Costs of treating diarrhoea in a children's hospital in Mexico City.

Research!Recherche Costs of treating diarrhoea in a children's hospital in Mexico City M. Phillips,' J. Kumate-Rodriguez,2 & F. Mota-Hernandez3 The treatment received by children aged under 5 years with diarrhoea was studied in the Hospital Infantil de Mexico (Federico Gom6z), Mexico City. The costs of treatment were calculated and estimates were made of how these had changed since the establishment of an oral rehydration unit in the hospital in 1985. The results indicate that drug treatment of outpatients was generally appropriate and inexpensive. In contrast, the cost of drugs for inpatients was considerably higher. The seriousness of the cases justified much of this additional expense for inpatients, but there is evidence that the costs could be reduced further without jeopardizing the quality of the care. Diagnostic tests were relatively expensive, frequently failed to identify diarrhoeal etiology, and their results correlated poorly with the treatment prescribed. The oral rehydration unit resulted in significant savings by causing a 25% fall in the number of inpatients with diarrhoea. Introduction The efficacy of oral rehydration therapy in the man- agement of childhood diarrhoea is well established (1). Despite this, inappropriate use of other treat- ments, such as intravenous fluids, antibiotics, and antidiarrhoeal agents, continues in many parts of the world. Some of these strategies have legitimate appli- cation in specific circumstances; however, in most instances they are unnecessary, ineffective, and some- times potentially harmful. Furthermore, these medi- cations are considerably more expensive than the usually more appropriate alternatives of no medica- tion or oral rehydration therapy. This is an impor- tant consideration, particularly because diarrhoea is one of the most common causes of morbidity in developing countries, and drugs often account for more than 20% and sometimes up to 60% of health expenditure in such countries (2). Rationalizing "drug" treatment for diarrhoea could, therefore, be expected to make an important impact on health ' Research Fellow, Evaluation and Planning Centre, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1 7HT, England. Requests for reprints should be sent to this address. 2 Secretary of State for Health, Mexico City, Mexico. 3 Chief, Education Division and Oral Rehydration Service, Hospi- tal Infantil de Mbxico (Federico Gomez), Mexico City, Mexico. Reprint No. 4976 budgets by releasing resources that are currently being expended on inappropriate treatment. Such is the case in Mexico, where diarrhoea among young children is a frequent and serious problem, and where there is evidence that inapprop- riate drug treatment is common. Objectives and methods The study examined the financial costs of diarrhoea treatment in the Hospital Infantil de Mexico (Federico Gomez), a large children's hospital in Mexico City. This hospital, one of the specialized hospitals of the Secretaria de Salud, caters specifi- cally for children from poorer families. It has about 300 beds and treats annually about 7000 inpatients and 130000 outpatients. Approximately 11% of emergency outpatients and 7.5% of inpatients are diagnosed as having diarrhoea. In April 1984 an oral rehydration "corner" was established in the hospital adjacent to the emergency department, and in 1985 this developed into a fully-fledged oral rehydration unit staffed by two doctors and two nurses. All emer- gency cases of diarrhoea were referred there. In addi- tion to the treatment received by their children in the unit, mothers were instructed on how to mix oral rehydration salts (ORS) and were provided with sachets of ORS to take home with them. The main purpose of the study was to estimate Bulletin of the World Health Organization, 67 (3): 273-280 (1989) (© World Health Organization 1989 273 M. Phillips et al. the cost of medication for patients who attended the hospital, either as inpatients or outpatients, and to measure the financial implications of the oral rehydration unit, principally in terms of savings in drug expenditure. As the study progressed, a further objective emerged-to explore the nature and cost of drugs used by children prior to attending the hospi- tal. The study employed a methodology developed by the WHO Diarrhoeal Diseases Control prog- ramme.a The clinical records of diarrhoea patients aged under 5 years were examined and the details of treatment provided or prescribed extracted. A sample of 40 to 50 records was taken for both 1986 and 1983 (or the year nearest 1983 for which records were available) from inpatients, emergency patients, and those who had attended the oral rehydration unit. For inpatients, diarrhoea was taken to include all diagnoses from 001 to 009 (except typhoid-002) in the list of three-digit categories, International Classification of Diseases (3). Records of patients who died or who had more than one serious condi- tion that was unrelated to diarrhoea were excluded. For each patient information on as many of the following as possible was obtained: age, sex, month when they visited the hospital, diagnosis, dehydra- tion status, malnutrition and breast-feeding status, types and amounts of medication and solutions given or prescribed in the hospital, types and numbers of clinical laboratory tests performed (and, for faecal cultures and parasitological tests, their results), as well as the types of medication taken before visiting the hospital. The prices of medicines used by inpatients were obtained from the hospital pharmacy records. For drugs prescribed in hospital, outpatients were assumed to have paid these prices plus 6% tax and 15% handling charges. The cost of drugs taken by patients prior to visiting the hospital was estimated from information obtained from a commercial phar- macy. Based on data obtained from the hospital and prices charged in a commercial clinic, it was conser- vatively estimated that the average cost per labor- atory test was 1000 Mexican pesos. All prices are expressed in December 1986 levels (1000 pesos = US$ 1.00). Results and discussion Emergency hospital treatment With the important exception of a greater than twofold increase in the use of ORS, drug treatment a Estimating costs for cost-effectiveness analysis-guidelines for managers of diarrhoeal disease control programmes. Unpub- lished document CDD/SER/88.3. for diarrhoea outpatients at the hospital changed little between 1984-85 and 1986 (Table 1). There appear to have been no major modifications to the general approach of treating emergency diarrhoea patients that could reduce costs while maintaining the quality of care: none of the outpatients surveyed were prescribed antidiarrhoeals (although Kaopectate®, for example, was still available in the hospital pharmacy); antibiotics were prescribed for less than 10% of patients and prescriptions for anti- pyretics fell from 20% in 1984 to 4% in 1986. The costs of medication are borne by the patient and for emergency outpatients are on average less than 200 pesos per person. Many of the children seen at the oral rehydration unit and provided with packets of ORS showed no signs of dehydration. It could be argued that the hospital incurs unnecessary expense by sending all emergency diarrhoea patients, whether or not they are suffering from dehydration, to the unit. However, this ignores the benefits derived from educating mothers and providing ORS in an environment where childhood diarrhoea is common and will probably be experienced again by the child or by a sibling. Laboratory tests were carried out on a higher proportion of diarrhoea outpatients in 1986 (40%) than in 1984 (12%) and, for those having any test, the average number of tests per patient increased from two to three; tests to identify diarrhoeal eti- ology accounted for most of this increase. Of the 50% of diagnostic tests for which the results were recorded, none was positive, and laboratory costs probably imposed a greater burden on patients than the cost of hospital-prescribed medication. Treatment of inpatients with diarrhoea Except for an increase in the proportion of males in the survey sample, there were no significant differ- ences in the characteristics of the inpatients in 1983 and 1986 (Table 2). Also, the pattern of medication prescribed changed little, apart from a substantial increase in the use of ORS and a shift in antibiotic use away from amikacin to amoxicillin and chlor- amphenicol (Table 3). Not unexpectedly, the cost of drug treatment for hospitalized cases was considerably greater than that for emergency outpatients. At less than 10000 pesos per patient, however, it was lower than that predict- ed from data on the total hospital expenditure on drugs and the total number of inpatientg, which sug- gested average drug costs of about 75000 pesos per inpatient (for all conditions). This difference may have arisen for any of the following reasons: the medication required to treat diarrhoea was cheaper 274 WHO Bulletin Vol. 67 1989 Costs of treating diarrhoea In a children's hospital In Mexico Clty Table 1: Distribution of treatment prescribed and tests given to samples of 50 outpatients and the estimated average cost, Hospital lnfantil de Mixico (Federico Gom.z), Mexico City, 1984 and 1986 Emergency ward Oral rehydration unit Treatment/test 1984 1986 1986 Treatment' Hygiene-diet advice 31 (62)b 19 (38) 0 Oral electrolytes: 14 (28) 35 (70) 42 (84) Quantity recommended (ml)c - - 353 Quantity given (ml) 272 Time spent (hours) - -2.4 Intravenous therapy 2 (4) 4 (8) 0 Antibiotics 3 (6) 4 (8) 0 Antipyretics 10 (20) 2 (4) 0 Other treatments 0 0 0 Average drug cost per patientd Borne by patient 150 180 Borne by hospital - - 130 Laboratory tests Total number 12 59 N.A.' Average number per patient 0.2 1.2 N.A. No. of patients who had any 6 (12) 20 (40) N.A. test Average number of tests per 2.0 3.0 N.A. outpatient tested Average cost of tests per patientd Assuming 1000 pesos per test 200 1200 N.A. Amount charged to parents' 155 980 N.A. ' No. of outpatients receiving treatment. b Figures in parentheses are percentages. c On average, considerably less than 1 litre of ORS was given to children while they were in the unit; however, mothers were provided with three packets of ORS to take home with them (40 pesos per packet). d Costs are shown in Mexican pesos (1000 pesos = US$ 1.00 in December 1986). * Based on the middle-income category of patient, i.e., total family income of 150000 pesos (US$ 150) per month. f N.A. = not available. Table 2: Characteristics of samples of 40 diarrhoea Inpatlents, Hospital Infantil de Mexico (Federico Gomez), Mexico City, 1983 and 19868 Characteristics 1983 1986 Average age (months) 9.6 6.5 No. of males 18 (45)b 31 (78) No. who were first-born children 12 (30) 12 (34); n = 35 No. currently breasffeeding 7 (19); n = 38 6 (16); n = 38 Nutritional status Normal 10 (26) 12 (30) Grade (mild malnutrition) 10 (26) 12 (30) Grade II (moderate malnutrition) 13 (33) 11 (28) Grade Ill (severe malnutrition) 6 (15); n = 39 5 (12) Dehydration status None 12 (33) 13 (37) Mild 6 (17) 9 (23) Moderate 13 (36) 8 (23) Severe 5 (14); n =36 6 (17); n =36 Diarrhoea-related conditions Hyponatraemia, acidosis, or disequilibrium 11 (28) 4 (10) Convulsions 5 (13) 7 (18) Shock 2 (5) 0 Medication poisoning 6 (15) 4 (10) Average length of hospitalization (days) 7.1 5.9 ' Information on all these characteristics was not available for all the patients; where the sample size was not 40, it is shown. b Figures in parentheses are percentages. 27SWHO Bulletin Vol. 67 1989 M. Phillips et al. Table 3: Distribution of the number of diarrhoea Inpatients In samples of 40 who received treatment and Its estimated average cost, Hospital Infantil de M6xico (Federico Gom4z), Mexico City, 1983 and 196 Intravenous rehydration Oral rehydration Antipyretics Anticonvulsants Vitamin K Antidiarrhoeals Antimicrobials At least one "systemic" At least two "systemic" Ampicillin Amikacin Amoxicillin Furazolidine Co-trimoxazole Gentamicin Penicillin Metronidazole Chloramphenicol Topical antibiotics No. of inpatients who received "systemic" antibiotics for <5 days Average cost of medication per inpatientb Laboratory tests Total number Average number per inpatient % of inpatients who had any test Estimated average cost per inpatient (pesos)c ' Figures in parentheses are percentages. b Costs are shown in Mexican pesos (1000 pesos = US$ 1.00 in December 1986). c Calculated assuming 1000 pesos per test. than that for the average inpatient; the sample sur- veyed was drawn from patients who were suffering from less complicated diarrhoea; or the estimates based on the clinical records did not take into account wastage, e.g., the discarding of partially-used ampoules and of drugs not stored properly or past the expiry date. The costs of the medication prescribed for diar- rhoea inpatients could probably be reduced without detriment to the patient. Although the proportion of these inpatients who received antibiotics fell some- what compared with 1983, it was still nearly 50% in 1986. Most of these received at least two antibiotics (usually ampicillin and amikacin). The prescription of antibiotics for nearly half the patients sampled is difficult to justify: none of them had a final diagnosis of septicaemia (the exclu- sion of patients who died probably partly accounts for this) and only a small number of patients had other infections for which systemic antibiotics might possibly be required (in 1986 two patients had bronchitis). In view of the potentially serious side- effects of amikacin, its relatively frequent use for diarrhoea inpatients merits further attention. The majority of diarrhoeas in 1983 and 1986 were of unidentified etiology, although in 1986, for example, faecal cultures were carried out on 30%, coproparasitology on 25%, and faecal smears on 70% of patients. There were no cholera cases and only one of the patients who was given antibiotics was diagnosed as infected with Shigella spp.-one of the few diarrhoeal etiologies for which antibiotics are recommended. In 1986, 25% of the patients in the sample were treated with antibiotics, despite having a negative culture result. An estimated 2 million pesos would have been saved in 1986 had these patients not received antibiotics. Interestingly, greater use was made of anti- biotics when faecal cultures had been carried out but were negative than in instances when no culturing was performed at all. This may reflect a tendency to either concurrently request tests to be carried out and to initiate use of antibiotics in situations where clinical evidence points to bacterial infections or to use cultures to confirm the absence of infection in order to determine whether treatment should stop. If, however, the apparent discrepancy between the treatment given and test results arose because the methods used had low sensitivity or because anti- biotics received prior to hospitalization were WHO Bulletin Vol. 67 1989 1983 1986 38 (95)' 0 1 (3) 4 (10) 14 (35) 0 23 (58) 17 (43) 18 (45) 15 (38) 0 1 (3) 2 (5) 1 (3) 3 (8) 1 (3) 0 7 (18) 14 (35) 7780 508 12.7 100 12700 34 (85) 16 (40) 0 4 (10) 9 (22) 0 18 (45) 16 (40) 17 (43) 10 (25) 6 (15) 0 2 (5) 1 (3) 0 0 3 (8) 6 (15) 5 (12) 9530 684 17.1 98 17 100 276 Costs of treating diarrhoea In a children's hospital in Mexico City masking the infection, the value of carrying out the tests is questionable-particularly because culture techniques are relatively expensive and may impose other costs by unnecessarily delaying the length of time patients stayed in hospital. Additional training of medical staff may be necessary if adequate under- standing of appropriate treatment of diarrhoea has led to overuse of antibiotics. Of the laboratory analyses carried out on inpa- tients with diarrhoea, the most frequent were general urine and blood tests and assessment of the level of serum electrolytes. The average number of tests per diarrhoea inpatient increased from 13 in 1983 to 17 in 1986. At a conservatively estimated average cost per test of 1000 pesos, the total cost of laboratory analyses per diarrhoea inpatient exceeded that of medication and in 1986 averaged approximately 17 000 pesos. Savings could also be made by reducing the fre- quency and duration of intravenous therapy. In 1986, 85% of inpatients received intravenous solu- tions, sometimes only in small volumes as a vehicle for administering intravenous drugs but usually also as part of rehydration therapy. This represents a reduction from the 95% of inpatients who received such solutions in 1983, but there is scope for further improvements: only 17% of sampled cases were severely dehydrated, while a further 22% suffered from electrolyte imbalances or convulsions that on clinical grounds might have justified the use of intra- venous over oral rehydration. Also, most patients appear to have received intravenous rehydration therapy for longer than the 6 hours generally con- sidered adequate. One reason for this frequent and extended use of intravenous rehydration could be that oral rehydration requires more time than nursing staff have at their disposal. If this is true, consideration should be given to relaxing restrictions on parental access to hospitalized children. The potential dangers of intravenous administration of medicaments make it all the more important that appropriate ways to reduce its use be explored (4). Impact of the oral rehydratlon unit The oral rehydration unit has had a positive impact on the hospital's expenditure on treatment for diar- rhoea patients by treating those who would other- wise have become inpatients. Compared with the situation before 1985, the average annual number of diarrhoea inpatients treated in the hospital fell by 180 in 1985-86, a reduction of nearly 25%. The total number of diarrhoea patients (inpatients and outpatients) for 1985-86 was somewhat lower than that for 1983-84, but the difference was not large enough to account for the fall in the number of in- patients. Application of the average drug costs per inpatient sampled in 1983 (7780 pesos) to the pre- sumed 180 hospitalized cases that were avoided in 1985-86 gives an estimated saving in drug costs of 1 400 400 pesos per annum. The fall in the number of hospitalized diarrhoea patients saves not only drugs but also other resources such as food, bedlinen, and the time of medical and administrative staff, which is potentially then freed for use with other patients. Estimates indi- cate that the savings in inpatient costs amounted to 56 million pesos per annum (Table 4). The estimate is conservative since it is based on the assumption that the average length of stay for those who would Table 4: Estimated annual savings resulting from the establishment of the oral rehydration unit, Hospital Infantil de M6xico (Federico Gomez), Mexico City Cost (pesos) (1000 pesos = US$ 1.00) Total hospital running expenses' 5800 millionb Estimated expenditure on inpatients 4640 millionc Cost per inpatient day 58000d Cost per inpatient staying 5 days 290000 For diarrhoea inpatients Cost of drugs per patient 7780 Cost of laboratory materials per patient 12700 Total cost per patient 310000 Total gross savings from averting 180 patients 55-8 million" ' Excluding drugs and laboratory, radiological, and photographic materials. b The actual expenditure during 1986 was 3970 million pesos. Given an inflation rate of 100%, and assuming an even distribution of hospital expenses and a constant rate of inflation throughout the year, this is equivalent to about 5800 million pesos in December 1986 prices. c Calculated assuming that this represents 80% of the total costs. d Based on the 80 000 inpatient days recorded in 1986. e The potential gross savings could be increased by more rational use of drugs and laboratory tests and more selective hospitalization. An estimation of the net savings would have to take account of the expense of establishing and running the oral rehydration unit, costs which were not measured in this study. WHO Bulletin Vol. 67 1989 277 M. Phillips et al. have been hospitalized but are now treated in the oral rehydration unit was 5 days. Estimates of net cost savings must also take into account the resources required for treatment in the unit. These costs have not, however, been included. Pre -hospital treatment of diarrhoea In 1986, about 25% of inpatients received some oral rehydration therapy prior to hospitalization, a sta- tistically insignificant increase over 1983 levels (Table 5). The dehydration status of these patients was on average, however, no better than that of the other patients, and in many instances was somewhat worse; this applied also to the sample of inpatients from 1983. This may reflect a tendency for those individuals with the most serious diarrhoea to use ORS or for the relatively small number of patients for whom ORS does not work to end up hospital- ized. However, it may be that improvements in the preparation, use, or nature of ORS are indicated. Approximately 50% of patients received at least one antimicrobial, and most of these more than one. Ampicillin, furazolidone, and metronidazole were the most commonly used, and many were taken as mix- tures with antidiarrhoeals, particularly kaolin and pectin. Kaopectate® was the most commonly used antidiarrhoeal: in 1986 more than 40% of inpatients and nearly 70% of outpatients purchased it in some form, usually mixed with antibiotics. Adsorbents such as kaolin are ineffective in preventing and treat- ing dehydration and are not indicated in the routine treatment of acute diarrhoeal disease.b There is also no convincing evidence that intestinal bacterial sup- plements (Lactobacillus cultures), which were used for pre-hospital treatment by 10% of emergency patients in 1986, are effective against diarrhoea. b A manual for the treatment of acute diarrhoea. Unpublished document WHO/CDD/SER/80.2, Rev. 1. Table 5: Distribution of the number of diarrhoea patients In samples of 40 who used medication prior to attending Hospital Infantil de Mexico (Federico Gom.z), Mexico City, 1983 and 19868 Inpatients Emergency ward Medication 1983 1986 1986 Total Intravenous rehydration 3 3 0 6 (5)b Oral rehydration 6 10 10 26 (20) Paedialyte® 5 4 5 14 (11) Antimicrobials At least one "systemic" 16 17 32 65 (50) At least two "systemic" 9 13 20 42 (32) Ampicillin 7 7 13 27 (21) Amikacin 2 3 1 6 (5) Gentamicin 5 4 6 15 (12) Co-trimoxazole 0 0 1 1 (1) Furazolidone 2 4 13 19 (15) Metronidazole 3 6 10 19 (15) Neomycin 0 4 2 6 (5) Penicillin 1 5 3 9 (7) Streptomycin 0 0 2 2 (2) Chloramphenicol 0 0 0 0 Antidiarrhoeals Kaopectates 2 4 12 18 (14) Kaopectate® + antibacterials 5 13 22 40 (31) Antimotility agents 4 10 6 20 (15) lodohydroxyquinoline 3 1 3 7 (5) Lactobacillus spp. 2 1 5 8 (6) Others 4 3 6 13 (10) Antipyretics 12 12 12 36 (28) Antiemetics 9 3 1 13 (10) Anticonvulsants 4 1 0 5 (4) Did not recall 3 3 2 8 (6) Not recorded/no medicines used 5 4 3 12 (9) No. of drugs per patient 2-4 2-8 2-6 2-6 Cost of medication per patient (pesos) 2230 2920 2630 2600 The percentages and average costs were calculated using as the denominator, all patients except those who explicitly "did not recall", i.e., assuming that where nothing was recorded, no medicine was taken. b Figures in parentheses are percentages. WHO Bulletin Vol. 67 1989278 Costs of treating diarrhoea In a children's hospital In Mexico City While both these groups of antidiarrhoeals are of questionable efficacy, neither has been found to have serious side-effects. The same cannot be said, however, for other antidiarrhoeals, particularly anti- motility drugs such as loperamide and mixtures of diphenoxylate and atropine. Not only are such drugs not indicated for the routine treatment of acute diar- rhoeal diseases but they can be very dangerous (even fatal) if prescribed for infants. Nevertheless, 15% of patients had received at least one antimotility drug. Of the patients studied, 5% had taken prep- arations containing iodohydroxyquinoline prior to hospitalization. Halogenated hydroxyquinolines can cause serious and irreversible injury to the nervous system, and it is now argued that the only possible justification for their oral administration is in the treatment of intestinal amoebiasis-and even then, their use is highly questionable since other effective and safer drugs are available to treat this condition (5). None of the patients who received iodohydroxy- quinoline had been diagnosed as having amoebiasis. The potential dangers of inappropriate drug use are highlighted by the more than 12% of inpatients in 1983 and 1986 who exhibited some kind of adverse effects to the medication they had received prior to hospitalization. Costs to patients Overall, patients contributed a relatively small amount to the running costs of the hospital (about 5% in 1986).c Nevertheless, the total outlay made by parents towards treating children with diarrhoea could represent a significant proportion of their incomes. The hospital caters for the least privileged members of the community, many of whom are likely to have incomes bordering on the minimum wage, which for Mexico City was about 3000 pesos per day. The hospital takes family income into account in determining most of the fees it charges; none the less, if a child with diarrhoea is hospitalized for a week, even a low-income family may have to pay up to 15000 pesos for drugs and medical care, and possibly more if certain laboratory tests are also carried out. In addition they would almost always have sought treatment prior to hospital admission. For medication alone, this would probably have cost 3000 pesos, and in many cases would have been double this-to this must be added any doctor's fees that may have been paid. Travelling and accommo- dation expenses add further to the financial burdens on households. Many patients have to travel con- siderable distances to the hospital; for example, in c Hospital Inbnffl de M6x1co. [Statements and reports, 31 December 1986, for the Government Bulletin]. Unpublished docu- ment, 1987 (in Spanish). 1983, 60% of patients came from outside Mexico City and almost half of these from other states in the country.d Conclusions The Hospital Infantil de Mexico (Federico Gomez) is probably one of the better-run children's hospitals in Mexico and the existence of the oral rehydration unit bears witness to the hospital's concern to develop appropriate treatment strategies for children with diarrhoea. Furthermore, the unit has freed scarce resources for other uses in the hospital. Child- ren with diarrhoea appear to be suitably treated as outpatients, and there is little scope for reducing drug costs. However, rationalization of some aspects of inpatient treatment could reduce the combined costs of medication and laboratory tests. A study of the decision-making processes leading to assignment of treatment in the hospital could be helpful in guiding the development of training programmes for medical staff. The nature of the antidiarrhoeal treatment received by children before coming to hospital was less than satisfactory. Research into the treatment provided in other health facilities, the extent and nature of self-medication, and the role of pharmacies in encouraging the use of inappropriate drugs for diarrhoea is clearly indicated. There are several strategies for combating over- use of inappropriate drugs. Where the dangers from side-effects are substantial and the efficacy of the drug is doubtful or minimal, e.g., diphenoxylate, there is a good case for banning its production or sale on a national scale. The sale and use of other drugs, such as antibiotics, with identifiable valuable applications but hazardous consequences if misused, need to be regulated. Medical personnel, phar- macists, and the public should be informed about all such drugs, and also about others that are harmless, but expensive and ineffective. Acknowledgements This study was funded by the WHO Diarrhoeal Diseases Control programme. We wish to thank Mr Sanchez and his staff from the archives section of the Hospital Infantil de Mexico (Federico Gom6z). Dr L6pez de Montero (Pan American Health Organization) and Dr Escalante (Universidad Nacional Aut6noma de Mexico) for their valued contributions to this investigation. d Department of Blostatltss, HospItal Infantil de Mexico. [Annual statistics 1983]. Mexico City, Hospital Infantil de Mexico, 1983 (in Spanish). WHO Bulletin Vol. 67 1989 279 M. Phillips et al. Resume Cout du traltement de la dlarrhee dans un hopital pedlatrique a Mexico Le traitement re,u par des enfants de moins de cinq ans atteints de diarrhee a ete etudie a l'Hospital Infantil de Mexico (Federico Gomez) sur un echantillon aleatoire de 50 malades ambula- toires et 40 malades hospitalises, pour les annees 1986 et 1983 (ou l'annee la plus proche de 1983 pour laquelle on possede des dossiers de malades). On a calcule le coat du traitement et examine ses variations depuis la creation d'un service de rehydratation orale a l'h6pital en 1985. Les resultats de l'etude montrent que si le traitement medicamenteux administre aux malades ambulatoires etait en general approprie et peu coOteux (US$ 0,20 par enfant), le cout des medicaments 'tait beaucoup plus eleve pour les malades hospitalises chez qui il s'elevait a US$ 10 par enfant (chiffres pour 1986 dans les deux cas). Le surcout etait en grande partie justifie par la gravite de la diarrhee chez les malades hospital- ises mais il semble qu'il aurait ete possible de le reduire sans nuire a la qualite des soins. Par exemple, en 1986, alors que dans la plupart des cas la diarrhee etait d'etiologie inconnue, on administrait des antibiotiques par voie generale a pres de 50% des patients et on traitait 90% des cas par rehydratation intraveineuse alors qu'au maximum 15% d'entre eux etaient gravement deshydrates. Les epreuves diagnostiques pra- tiqu6es etaient relativement couteuses, souvent inefficaces et correspondaient peu au traitement prescrit. La creation d'un service de rehydratation orale a permis de realiser des economies subs- tantielles. Le nombre de malades hospitalises pour diarrh6e a baisse de 25%, ce qui correspond a une economie annuelle evaluee en gros a 56 millions de pesos et a une economie de medica- ments d'environ 1,4 million de pesos par an. La plupart des enfants avaient deja requ un traitement antidiarrheique avant d'etre conduits a l'h6pital. Ce traitement etait souvent d'efficacite douteuse et parfois meme dangereux: 15% des malades hospitalises pour diarrhee en 1983 et 10% en 1986 presentaient une intoxication m6dica- menteuse. Des recommandations concernant la fa9on dont le traitement de la diarrhee chez les malades hospitalis6s pourrait etre rationalise et les couts abaisses tout en preservant la qualite des soins ont ete formulees. Dans 1'ensemble, il est toutefois manifeste que cet hopital fonctionne de maniere satisfaisante et que 1'existence d'un service de rehydratation orale temoigne d'un reel souci d'amelioration de la strategie de soins aux enfants diarrheiques. Les donnees concernant les traite- ments re4us par les enfants avant leur arrivee a l'h6pital montrent en revanche que les soins fournis ailleurs sont loin d'etre aussi satisfaisants. References 1. Hlrshhorn, N. The treatment of acute diarrhea in child- ren: an historical and physiological perspective. American journal of clinical nutrition, 33: 637-663 (1980). 2. Mamdami, M. & Walker, G. Essential drugs in developing countries: a review and selected annotat- ed bibliography. London, London School of Hygiene and Tropical Medicine, 1985 (EPC Publication No. 8). 3. International Classification of Diseases, volume 1, 1975 revision. Geneva, World Health Organization, 1977. 4. Collins, R.N. et al. Risk of local and systemic infection with polyethylene intravenous catheters. New England journal of medicine, 279: 340-343 (1968). 5. Health Action International. Hydroxyquinolines. The Hague, Netherlands, International Organization of Consumers Unions, 1985. 260 WHO Bulletin Vol. 67 1989