Quality of care for under-fives in first-level health facilities in one district of Bangladesh

Transcription

1 Qulity of cre for under-fives in first-level helth fcilities in one district of Bngldesh S.E. Arifeen, 1 J. Bryce, 2 E. Gouws, 3 A.H. Bqui, 4 R.E. Blck, 4 D.M.E. Hoque, 1 E.K. Chowdhury, 1 M. Yunus, 1 N. Begum, 1 T. Akter, 1 & A. Siddique 1 Objectives The multi-country evlution of Integrted Mngement of Childhood Illness (IMCI) effectiveness, cost nd impct (MCE) is globl evlution to determine the impct of IMCI on helth outcomes nd its cost-effectiveness. MCE studies re under wy in Bngldesh, Brzil, Peru, Ugnd nd the United Republic of Tnzni. The objective of this nlysis from the Bngldesh MCE study ws to describe the qulity of cre delivered to sick children under 5 yers old in first-level government helth fcilities, to inform government plnning of child helth progrmmes. Methods Generic MCE Helth Fcility Survey tools were dpted, trnslted nd pre-tested. Medicl doctors trined in IMCI nd these tools conducted the survey in ll 19 helth fcilities in the study res. The dt were collected using observtions, exit interviews, inventories nd interviews with fcility providers. Findings Few of the sick children seeking cre t these fcilities were fully ssessed or correctly treted, nd lmost none of their cregivers were dvised on how to continue the cre of the child t home. Over one-third of the sick children whose cre ws observed were mnged by lower-level workers who were significntly more likely thn higher-level workers to clssify the sick child correctly nd to provide correct informtion on home cre to the cregiver. Conclusion These results demonstrte n urgent need for interventions to improve the qulity of cre provided for sick children in first-level fcilities in Bngldesh, nd suggest tht including lower-level workers s trgets for IMCI cse-mngement trining my be beneficil. The findings suggest tht the IMCI strtegy offers promising set of interventions to ddress the child helth service problems in Bngldesh. Keywords Child helth services/stndrds; Primry helth cre; Qulity of helth cre; Qulity indictors, Helth cre; Helth cre surveys; Child, Preschool; Bngldesh (source: MeSH, NLM). Mots clés Service snté infntile/normes; Soins snté primire; Qulité soins; Indicteurs qulité snté; Enquête système de snté; Enfnt âge pré-scolire; Bengldesh (source: MeSH, INSERM). Plbrs clve Servicios de slud infntil/norms; Atención primri de slud; Clidd de l tención de slud; Indicdores de clidd de l tención de slud; Encuests de tención de l slud; Preescolr; Bngldesh (fuente: DeCS, BIREME). Bulletin of the World Helth Orgniztion 2005;83: Voir pge 266 le résumé en frnçis. En l págin 266 figur un resumen en espñol..267 Introduction Ech yer over 10 million children in low- nd middle-income countries die before their fifth birthdy (1, 2). About hlf of these deths re due to pneumoni, dirrhoe, mlri nd mesles (1). Undernutrition contributes to bout 50% of these deths (1, 3). Integrted Mngement of Childhood Illness (IMCI) is strtegy developed by WHO, the United Ntions Children s Fund (UNICEF) nd other technicl prtners to reduce mortlity due to these cuses nd to undernutrition (4). By December 2002, IMCI hd been introduced in 109 developing countries (5). The IMCI strtegy includes guidelines for the mngement of sick children t first-level fcilities. The guidelines re intended to improve cre by ensuring complete ssessment of the child s helth, nd by providing lgorithms tht combine presenting symptoms into set of illness clssifictions for mngement. The generic guidelines were vlidted nd found to produce outcomes similr to those from expert peditricins (4). The multi-country evlution of IMCI effectiveness, cost nd impct (MCE) is globl evlution to determine the impct of IMCI on child mortlity, helth nd nutrition. MCE studies re under wy in five countries (6, 7). Findings from other MCE sites hve demonstrted tht the introduction 1 Interntionl Centre for Dirrhoel Diseses Reserch, Bngldesh (ICDDR,B), PO Box 128, Dhk, Bngldesh. Correspondence should be sent to Dr Arifeen t this ddress (emil: shms@icddrb.org). 2 Ithc, New York, USA. 3 Deprtment of Child nd Adolescent Helth, World Helth Orgniztion, Genev, Switzerlnd. 4 Johns Hopkins Bloomberg School of Public Helth, Bltimore, MD, USA. Ref. No (Submitted: 22 July 2004 Finl revised version received: 4 Jnury 2005 Accepted: 5 Jnury 2005) 260 Bulletin of the World Helth Orgniztion April 2005, 83 (4)

2 S.E. Arifeen et l. of IMCI in helth fcilities cn improve the qulity of helth cre provided for children (8 10). Only in the Bngldesh site, however, ws n ssessment of the qulity of cre crried out prior to introducing IMCI. This study is prticulrly importnt in the light of severl recent reports tht hve highlighted limited country-specific plnning for implementtion s mjor reson why most countries hve not successfully scled up IMCI to rech coverge levels tht would hve n impct t the popultion level (11, 12). The MCE evlution in Bngldesh is 7-yer study tht begn in The study is being implemented in Mtlb thn, rurl subdistrict in southern Bngldesh. Mtlb hs popultion of bout nd mortlity rte in children ged under 5 yers of pproximtely 89 per 1000 live births. About 35% of these deths cn be ttributed to cuses directly ddressed by IMCI, nmely pneumoni, dirrhoe, mesles nd mlnutrition (13). The IMCI evlution is tking plce in the four-fifths of the thn where helth services re run by the Government of Bngldesh. First-level government fcilities in rurl Bngldesh, nd in the study re, re usully stffed by prmedic (medicl ssistnt/sub-ssistnt community medicl officer (MA/SACMO)) usully mle who hs hd 4 yers of clinicl trining, nd femle reproductive helth worker (fmily welfre visitor (FWV)) who hs hd 18 months of trining in provision of mternl nd child helth nd fmily plnning services. In some fcilities there is position for doctor, but in most cses, these positions remin vcnt (14). There re no officil user-fees for government-provided child helth services. Bngldesh lso hs wide rnge of locl prctitioners of indigenous nd western medicine, nd drug stores. Becuse IMCI hd not yet been fully implemented in Bngldesh, Mtlb thn provided n excellent opportunity for probbility-design ssessment of IMCI impct. The smpling frme included ll 20 first-level government helth fcilities in the thn outside the Interntionl Centre for Dirrhoel Diseses Reserch, Bngldesh (ICDDR,B) intervention re, hlf of which were rndomly selected for the implementtion of IMCI. The dt sources for the evlution included bseline household nd helth fcility surveys. Mid-term nd finl surveys will be conducted fter IMCI hs been fully implemented (6). The objective of this nlysis from the Bngldesh MCE study ws to describe the qulity of cre delivered to sick children Reserch ged under 5 yers in the 20 first-level government helth fcilities to inform government plnning nd lter evlutions of improvements in child helth services nd outcomes tht my be ssocited with the introduction of IMCI. Methods The bseline survey of helth fcilities ws crried out between August nd September The trget ws to observe 15 sick children t ech fcility (to detect difference of 20 percentge points between IMCI nd comprison fcilities for priority MCE indictors; lph = 5%; power = 80%; one-tiled design; design effect = 2.0). The generic helth-fcility survey tool ws dpted for this survey (7, 15). Informtion ws collected through observtion of cse-mngement using stndrd checklist, exit interviews with cregivers, re-exmintion of ech child by gold-stndrd surveyor, interviews with helth-cre providers nd n udit of the supplies nd equipment vilble t the helth fcilities. Indictors The performnce of helth-cre providers ws ssessed using the qulity of cre indictors designed for the MCE. IMCIspecific indictors hve been indicted in the Tbles by footnote s their inclusion in the ssessment would otherwise represent negtive bis ginst the qulity of cre ssessment of the fcilities (becuse the cre my be cliniclly correct but not conform to IMCI guidelines). Indictor definitions cover ssessment tsks, correct clssifiction, correct tretment nd correct counselling of cregivers (7). A composite indictor ws used to summrize the performnce of 10 ssessment tsks (16). Helth fcility support indictors including vilbility of essentil orl tretments (excluding ntimlrils) nd vilbility of essentil equipment nd mterils were lso ssessed. The stisfction of cregivers with fcility services ws ssessed, but is not reported here. Procedures Two field tems, ech consisting of three physicins, conducted the survey. They hd been trined in IMCI cse-mngement nd study procedures. The exit interview form ws trnslted into Bengli nd verified by the principl investigtor. UNISCALE weighing scles, mesuring tpes, thermometers nd timers were used in the survey. Tble 1. Helth fcilities ssessed by type nd cdre of helth provider present during survey observtions in Mtlb thn, Bngldesh, 2000 No. of fcilities with: Fcility type Totl number Medicl ssistnt/ Fmily welfre Phrmcist of fcilities sub-ssistnt visitor community medicl officer Type 1: Union helth nd fmily welfre centre Type 2: Union subcentre Type 3: Co-locted union helth nd fmily welfre centre nd union subcentre Totl In three instnces, the union helth nd fmily welfre centre operted only s stellite clinic. Bulletin of the World Helth Orgniztion April 2005, 83 (4) 261

3 Reserch One tem spent three consecutive dys t ech fcility, rriving before consulttions begn in the morning. Ech dy, the tems ttempted to enrol the first five sick children ged 2 59 months presenting for cre. One surveyor ccompnied the child nd cregiver throughout the cse-mngement process, to record wht hppened without interfering with routine services. When the consulttion ws complete, second surveyor, blinded s to the originl ssessment, conducted gold stndrd re-exmintion of the sick child to provide the IMCI-specific clssifiction of illness used in this nlysis. An exit interview ws conducted with the child s cregiver t this time. Menwhile, the first surveyor observed the csemngement process for the next child in the witing room. It ws not lwys possible to observe ll children, becuse some rrived while dt were still being collected on the previous child. A complete list of ll sick children seeking cre t the fcility on the survey dys ws mde. Interviews with helth providers nd udits of fcility supplies nd equipment were crried out during slck periods. At the end of ech dy the dt-collection tem nd the tem leder reviewed ll forms for consistency nd completeness. Ethicl pprovl for the study ws obtined from the ethicl review committees of ICDDR,B, nd WHO in Genev, Switzerlnd. Informed verbl consent ws obtined from the cregivers of the sick children who presented t the helthcre fcilities. Dt processing nd nlysis Double dt entry ws crried out by two different people using EPIInfo 6.0 (17); the two versions were compred nd discrepncies resolved by referring to the originl forms. Stndrd rnge nd consistency checks were mde. Sttisticl nlysis ws crried out using STATA version 7 for Windows (18). For cse-mngement indictors, the nlysis ws weighted so tht the smple of children observed ws representtive of typicl dy s ttendnce t the helth fcility. The weighting used ws the inverse of the totl number of children ctully included in the survey divided by the totl number of sick children who ttended the fcility on the three dys of observtion. For comprison between cdres of helth-cre providers the chi-squre test ws used with Ro Scott correction to djust for clustering of children t the helth fcility level. Results The bseline helth fcility survey ws crried out in 19 of the 20 first-level fcilities in the study re. One fcility could not be included in the survey becuse no stff member ws present during the survey period. Tble 1 provides detils of fcility types nd stffing ptterns nd Tble 2 presents the ge nd sex distribution of the 284 sick children seen t the fcilities during the study period. The clssifiction of illness most frequently presenting ws fever (80%), followed by cough or cold (49%), pneumoni (25%) nd dirrhoe (19%) (Tble 3). No child presented with dirrhoe with dehydrtion or with mlri. Thirty-three per cent of children presented with other illnesses not included in IMCI clssifictions, lthough ll but one of these children (94/95) lso presented with n IMCI-relted illness. The most common non-imci symptom ws skin infection (46/95). Eighty-seven per cent of ll children included in the survey received more thn one clssifiction; 22% received four or more clssifictions. Both cough nd fever were present, with or without S.E. Arifeen et l. Tble 2. Chrcteristics of children observed nd cdre of helth cre provider in Mtlb thn, Bngldesh, 2000 No. of children observed n = 284 (percentge) Age in months (30) (26) (18) (16) (10) Sex Mle 158 (56) Femle 126 (44) Cregiver Mother 242 (88) Other 33 (12) Cdre of provider Medicl ssistnt 184 (65) Fmily welfre Visitor 94 (33) Phrmcist 6 (2) Nine missing (no cregiver interview). Tble 3. Presenting conditions of 284 sick children ged under 5 yers ttending first-level helth fcilities in Mtlb thn, Bngldesh, 2000 Clssifiction of n Percentge presenting illness Fever Cough or cold Pneumoni Dirrhoe Very low weight Anemi Acute er infection 25 9 Chronic er infection 21 7 Dysentery 15 5 Mesles 8 4 Other Percentges re weighted estimtes wheres the numbers re ctul, nd therefore they will not exctly correspond. other symptoms, in 40% of the children (17% hd both cough nd fever with no other symptoms). Twenty three per cent of the children were clssified s hving both pneumoni nd fever. Assessment nd clssifiction of the sick child The qulity of the ssessment of the child s illness, when mesured ginst IMCI stndrds, ws very low in ll fcilities (Tble 4). None of the children ws checked for the three dnger signs (unble to drink or brestfeed, vomiting ll feeds, or convulsions) or hd their weight checked ginst growth chrt. Very few were checked for the presence of cough, dirrhoe nd fever, or other problems. Almost none of the children who were identified s hving very low weight on subsequent re-exmintion by the surveyors were ssessed for feeding prctices by the fcility providers. The overll index of ssessment hd men score of 23 out of mximum of Bulletin of the World Helth Orgniztion April 2005, 83 (4)

4 S.E. Arifeen et l. Reserch Tble 4. Proportions of children nd cregivers for whom specific cse-mngement tsks were performed by providers in first-level helth fcilities in Mtlb thn, Bngldesh, 2000 (weighted estimtes) Indictor No. of children or no. of Percentge (n) or men for crers eligible for tsk whom tsk ws performed Assessment of the sick child Child checked for three dnger signs b (0) Child checked for the presence of cough, dirrhoe nd fever (43) Child s weight checked ginst growth chrt (0) Child under 2 yers of ge ssessed for feeding prctices b (0) Child checked for other problems (15) Child with very low weight ssessed for feeding problems b (1) Index of integrted ssessment (men) (rnge 0 100) b c Clssifiction of the sick child Child ws correctly clssified (56) Child with very low weight ws correctly clssified b (1) Tretment of the sick child Child with pneumoni treted correctly (8) Child with dehydrtion treted correctly 0 0 (0) Child with nemi treted correctly 43 0 (0) Child needing n orl ntibiotic ws prescribed the drug correctly (12) Child not needing ntibiotics left the fcility without ntibiotic (68) Child received first dose of tretment t the fcility b 97 0 (0) Child needing referrl ws referred (4) Advice nd counselling given to cregiver of sick child Cregiver of sick child ws dvised to give extr fluids (12) nd continue feeding b Child prescribed orl mediction: cregiver ws dvised on how (13) to dminister the tretment Sick children whose cretker ws dvised on circumstnces (1) indicting need to return immeditely to helth cre fcility b Child with very low weight whose cregiver received correct (1) counselling b Cregiver of child prescribed orl rehydrtion solution, nd/or n (16) orl ntibiotic knows how to give the tretment The dult (usully prent) who ccompnied the sick child to the helth fcility. b IMCI-specific indictor. c Mens nd rnges re provided for the composite indices. Note: Percentges re weighted estimtes wheres the numbers (n) re ctul, nd therefore will not exctly correspond. The helth providers clssified the child s illness correctly for only one in five children (Tble 4). As expected, lmost none of the children whose weights were very low were correctly clssified. Tretment of the sick child Tretment prctices were generlly poor (Tble 4). None of the children presenting with nemi ws treted correctly nd only 13% of children with pneumoni were treted correctly. Among the children who needed n orl ntibiotic, 81% received one (dt not shown) nd 11% received correct prescription for n pproprite ntibiotic t the correct dose. About one in three (39%) of children who did not need n ntibiotic left the fcilities without one; over hlf of the prescriptions for ntibiotics were unnecessry. None of the children received the first dose of their tretment t the fcility. Four of eight very sick children (45%, weighted estimte) who needed referrl bsed on ssessment by the gold stndrd surveyor were lso identified s needing referrl by the helth worker. Bulletin of the World Helth Orgniztion April 2005, 83 (4) Advice nd counselling given to cregiver The helth providers in these fcilities mde little effort to explin the necessry home tretment or to counsel the cregiver (Tble 4). One in every 20 cregivers ws dvised to give extr fluids nd to continue feeding the sick child, nd this finding held true even mong the subset of children who presented with dirrhoe (dt not shown). Fewer thn one in 10 cregivers of children prescribed n orl mediction were dvised on how to dminister the tretment. None of the cregivers of children who were prescribed orl rehydrtion solution, nd/or n orl ntibiotic could report correctly, s they left the fcility, on how to give the tretment to the child t home. Only one of 274 cregivers ws dvised by the helth provider bout signs indicting need for immedite return to the helth fcility. Fcility prepredness The fcilities were not well equipped nd supported (Tble 5). Although reltively high proportions of helth fcilities were found to hve the essentil orl tretments for modertely 263

5 Reserch Tble 5. Mesures of helth fcility prepredness for the mngement of sick children in 19 first-level helth fcilities in Mtlb thn, Bngldesh, 2000 Indictor % (n) or men for whom tsk ws performed Index of vilbility of essentil orl 69.9 b tretments (men) (rnge 0 100) Helth fcility hs essentil 0 (0) equipment nd mterils Helth fcility hd received t lest 0 (0) one supervisory visit tht included observtion of cse mngement during the previous 6 months IMCI-specific indictor. b Mens nd rnges re provided for the composite indices. Note: Percentges re weighted estimtes wheres the numbers (n) re ctul, nd therefore will not exctly correspond. ill children, the injectble drugs needed to tret more severe disese were not generlly vilble. None of the fcilities hd received supervisory visit tht hd included observtion of cse mngement in the previous 6 months. Compring performnce by provider type Tble 6 shows the comprison of the performnce of the MAs/ SACMOs or FWVs who together sw more thn 90% of children in the survey using the subset of indictors performed correctly for t lest 5% of children (see Tble 4). FWVs performed mrginlly better thn MAs/SACMOs in clssifying illnesses correctly, nd significntly better thn MAs/SACMOs in the rtionl prescription of ntibiotics nd on two mesures of provision of correct dvice to cregivers. MAs/SACMOs showed tendency to conduct more complete ssessment of the child by checking for cough, dirrhoe nd fever, but the difference ws not sttisticlly significnt. No differences were found between the two types of helth-cre providers in the performnce of other cse-mngement tsks, possibly becuse too few observtions for FWVs were mde for some indictors. For the four indictors tht showed lrge but non-significnt differences between providers, two fvoured the MAs/SACMOs nd two the FWVs. Discussion The present study ws n observtion-bsed ssessment of the qulity of cre provided to sick children in ll first-level government helth fcilities in n re of Bngldesh. Governmentprovided helth services in the study re were similr to those provided throughout the country (14). Low utiliztion of these services becuse of preference for locl privte sources of cre is common in ll res of Bngldesh (19). The vlidity of our findings on provider performnce ws likely to be ffected by the presence of n observer (20), but given the overll low performnce levels this potentil bis is not considered to hve serious implictions for the interprettion nd use of the survey findings. Sick children re receiving indequte cre The findings indicte tht the current qulity of cre offered to sick children in these fcilities is very poor, even when IMCIspecific items re excluded from the ssessment. Sick children re incompletely ssessed, nd their illnesses re erroneously or S.E. Arifeen et l. incompletely clssified. The mjority of children receive incorrect tretment. Antibiotics re frequently overused nd sometimes underused. Severely ill children re not lwys referred to higher levels of cre by helth worker. Some of the fcility supports for correct cse-mngement of sick children re present in these fcilities: high proportion hd the essentil orl drugs vilble to mnge ll but the most severe diseses. However, most fcilities lcked t lest some items of essentil equipment nd few hd the injectble drugs needed to mnge severe illness. In summry, indequcies in the cse mngement of less severe disese cnnot be ttributed to the lck of the necessry supplies nd equipment, nd should therefore benefit from IMCI interventions designed to improve performnce of helth workers. Comorbidity mong children presenting for cre Two of the ssumptions underlying the IMCI strtegy re tht: children re often sick with more thn one illness t the sme time; nd disese-specific progrmmes such s those developed in the pst for tckling dirrhoe nd pneumoni were not designed to hndle children with multiple illnesses ppropritely nd fully (21). The findings of the present study provide evidence tht supports these ssumptions. Most sick children presented t the fcilities with severl concurrent illnesses. This is consistent with findings from the bseline household survey conducted mong the sme popultion (S.E. Arifeen, personl communiction), nd underlines the relevnce of IMCI to child helth needs in Bngldesh. Although ll but one of the children with severl concurrent illnesses presented with n IMCI clssifiction, bout one-third lso presented with condition not included in the Bngldesh IMCI guidelines most often with skin infections. Future reserch should investigte the mix of presenting symptoms nd its reltionship to morbidity in children ged under 5 yers in specific settings to help in the dpttion of the generic IMCI guidelines, in wy similr to tht used to determine the distribution of mortlity by cuse (1). Our ssumption, which would need to be tested, is tht if helth workers re trined to ddress the symptoms tht re most disturbing to cregivers, utiliztion of the services would increse for both life-thretening nd less severe conditions. Uptke of study findings by the Government of Bngldesh The existing policies of the Government of Bngldesh give priority for trining first to physicins, nd second to nurses nd MAs/SACMOs. At the time of our study the government s plns did not include IMCI trining for FWVs or phrmcists. This study showed tht one-third of the sick children were being mnged by FWVs who were ineligible for IMCI trining nd tht these FWVs performed s well s, if not better thn, the MAs/SACMOs. The results suggested tht the plns of the Government of Bngldesh for IMCI cse-mngement trining should be expnded to include ll the mjor ctegories of helthcre providers who were ctully mnging sick children in firstlevel helth fcilities. In response, the Government of Bngldesh greed in principle to offer IMCI trining for FWVs to increse the popultion-level impct of IMCI. The government is lso promoting nd supporting the implementtion of IMCI in the helth fcilities of nongovernmentl orgniztions. 264 Bulletin of the World Helth Orgniztion April 2005, 83 (4)

6 S.E. Arifeen et l. Reserch Tble 6. Performnce of selected cse-mngement tsks by cdre of helth-cre provider in Mtlb thn, Bngldesh, 2000 Medicl ssistnts/sub-ssistnt community medicl officers Fmily welfre visitors Indictor n (n) b % 95% CI c n (n) b % 95% CI P-vlue Child checked for the presence of cough, 184 (35) (8) dirrhoe nd fever Child checked for other problems 79 (8) (6) Child ws correctly clssified 179 (30) (26) Child with pneumoni correctly treted 45 (6) (1) Child needing n orl ntibiotic ws prescribed 74 (7) (4) the drug correctly Child not needing ntibiotics left the fcility 105 (35) (33) without ntibiotic Cregiver of sick child ws dvised on how to give 179 (4) (8) extr fluids nd continue feeding d Child prescribed orl mediction whose cregiver 134 (4) (9) ws dvised on how to dminister tretment n Men 95% CI n Men 9.5% CI Index of integrted ssessment (men) (rnge 0 100) d, e N = eligible smple. b n = numertor. c CI = confidence intervl. d IMCI-specific indictor. e Mens nd rnges re provided for the composite indices. Note: Percentges re weighted estimtes while the numbers (n) re ctul, nd therefore will not exctly correspond. The finding of bysml qulity of service hd the benefit of showing tht there ws much room for improvement. The ongoing MCE intervention study now hs the opportunity to show substntil impct becuse the findings presented here hve defined the specific chllenges tht need to be ddressed. None of the fcilities surveyed hd received supervision visit tht included observtion of cse-mngement in the previous 6 months. Although fcilities were found to be reltively well-supplied with orl drugs, the injectble drugs needed for mnging severe illness were rrely vilble. The Government of Bngldesh used the findings of this study to refine nd improve their plns for the ntionwide implementtion of IMCI. Continued work in the study district hs provided mple support for these decisions, s continuous monitoring hs demonstrted incresed utiliztion of the IMCI intervention fcilities reltive to the comprison fcilities, nd importnt nd sustined improvements in the qulity of cre provided to sick children (9). Conclusions The study findings provide useful guidnce not only for policymkers in Bngldesh, but for those in other settings with similr epidemiologicl profiles nd similr qulity of service. The IMCI strtegy provides n pproprite wy forwrd for the Government of Bngldesh to improve fcility-bsed services. Country-specific dpttions of the generic IMCI csemngement guidelines will need to be periodiclly reviewed to reflect chnges in the mix of presenting illnesses nd cuses of mortlity. The cdres of helth cre providers trgeted for IMCI cse mngement trining must include those who re ctully mnging sick children in helth fcilities. Locl-level support for correct performnce of helth workers, including supportive supervision, must be estblished nd sustined. Field-bsed evlutions such s this one, conducted in close collbortion with ministries of helth, cn provide the evidence needed to guide the development nd full implementtion of more effective child survivl progrmmes. The MCE is committed to providing techniclly sound feedbck nd responding to the dt needs of those responsible for mking decisions on mternl nd child helth t the locl, ntionl nd interntionl levels. This study will be complemented by others being undertken in Bngldesh nd t other MCE sites to exmine the effects of IMCI on the performnce of helthcre providers nd fmilies, nd their costs. O Acknowledgements Thnks re due to the helth fcility stff in the survey re for their willingness to be observed, especilly prior to the introduction of IMCI. We lso thnk the mothers nd children who prticipted in the study. We wish to thnk Dr Thierry Lmbrechts of WHO, Genev, for trining the tems of surveyors, nd for his erly work on the development of the MCE indictors nd helth fcility survey methodology. The MCE Technicl Advisers, nd prticulrly Dr Cesr Victor nd Dr Jonn Schellenberg, nd Dr Robert Scherpbier of WHO provided importnt suggestions nd guidnce throughout the study. This study ws conducted t the ICDDR,B: Centre for Helth nd Popultion Reserch with support from the Bill nd Melind Gtes Foundtion through grnt to the WHO Bulletin of the World Helth Orgniztion April 2005, 83 (4) 265

7 Reserch Deprtment of Child nd Adolescent Helth nd Development nd of Coopertive Agreement #388-A from the United Sttes Agency for Interntionl Development. ICDDR,B cknowledges with grtitude their commitment to the Centre s reserch effort. This work is prt of the Multi-Country Evlution of IMCI Effectiveness, Cost nd Impct (MCE), rrnged, coordinted nd funded by WHO s Deprtment of Child nd Adolescent Helth nd Development, nd with the finncil S.E. Arifeen et l. support of the Bill nd Melind Gtes Foundtion nd the US Agency for Interntionl Development. Finlly, our hertfelt thnks go to those working in the re of child helth from the Government of Bngldesh nd its prtners. Their commitment to using the results of the reserch to improve child helth plns nd policies in Bngldesh ws continuing source of motivtion. Competing interests: none declred. Résumé Qulité des soins dispensés ux enfnts de moins de cinq ns dns les instlltions de snté de premier niveu d un district du Bengldesh Objectif L évlution multintionle de l efficcité, des coûts et de Résultts Prmi les enfnts soignés dns ces instlltions, peu l impct de l Prise en chrge intégrée des mldies de l enfnce ont bénéficié d une évlution complète ou d un tritement correct (PCIME) est une évlution à l échelle mondile, destinée à et presque ucune des personnes s occupnt de ces enfnts déterminer l incidence de ce progrmme sur les résultts snitires n reçu de conseils sur l mnière de poursuivre les soins à et son rpport coût-efficcité. De évlutions de ce type sont en domicile. Environ un tiers des enfnts mldes dont les soins ont cours u Bengldesh, u Pérou, en Ougnd et en République- été éxminés ont été pris en chrge pr des employés disposnt Unie de Tnznie. L présente nlyse de l étude menée u d un niveu de formtion peu élevé, pour lesquels l probbilité Bengldesh vise à décrire l qulité des soins dispensés ux de clsser convenblement les enfnts mldes et de fournir ux enfnts mldes de moins de 5 ns dns les instlltions de snté personnes s occupnt d eux des informtions correctes sur les soins publiques de premier niveu, en vue d informer les responsbles à dispenser à domicile étit nénmoins nettement plus élevée que u niveu de l Ett de l plnifiction des progrmmes snitires pour le personnel soignnt ynt reçu une formtion supérieure. destinés ux enfnts. Conclusion Ces résultts démontrent qu il est urgent d intervenir Méthodes Des outils génériques permettnt d évluer les pour méliorer l qulité des soins dispensés ux enfnts mldes instlltions de snté sous l ngle de l efficcité, des coûts et de dns les instlltions de premier niveu u Bengldesh et lissent l impct ont été dptés, trduits et soumis à des tests prélbles. à penser qu il pourrit être profitble d étendre les formtions à l Des médecins formés à l PCIME et à l utilistion de ces outils ont prise en chrge des cs conformément à l PCIME ux employés mené l enquête dns 19 instlltions de snté des zones étudiées. disposnt d un fible niveu de formtion. Ils semblent indiquer Les données ont été collectées à prtir d observtions, d entretiens églement que l strtégie PCIME offre un éventil prometteur de sortie, d inventires et d entretiens vec les presttires de d interventions pour fire fce ux problèmes liés ux services de soins des instlltions. snté péditriques u Bengldesh. Resumen Clidd de l tención dispensd los menores de cinco ños en los centros snitrios de primer nivel de un distrito de Bngldesh Objetivo L evlución multipíses de l eficci, el costo estblecimientos pr recibir tención snitri, fueron pocos los y el impcto de l Atención Integrd ls Enfermeddes exmindos de form exhustiv y trtdos decudmente, y csi Prevlentes de l Infnci (AIEPI) es un estudio relizdo ninguno de sus cuiddores recibió instrucciones sobre l mner escl mundil con el fin de determinr el impcto de l AIEPI de seguir tendiendo l niño trs su regreso l domicilio. Más de en los resultdos snitrios y su eficci en relción con el costo. un tercio de los csos en que se observó cómo tendín l niño Se están llevndo cbo estudios de ese tipo en Bngldesh, enfermo fueron mnejdos por personl de nivel inferior que Brsil, Perú, Ugnd y Repúblic Unid de Tnzní. En el demostró un cpcidd significtivmente myor que l de otros cso de Bngldesh, el objetivo er describir l clidd de l trbjdores de nivel superior pr clsificr los niños enfermos tención que reciben los menores de cinco ños enfermos en correctmente y pr proporcionr l cuiddor l informción los estblecimientos snitrios públicos de primer nivel, con el necesri sobre l sistenci domiciliri. fin de portr informción l Gobierno de cr l plnificción Conclusión Estos resultdos ponen de mnifiesto l necesidd de los progrms de slud infntil. urgente de emprender intervenciones de mejor de l clidd de l Métodos Los instrumentos genéricos utilizdos pr relizr sistenci dispensd los niños enfermos en los centros snitrios este estudio en los estblecimientos snitrios fueron dptdos, de primer nivel de Bngldesh, y sugieren que puede ser positivo trducidos y sometidos pruebs preliminres. Médicos con incluir trbjdores de nivel inferior entre los destintrios de formción sobre l AIEPI y dichos instrumentos llevron cbo el l formción pr el trtmiento de csos en el mrco de l AIEPI. estudio en los 19 centros snitrios de ls zons nlizds. Los Los resultdos obtenidos permiten pensr que l estrtegi de l dtos fueron obtenidos medinte observciones, entrevists l AIEPI ofrece un conjunto de intervenciones muy prometedor pr slid, inventrios y encuentros con los dispensdores de slud. hcer frente los problems que pdecen los servicios de slud Resultdos Entre los niños enfermos que cudieron infntil de Bngldesh. 266 Bulletin of the World Helth Orgniztion April 2005, 83 (4)

8 S.E. Arifeen et l. Reserch References 1. Blck RE, Morris SS, Bryce J. Where nd why re 10 million children dying every yer? Lncet 2003;361: Ahmed OB, Lopez AD, Inoue M. The decline in child mortlity: repprisl. Bulletin of the World Helth Orgniztion 2000;78: Culfield LE, de Onis M, Blossner M, Blck RE. Undernutrition s n underlying cuse of child deths ssocited with dirrhe, pneumoni, mlri nd mesles. Americn Journl of Clinicl Nutrition 2004;80: Gove S. Integrted mngement of childhood illness by outptient helth providers: technicl bsis nd overview. Bulletin of the World Helth Orgniztion 1997;75 Suppl 1: Deprtment of Child nd Adolescent Helth nd Development, World Helth Orgniztion. Globl Monitoring of IMCI Implementtion, December Genev, WHO. Avilble from: child-dolescent-helth 6. Bryce J, Victor C, Hbicht JP, Blck RE, Vughn P. The Multi-Country Evlution of the Integrted Mngement of Childhood Illnesses Strtegy. Americn Journl of Public Helth 2004;94: World Helth Orgniztion. Multi-country evlution: integrted mngement of childhood illness. Genev, WHO. Avilble from: 8. Gouws E, Bryce J, Hbicht JP, Amrl J, Priyo G, Schellenberg JA, et l. Improving the use of ntimicrobils through IMCI cse mngement trining. Bulletin of the World Helth Orgniztion 2004;82: Arifeen SE, Blum LS, Hoque DME, Chowdhury EK, Khn R, Blck RE, et l. Utiliztion of public helth fcilities responds to implementtion of integrted mngement of childhood illness: erly experience from Bngldesh. Lncet 2004;364: Amrl J, Gouws E, Bryce J, Leite AJM, Cunh ALA, Victor CG. Effect of Integrted Mngement of Childhood Illness (IMCI) on helth worker performnce in Northest-Brzil. Cd Sude Public 2004;20: Victor CG, Hnson K, Bryce J, Vughn JP. Achieving universl coverge with helth interventions. Lncet 2004;364: Bryce J, El Arifeen S, Priyo G, Lnt C, Gwtkin D, Hbicht JP nd the Multi-Country Evlution of IMCI Study Group. Reducing child mortlity: Cn public helth deliver? Lncet 2003;362: Interntionl Centre for Dirrhoel Diseses Reserch, Bngldesh (ICDDR,B): Centre for Helth nd Popultion Reserch. Helth nd Demogrphic Surveillnce System-Mtlb, v.34. Registrtion of helth nd demogrphic events Dhk: ICDDR,B; Hider SJ, Stretfield K, Krim MA. Comprehensive guidebook to the Bngldesh Fmily Plnning-MCH Progrm. Dhk: Reserch Evlution Assocites for Development, The Popultion Council-Bngldesh, Ministry of Helth nd Fmily Welfre; World Helth Orgniztion. Helth Fcility Survey for Integrted Child Helth Services (IMCI). Finl drft, Mrch Genev: WHO. Avilble from: Gouws E, Bryce J, Schellenberg J, Priyo G, Hbicht JP. Mesuring the qulity of child helth cre t first level fcilities. Socil Science nd Medicine. In press. 17. Epi Info 6.0. Avilble from: Sttcorp. Stt Sttisticl Softwre: Relese 7.0. College Sttion, TX: Stt Corportion; Amin R, Chowdhury SA, Kml GM, Chowdhury J. Community helth services nd helth cre utiliztion in rurl Bngldesh. Socil Science nd Medicine, 1989;29: Anker M, Guidotti RJ, Orezeszyn S, Spirie SA, Thuriux MC. Rpid evlution methods (REM) of helth services performnce: methodologicl observtions. Bulletin of the World Helth Orgniztion, 1993;71: Tulloch J. Integrted pproch to child helth in developing countries. Lncet, 1999;354 Suppl 2: Bulletin of the World Helth Orgniztion April 2005, 83 (4) 267