaths reported. Perceived treatment failure was Single 23 (3.8) 27 (4.1) significantly lower in the intervention arm compared to the control arm (RR = 0.58, 95%CI = 0.46-0.73) (Table 5). Divorced 19 (3.1) 37 (5.6) Among children with pneumonia symptoms, children Widowed 6 (1.0) 11 (1.7) in the control arm were more likely to receive additional Rural residence, n (%) 499 (81.9) 600 (90.0) anti-malarials (3% control, 0% intervention). * missing data (2 control, 1 intervention). ** missing data (2 intervention). # significantly different. Additional medicines received by children Prompt and appropriate treatment for pneumonia The use of additional medicines was infrequent. Among symptoms the children that received additional medicines after re- In the intervention arm, 63% of the children with pneu- ceiving care from the CHW, paracetamol was the most monia symptoms received antibiotics while in the con- frequently used medicine in children from both the trol arm 51% were referred to health facilities. However, intervention (76%) and control (74%) arms followed by 45% of children with pneumonia symptoms in the inter- cotrimoxazole (14% intervention, 16% control) and vention arm and 11% in the control arm received cough syrups (10% intervention, 6% control) (Table 6). Table 2 Illness characteristics of 1276 children treated by CHWs in Iganga-Mayuge HDSS Characteristic Intervention Control RR (95% CI) History of fever, n (%) 589 (96.7) 647 (97.0) 1.00 (0.98-1.02) Temperature ≥ 37.5 C, n (%) 90 (14.8) 94 (14.1) 0.98 (0.71-1.34) Cough, n (%) 500 (82.1) 541 (81.1) 1.00 (0.94-1.06) Fast breathing by history, n (%) 182 (29.9) 199 (29.8) 0.99 (0.84-1.18) Fast breathing based on respiratory rate counted by field assistant, n (%) 122 (20.0) 76 (11.4) 1.44 (1.09-1.91) Difficult breathing by history, n (%) 174 (28.6) 194 (29.1) 0.94 (0.79-1.13) Diarrhoea by history, n (%) 193 (31.7) 218 (32.7) 0.90 (0.76-1.07) Convulsions by history, n (%) 40 (6.6) 40 (6.0) 0.95 (0.62-1.45) Vomiting by history, n (%) 232 (38.2) 262 (39.5) 0.89 (0.76-1.04) Kalyango et al. Malaria Journal 2013, 12:340 Page 8 of 13 http://www.malariajournal.com/content/12/1/340 Table 3 Prompt and appropriate antibiotics among children with pneumonia symptoms Outcome indicator Intervention Control RR (95% CI) n (%) n (%) CHW gave antibiotic/referred* 158 (63.0) 142 (50.9) 1.69 (1.23-2.32) CHW gave antibiotic/referred** 205 (81.7) 142 (50.9) 2.01 (1.49-2.71) Appropriate antibiotics from any source 166 (66.1) 69 (24.7) 2.41 (1.91-3.04) Prompt and appropriate antibiotics 114(45.4) 30 (10.8) 3.47 (2.41-4.99) * CHWs in intervention arm gave antibiotics and those in control arm referred to health facility. ** CHWs in intervention arm gave antibiotics or referred and those in control arm referred to health facility. Discussion This study has demonstrated that integrated commu- This study has demonstrated that integrated community nity case management of malaria and pneumonia in- case management of malaria and pneumonia increases creases prompt and appropriate antibiotics for self- prompt and appropriate treatment for self-reported reported pneumonia symptoms in a rural area with a pneumonia symptoms. In addition, children in the con- large number of private clinics and drug shops but trol arm were more likely to have a high temperature on which may be manned by unqualified health providers. day four of treatment seeking. There was a greater re- These findings are similar to those reported in a study in duction in fast breathing in children in the intervention Zambia where prompt and appropriate treatment was compared to the control arm. higher among children treated by CHWs that could treat Table 4 Factors associated with prompt and appropriate treatment among children with self-reported pneumonia symptoms in Iganga-Mayuge DSS Characteristic(N)+ Prompt & appropriate, n(%) Unadjusted RR (95%CI) P-value Adjusted RR* (95%CI) P-value Cluster arm† Intervention (251) 114(45.4) 3.60 (1.92-6.73) <0.001 3.51 (1.75-7.03) <0.001 Control (279) 30(10.8) 1.00 1.00 Primary caregiver Male (35) 15(42.9) 1.45(0.85-2.50) 0.18 Female (495) 129(26.1) 1.00 Education, caregiver† ≤Primary (415) 96 (23.1) 1.00 0.001 1.00 0.03 Post primary (114) 48 (42.1) 1.73 (1.25-2.40) 1.56 (1.04-2.36) No occupation/housewife/farmer (434) 97(22.4) 1.00 0.013 Other (96) 47(49.0) 1.85 (1.14-3.01) Child had diarrhoea† Yes (193) 44(22.8) 1.00 0.006 1.00 0.04 No (337) 100(29.7) 1.46 (1.12-1.90) 1.33 (1.01-1.76) Child had convulsions Yes (43) 10(23.3) 1.00 0.09 No (486) 133(27.4) 1.31 (0.96-1.78) Child was vomiting Yes (219) 69(31.5) 1.20 0.11 No (310) 75(24.2) 1.00 (0.96-1.52) Residence† Urban (79) 43(54.4) 2.08 (1.25-3.45) 0.005 1.56 (1.20-2.00) 0.001 Rural (451) 101(22.4) 1.00 1.00 +N represents the number in each category. * Adjusted analysis includes all variables with †. Kalyango et al. Malaria Journal 2013, 12:340 Page 9 of 13 http://www.malariajournal.com/content/12/1/340 Table 5 Treatment outcomes of children treated by CHWs in Iganga-Mayuge DSS Characteristic Intervention Control RR (95% CI) All children treated by CHWs N = 609 N = 667 Overall treatment failure 90 (14.8%) 101 (15.1%) 0.88 (0.67-1.18) Received additional* antibiotics 7 (1.2%) 13 (2.0%) 0.77 (0.32-1.86) Received additional** anti-malarials 4 (0.7%) 10 (1.5%) 0.44 (0.15-1.28) Hospitalization 10 (1.6%) 9 (1.4%) 1.47 (0.64-3.38) Temperature ≥ 37.5°C on day 4 7 (1.2%) 23 (3.5%) 0.29 (0.11-0.78) Had fast breathing on day 4 66 (10.8%) 48 (7.2%) 1.29 (0.87-1.92) Difference in fast breathing, day 1& 4† 9.2% 4.2% p-value = 0.01 Self reported treatment failure 83 (13.7%) 170 (25.5%) 0.58 (0.46-0.74) Children with malaria symptoms Characteristic Intervention Control RR (95% CI) n = 589 n = 647 Overall treatment failure in malaria 84 (14.3%) 99 (15.3%) 0.85 (0.64-1.13) Received additional* antibiotics 6 (1.0%) 13 (2.0%) 0.65 (0.26-1.65) Received additional** anti-malarials 3 (0.5%) 10 (1.6%) 0.35 (0.11-1.11) Hospitalization 9 (1.5%) 9 (1.4%) 1.36 (0.59-3.16) Temperature ≥ 37.5°C on day 4 7 (1.2%) 22 (3.4%) 0.30 (0.11-0.81) Had fast breathing on day 4 62 (10.5%) 48 (7.4%) 1.22 (0.82-1.82) Difference in fast breathing, day 1& 4,† 9.5% 4.0% p-value = 0.006 Self reported treatment failure 78 (13.3%) 168 (26.0%) 0.56 (0.44-0.71) Children with self reported pneumonia symptoms Characteristic Intervention Control RR (95% CI) n = 251 n = 279 Overall treatment failure in “pneumonia” 43 (17.1%) 56 (20.1%) 0.77 (0.52-1.13) Received additional* antibiotics 2 (0.8%) 10 (3.6%) 0.30 (0.08-1.21) Received additional** anti-malarials 0 (0) 8 (2.9%) - Hospitalization 4 (1.6%) 3 (1.1%) 2.42 (0.60-9.84) Temperature ≥ 37.5°C on day 4 5 (2.0%) 11 (3.9%) 0.41 (0.10-1.75) Had fast breathing on day 4 34 (13.6%) 25 (9.0%) 1.30 (0.76-2.21) Difference in fast breathing, day 1& 4 † 18.3% 3.6% p-value < 0.001 Self reported treatment failure 22 (8.8) 74 (26.5) 0.35 (0.22-0.56) * Additional antibiotics received after CHW treatment in intervention arm or after treatment from other source where children were referred in control arm. The CHWs in the control arm did not have antibiotics and children with pneumonia symptoms are expected to receive antibiotics from other health providers. ** Additional anti-malarials received after treatment by CHWs. † difference between percentage of children with fast breathing on day one and four, p-value comparing difference in intervention and control. both malaria and pneumonia [12]. This similarity is participation in the study at the time of seeking care. In found despite the differences in study designs and con- addition, the current study was done in a setting with texts of the Zambian and the current study. In the several drug shops and clinics which are likely to be add- Zambian study, RDTs were employed for malaria diag- itional sources of antibiotics. As expected in such a set- nosis which was likely to have improved the accuracy ting, the difference between the intervention and control of the illness classification. The Zambian study was arms was lower compared to that in the Zambian study also in more controlled conditions where children were though not significantly. recruited as they sought care from the CHWs compared The proportion of children that received prompt and to the current study which is better classified as an ef- appropriate antibiotics was higher among children fectiveness study because children were recruited after whose caregivers had received post primary education. they sought care and were therefore not aware of their This may be due to better recognition of symptoms Kalyango et al. Malaria Journal 2013, 12:340 Page 10 of 13 http://www.malariajournal.com/content/12/1/340 Table 6 Additional medicines received by children after caregivers with primary or no education [32] that were treatment by CHW mostly living in the rural areas. Caregivers in urban Medicine Intervention Control areas have better knowledge of childhood illness and also have better health-seeking behaviour [33]. Dispar- (n = 42) (n = 96) ities in access to health care have been observed between Paracetamol 32 (76.2) 71 (74.0) rural and urban areas [34]. Cotrimoxazole* 6 (14.3) 15 (15.6) There was no difference in overall treatment failure Cough syrups 4 (9.5) 6 (6.3) among all children treated by CHWs in the intervention Amoxicillin* 1 (2.4) 5 (5.2) and control areas, similar to a study in Zambia [12]. Chlorphenramine 6 (14.3) 5 (5.2) However, contrary to the Zambian study where the pro- Metronidazole 2 (4.8) 4 (4.2) portion of children with persistent fever, fast or difficult breathing at follow up did not differ, the current study Quinine† 0 (0) 3 (3.0) found a higher proportion of febrile children in the con- Ampicillin* 0 (0) 2 (2.0) trol arm on day four despite both groups having similar Artemether-lumefantrine† 3 (7.1) 2 (2.0) proportions of children with high temperature on day Benzylpenicillin* 0 (0) 2 (2.0) one. In addition, although the children in the interven- Chloramphenicol* 0 (0) 2 (2.0) tion arm had a higher proportion of children with fast Chloroquine† 1 (2.4) 2 (2.0) breathing on day one and day four (although not signifi- cant on day four), there was a greater reduction in the Combination flue medicines 0 (0) 2 (2.0) proportions of children with fast breathing in the inter- Dexamethasone 1 (2.4) 2 (2.0) vention arm. The differences seen in the two studies Unknown antimalarial syrup† 0 (0) 2 (2.0) could have been a result of the way the outcomes were Ampiclox* 0 (0) 1 (1.0) presented. In the Zambian study persistent fever and fast Anti-emetic 1 (2.4) 1 (1.0) or difficult breathing were grouped together which could Cetrizine 0 (0) 1 (1.0) have masked outcome-specific differences. The initial higher fast breathing in the intervention arm could have Erythromycin* 0 (0) 1 (1.0) resulted from more caregivers in the control arm whose Ibuprofen 1 (2.4) 1 (1.0) children had pneumonia symptoms bypassing the CHWs Mebendazole 0 (0) 1 (1.0) and taking their children directly to other health pro- ORS 1 (2.4) 1 (1.0) viders. Similar differences in proportions of children Vitamins 1 (2.4) 0 (0) with respiratory symptoms were also noted in the study * considered as additional antibiotics. in Zambia. These findings suggest improved treatment † considered as additional anti-malarials. outcomes when children can access effective treatments promptly. Many of the children in the control arm that got antibiotics received cotrimoxazole against which among the more educated caregivers or better afford- high levels of resistance of pneumonia-causing bacteria ability of treatment resulting in more prompt care seek- have been reported [35,36]. This finding supports the ing. The study in Zambia found a tendency to more need to have the effective medicines for common child- prompt and appropriate treatment among caregivers hood illnesses at the first level of care because the chil- with primary or secondary education compared to those dren need them. This may prevent the delays that did not have any education although it was not sta- experienced in receiving necessary treatment. tistically significant [12]. Differences in receiving prompt There was no difference in overall treatment failure and appropriate antibiotics may point to inequalities that among children with pneumonia symptoms contrary to may exist in accessing health care in persons of different the Zambian study [12] where children classified as hav- social standing. Low education has been associated with ing pneumonia in the intervention arm were less likely low income and low knowledge on aspects that affect to have treatment failure compared to their counterparts use of health services [31]. in the control arm. The difference in findings could be a The children in urban areas were more likely to re- result of how pneumonia was classified in the two stud- ceive prompt and appropriate antibiotics. This may be ies. The current study used caregiver reports of pneumo- due to easier access to treatment in the urban areas. It nia symptoms whereas in the Zambian study the may also be related to caregivers’ knowledge of disease classification of pneumonia was based on the CHWs symptoms. There were higher proportions of caregivers classification; which may have led to lower misclassifica- with post primary education in the urban areas and tion of children as having pneumonia when they did not these may have better health knowledge than the [37]. This would make it easier to detect differences in Kalyango et al. Malaria Journal 2013, 12:340 Page 11 of 13 http://www.malariajournal.com/content/12/1/340 treatment failure in children treated by CHWs who the intervention and control arms compared to the re- treated and those that did not treat pneumonia. How- sults obtained using self reported pneumonia symptoms. ever, reduction in fast breathing (based on respiratory This suggests that the use of self reported pneumonia rates measured by the data collectors) between day one symptoms in the current study could have led to under- and four among children with pneumonia symptoms in rather than over-estimation of the effect of integrated the intervention arm of the current study (18%) com- malaria and pneumonia management because self- pared to reduction in fast breathing in the control arm reported pneumonia symptoms may be more sensitive (3.6%) is even higher than that seen among all children but less specific than the respiratory rate assessed by treated by CHWs (9.2% versus 4.2% respectively). data collectors on day one. While this study assessed only symptoms and treat- The analysis presented was done at the individual level ment of malaria and pneumonia, current WHO recom- whereas it would have been better to do cluster level mendations are to integrate also diarrhoea treatment analysis since there were fewer than 15 clusters in each into integrated community case management. The find- arm. However with analysis at the cluster level it would ings from this study support this recommendation to in- have been difficult to simultaneously control for other corporate ORS and zinc since a considerable proportion variables [30]. The results obtained from analysis at the of children had diarrhoea (about 30%) but only two chil- cluster and individual levels at unadjusted analysis were dren received ORS and likely no child zinc. compared and found similar. Adjusted analysis based on residuals for the primary outcome was done and found Methodological issues that it was still significant. The analysis done at the indi- Self-reported symptoms were used to classify children vidual level with adjustments for the cluster into whether they had possible pneumonia and thus randomization to show effects of other variables has, children may have been misclassified. The IMCI guide- therefore, been presented. In addition, there were imbal- lines, which CHWs use to make classifications, define ances in the number of children treated in the clusters pneumonia as cough or difficult breathing and fast which could have resulted in loss of power. However, breathing [20]. The classification of fast breathing is the effect may have been minimal since a difference in based on age-specific thresholds of breath counts which the main outcome variable has been detected between could not be ascertained based on the caregiver reports. the intervention and control areas. Nevertheless, other It was not possible to take the baseline assessments of associations may have been undetectable. respiratory rates in the children because they were lo- There were no buffer zones between the intervention cated on day one of treatment-seeking and for some of and control areas which may have led to dilution of the the children the illness may have changed by the time intervention effects since some caregivers in the control they were seen. In addition, the CHWs’ records could areas could have accessed antibiotics from the interven- not be used for classification of pneumonia because in tion areas. However, a three-fold difference between the the control arm, the CHWs do not assess and classify intervention and control areas is still demonstrated on pneumonia symptoms. They had no record of which receiving prompt and appropriate antibiotics. children had presented with pneumonia symptoms. Use Treatment failure may have been underestimated be- of caregivers’ reports to classify pneumonia symptoms cause danger signs of illness were not assessed on day has been used by Multiple Indicator Cluster Surveys four. This could have underestimated treatment effects (MICS) and Demographic and Health surveys (DHS) as in the integrated care arm. well as other studies to identify children with possible Another limitation of this study is that it did not in- pneumonia in the community [25,26], but it is likely to clude all children treated by the CHWs in the cluster- over-estimate the proportion of children with true pneu- randomized trial but only assessed those that were treated over a period of five weeks. There may have been monia. Therefore, the results of appropriate treatment obtained may reflect over-treatment with antibiotics variations in outcomes of children treated over the [37]. The desirable treatment rate may therefore not be whole year. 100%, until a more specific indicator has been found [37]. However, this measure though unspecific, does not Conclusion introduce differential bias between the study arms, thus Implementation of integrated community case manage- allowing for relative comparisons. A further classifica- ment is an effective approach in improving access to tion of pneumonia was done based on the respiratory prompt and appropriate antibiotics for pneumonia rate assessments of children on day one by the field as- symptoms and may improve treatment outcomes. This sistants. The results showed an even bigger difference in may reduce mortality in children less than five years the proportion of children that received prompt and ap- since pneumonia is one of the major killers of children propriate treatment for pneumonia symptoms between globally. Kalyango et al. Malaria Journal 2013, 12:340 Page 12 of 13 http://www.malariajournal.com/content/12/1/340 Abbreviations 5. Källander K, Nsungwa-Sabiiti J, Peterson S: Symptom overlap for malaria ACT: Artemisinin-based combination therapy; AL: Artemether-lumefantrine; and pneumonia—policy implications for home management strategies. CHWs: Community health workers; CMDs: Community medicine distributors; Acta Trop 2004, 90:211–214. HDSS: Health and demographic surveillance site; HMM: Home management 6. Källander K, Hildenwall H, Waiswa P, Galiwango E, Peterson S, Pariyo G: of malaria; ICCM: Integrated Community Case Management of Childhood Delayed care seeking for fatal pneumonia in children aged under five years Illnesses; IMCI: Integrated Management of Childhood Illness; NGO: Non in Uganda: a case-series study. Bull World Health Organ 2008, 86:332–338. Governmental Organization; ORS: Oral rehydration salts; PFP: Private for Profit; 7. Pneumonia: The forgotten killer of children. http://www.unicef.org/mdg/ PNFP: Private not for Profit; PPF: Procaine Penicillin Fortified; RDTs: Rapid mortalitymultimedia/Pneumonia_The_Forgotten_Killer_of_Children.pdf diagnostic tests; RR: Relative risk; SD: Standard deviation; UNICEF: United (Accessed November 18, 2012). Nations Children’s Fund; WHO: World Health Organization. 8. WHO: Child and Adolescent Health Program: Quarterly Report. Geneva: World Health Organization; 2011. Competing interests 9. Chinbuah MA, Kager PA, Abbey M, Gyapong M, Awini E, Nonvignon J, The authors declare that they have no competing interests. Adjuik M, Aikins M, Pagnoni F, Gyapong JO: Impact of community management of fever (using antimalarials with or without antibiotics) on childhood mortality: a cluster-randomized controlled trial in Ghana. Am J Authors’ contributions Trop Med Hyg 2012, 87:11–20. JNK, TA, SP, CK, and ER took part in: designing the study, development of 10. Seidenberg PD, Hamer DH, Iyer H, Pilingana P, Siazeele K, Hamainza B, study tools, data analysis and manuscript writing. KM took part in MacLeod WB, Yeboah-Antwi K: Impact of integrated community case development of study tools. JNK and KM participated in data collection. All management on health-seeking behavior in rural Zambia. Am J Trop Med authors read and approved the final manuscript. Hyg 2012, 87:105–110. 11. Kalyango JN, Lindstrand A, Rutebemberwa E, Ssali S, Kadobera D, Karamagi Acknowledgements C, Peterson S, Alfven T: Increased use of community medicine distributors We are very grateful to the CHWs and Iganga-Mayuge Health and and rational use of drugs in children less than five years of age in Demographic Surveillance Site administration and staff, most notably Mr Uganda caused by integrated community case management of fever. Edward Galiwango, Mrs Judith Kaija, Mr Michael Kibuuka, Mr Daniel Am J Trop Med Hyg 2012, 87:36–45. Kadobera, Mr Zach Ojiek, and the field team who made this study possible. 12. Yeboah-Antwi K, Pilingana P, Macleod WB, Semrau K, Siazeele K, Kalesha P, We also thank Dr Sheila Tusubira and Amanda Wanyana for their Hamainza B, Seidenberg P, Mazimba A, Sabin L, Kamholz K, Thea DM, contribution to data collection; and the data entrants for their hard work on Hamer DH: Community case management of fever due to malaria and this research. We thank the anonymous reviewer for the valuable comments pneumonia in children under five in Zambia: A cluster randomized and suggestions. controlled trial. PLoS Med 2010, 7:e1000340. 13. Hamer DH, Marsh DR, Peterson S, Pagnoni F: Integrated community case Financial support management: next steps in addressing the implementation research This study received financial support from Swedish Institute for Development agenda. Am J Trop Med Hyg 2012, 87:151–153. Agency (SIDA) and UNICEF/UNDP/World Bank/WHO Special Program for 14. Marsh DR, Hamer DH, Pagnoni F, Peterson S: Evidence for the Research and Training in Tropical Diseases. implementation, effects, and impact of the integrated community case management strategy to treat childhood infection. Am J Trop Med Hyg Disclaimer 2012, 87:2–5. The findings and conclusions in this article are those of the authors and do 15. Uganda national communication strategy for promoting rational use of not necessarily represent the views of SIDA or UNICEF/UNDP/World Bank/ medicines. http://apps.who.int/medicinedocs/documents/s16511e/s16511e. WHO Special Programme for Research and Training in Tropical Diseases. pdf (Accessed April20, 2013). 16. Awor P, Wamani H, Bwire G, Jagoe G, Peterson S: Private sector drug shops in Author details integrated community case management of malaria, pneumonia, and Department of Public Health Sciences, Global Health (IHCAR), Karolinska diarrhea in children in Uganda. Am J Trop Med Hyg 2012, 87:92–96. Institutet, SE 17177, Stockholm, Sweden. Clinical Epidemiology Unit, 17. Campbell MK, Piaggio G, Elbourne DR, Altman DG: Consort 2010 Makerere University College of Health Sciences, P.O. Box 7072, Kampala, statement: extension to cluster randomised trials. BMJ 2012, 345:e5661. Uganda. Department of Pharmacy, Makerere University College of Health 18. Malaria case management: home based management of fever. http:// Sciences, P.O. Box 7072, Kampala, Uganda. Department of Paediatrics, Sach’s www.health.go.ug/mcp/mt.html (Accessed September 10. 2010). Children’s Hospital, Södersjukhuset, Stockholm, Sweden. International 19. Kalyango JN, Rutebemberwa E, Alfven T, Ssali S, Peterson S, Karamagi C: Maternal and Child Health, Department of Women and Children’s Health, Performance of community health workers under integrated community Uppsala University, Uppsala, Sweden. African Field Epidemiology Network case management of childhood illnesses in eastern Uganda. Malar J (AFENET), P.O Box 12874, Kampala, Uganda. Department of Paediatrics and 2012, 11:282. Child Health, Makerere University College of Health Sciences, P.O. Box 7072, 20. Integrated Management of Childhood Illness chart booklet. http://whqlibdoc. Kampala, Uganda. Department of Health Policy, Planning and Management, who.int/publications/2008/9789241597289_eng.pdf (Accessed November School of Public Health, Makerere University College of Health Sciences, P.O. 20, 2011). Box 7072, Kampala, Uganda. 21. Guidelines for the treatment of malaria. http://whqlibdoc.who.int/ publications/2010/9789241547925_eng.pdf (Accessed September 15, 2012). Received: 6 June 2013 Accepted: 18 September 2013 22. ISCAP study group: Three day versus five day treatment with amoxicillin Published: 22 September 2013 for non-severe pneumonia in young children: a multicentre randomised controlled trial. BMJ 2004, 328:791. References 23. Pakistan Multicentre Amoxycillin Short Course Therapy (MASCOT) 1. UNICEF: In Commiting to child survival: A promise renewed. Edited by pneumonia study group: Clinical efficacy of 3 days versus 5 days of oral Anthony D, Mullerbeck E. New York: UNICEF; 2012:44. amoxicillin for treatment of childhood pneumonia: a multicentre 2. WHO: Health related Millennium Development Goals. Geneva: World Health double-blind trial. Lancet 2002, 360:835–841. Organization; 2011. 24. Hayes RJ, Bennett S: Simple sample size calculation for cluster- 3. CORE Group, Save the Children, BASICS and MCHIP: Community case randomized trials. Int J Epidemiol 1999, 28:319–326. management essentials: Treating common childhood illnesses in the 25. Pneumonia. www.childinfo.org/pneumonia.html (Accessed September 5, 2013). community. A guide for program managers. Washington DC: CORE group; 2010. 26. Diaz T, George A, Rao S, Bangura P, Baimba J, McMahon S, Kabano A: 4. Staedke SG, Mwebaza N, Kamya MR, Clark TD, Dorsey G, Rosenthal PJ, Healthcare seeking for diarrhoea, malaria and pneumonia among Whitty CJM: Home management of malaria with artemether-lumefantrine children in four poor rural districts in Sierra Leone in the context of free compared with standard care in urban Ugandan children: a randomised health care: results of a cross-sectional survey. BMC Publ Health 2013, controlled trial. Lancet 2009, 373:1623–1631. 13:157. Kalyango et al. Malaria Journal 2013, 12:340 Page 13 of 13 http://www.malariajournal.com/content/12/1/340 27. Department of Health Policy Planning and Management: Community Medicine Distributor training guide on Integrated Management of Malaria and Pneumonia. Kampala: School of Public Health Makerere University; 2010:24. 28. Joint Formulary Committee: British National Formulary. 53rd edition. London: British Medical Association and Royal Pharmaceutical Society of Great Britain; 2007. 29. Ministry of Health (Uganda): Uganda Clinical Guidelines. 2nd edition. Kampala: Ministry of Health, Uganda; 2010. 30. Hayes RJ, Moulton LH: Cluster randomized trials. 1st edition. Boca Raton (Florida): CRC Press; 2009. 31. Victora CG, Wagstaff A, Schellenberg JA, Gwatkin D, Claeson M, Habicht J-P: Applying an equity lens to child health and mortality: more of the same is not enough. Lancet 2003, 362:233–241. 32. Njama D, Dorsey G, Guwatudde D, Kigonya K, Greenhouse B, Musisi S, Kamya MR: Urban malaria: primary caregivers’ knowledge, attitudes, practices and predictors of malaria incidence in a cohort of Ugandan children. Trop Med Int Health 2003, 8:685–692. 33. Uzochukwu BSC, Onwujekwe EO, Onoka CA, Ughasoro MD: Rural–urban differences in maternal responses to childhood fever in south East Nigeria. PLoS One 2008, 3:e1788. 34. Kahabuka C, Kvale G, Hinderaker SG: Care-seeking and management of common childhood illnesses in Tanzania - Results from the 2010 demographic and health survey. PLoS One 2013, 8:e58789. 35. Hoa N, Trung N, Larsson M, Eriksson B, Phuc H, Chuc N, Lundborg C: Decreased Streptococcus pneumoniae susceptibility to oral antibiotics among children in rural Vietnam: a community study. BMC Infect Dis 2010, 10:85. 36. Joloba ML, Bajaksouzian S, Palavecino E, Whalen C, Jacobs MR: High prevalence of carriage of antibiotic-resistant Streptococcus pneumoniae in children in Kampala Uganda. Int J Antimicrob Agents 2001, 17:395–400. 37. Campbell H, el Arifeen S, Hazir T, O’Kelly J, Bryce J, Rudan I, Qazi SA: Measuring coverage in MNCH: challenges in monitoring the proportion of young children with pneumonia who receive antibiotic treatment. PLoS Med 2013, 10:e1001421. doi:10.1186/1475-2875-12-340 Cite this article as: Kalyango et al.: Integrated community case management of malaria and pneumonia increases prompt and appropriate treatment for pneumonia symptoms in children under five years in Eastern Uganda. Malaria Journal 2013 12:340. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit

Journal

Malaria JournalSpringer Journals

Published: Sep 22, 2013

There are no references for this article.