Walsh,

J., and Warren, K. (1979). "Selective PHC -an interimstrategy for disease control in developing countries." The NewEngland Journal of Medicine 30(18): 967 -974

Vol. 301 No 18 DISEASE CONTROL IN DEVELOPING COUNTRIES -WALSH A:--'D WARREN

SPECIAL ARTICLE

SELECTIVE PRIMARY HEALTH CARE

An Interim Strategy for Disease Control in Developing Countries

JULIA A. WALSH, M.D., AND KENNETH S. WARREN, Yt.D.

Abstract Priorities among the infectious diseases af-fecting the three billion people in the less developedworld have been based on prevalence. morbidity. mor-tality and feasibility of control. With these priorities inmind a program of selective primary health careis compared with other approaches and suggest-ed as the most cost-effective form of medical inter-vention in the least developed countries. A flexi-ble program delivered by either fixed or mobileunits might include measles and diphtheria-per-

tussis-tetanus vaccination, treatment for febrilemalaria and oral rehydration for diarrhea in chil-dren, and tetanus toxoid and encouragement ofbreast feeding in mothers. Other interventions mightbe added on the basis of regional needs and new de-velopments. For major diseases for which con-trol mea?ures are inadequate. research is an inex-pensive approach on the basis of cost per infect-ed person per year. (N Engl J Med 301:967-974,1979)

T HE three billion people of the less developedworld suffer from a plethora of infectious dis-

.,eases. Because these infections tend to nourish at thepoverty. level, they are important indicators of a vast~ate of collectiv,= ill health. The concomitant dis-'ability has an adverse effect on agricultural and in-dustrial development, and the infant and child mor-tality inhibits attempts to control population growth.

What can be done to help alleviate a nearly un-broken cycle of exposure, disability and death? The~t solution, of cQurse, is comprehensive primaryhealth care, defined at the World Health Organiza-tion conference held at Alma Ata in 1978 as

Absolute poverty is a condition of life so characterized bymalnutrition, illiteracy. disease, high infant mortality and low lifeexp~ctancy as to b~ beneath any reasonable definition of human

decency'

How then, in an age of diminishing resources, canthe health and well-being of those "trapped at the bot-tom of the scale" be improved before the year 2OOO? Avalid approach to this overwhelming problem can bebased on the realization that the state of collective illhealth in many of the less developed countries is not asingle problem. Traditional indicators, such as infantmortality or life expectancy, do not permit a grasp ofthe issues involved, since they are actually compositesof many different health problems and disorders.Each of the many diseases endemic to the lessdeveloped countries (Table 1) has its own uniquecause and its own complex societal and scientificfacets; there may be several points in the process forwhich interventions could be considered.'-s

Thus, a rationally conceived, best-data-based,selective attack on the most severe public-healthproblems facing a region might maximize improve-ment of health and medical care in less developedcountries. In the discussion that follows, we try toshow the rationale and need for instituting selectiveprimary health care directed at preventing or treatingthe few diseases that are responsible for the greatestmortality and morbidity in less developed areas andfor which interventions of proved efficacy exist.

the attainment by all peoples of the world by the year 2000 of a~l of health that will pcnnit them to lead a socially and

.economically productive life. Primary health ~ includes atlcast: education concerning prevailing health problems and themethods of preventing and controlling them; promotion of foodsupply and proper nutrition, an adequate supply of safe water&nd basic sanitation; maternal and child health ca~, includingfamily planning; immunization against the major infectious dis-

c cases; prevention and control of 1~lIy endemic diseases; ap-, propriate t~tmcnt of common diseases and injuries; and provi-

sion of essential drugs.'

The goal set at Alma Ata is above reproach, yet itsvery scope makes it unattainable because of the costand numbers of trained personnel required. Indeed,the World Bank has estimated that it would cost bil-lions of dollars to provide minimal, basic (not com-prthtnsive) health services by the year 2000 to all thepoor in developing countries. Tht bank 's pr~sident,Robert McNamara, offered this somber prognosis in~ annual report in 1978:

ESTABLISilINC PRIORITIES FOR HEALTH CARE

Faced with the vast number of health problems ofmankind, one immediately becomes aware that all ofthem cannot be attacked simultaneously. In manyregions priorities for instituting control measuresmust be assigned, and measures that use the limitedhuman and financial resources available most effec-tively and efficiently must be chosen. Health planningfor the developing world thus requires two essentialsteps: selection of diseases for control and evaluationof different levels of medical intervention from themost comprehensive to the most selective.

Even if the projected -and optimistic -growth rates in thedeveloping world arc achieved, some 600 million individuals atthe end of the century will remain trapped in absolute poverty

0

~

~ From Ibe Rockefeller Foundation, 1133 Avenue of Ihe Americas, New'!rOfk. NY 10036, whcre reprint requests may bc addressed 10 Dr Warren.~ Pracnlcd al a meeting on Heallh and Population in Dcveloping

lrics, cosponsored by Ihe Ford Foundalion,lhe International Dcvelop-~.1 Research Center and the Rockefeller Foundation and held at Ihe Bel-~ Study and Confcrencc Center. lake Como. Iiaiv. April. 1979

.Noy I, 197~ ~THE NEW ENGLAND JOURNAL OF MEDICINE968

Selecting Diseases for Control long-~erm improvem~nts i~ sanitary and agriculturalpractices to reduce reinfection. In View of the difficultyof eliminating exposure to the roundworm and the lowmorbidity associated with the infection, ascariasisdeserves less attention than its ubiquity seems to sug-

gest.Malaria is associated with a far smaller mortality

rate than that of Lassa fever and a far lowerprevalence that that of ascariasis. Yet its mode oftransmission is well known, and it produces muchrecurring illness and death; about one milliQnchildren in Africa alone die annually from malaria.'What also distinguishes malaria from Lassa fever andascariasis is that it can be controlled through regularmosquito-spraying programs or chemoprophylaxis.'.'Of these three infections, then, malaria would be as-signed the highest priority for prevention in the mOSteffective approach to reducing morbidity and mor-

tality.By means of the process outlined above for Lassa

fever, ascariasis and malaria, the major infectionsendemic to the developing world (Table 1) wC!"tevaluated and assigned high (I), medium (II) or low(III) priorities. Within categories exact rank is not dmajor importance, and rank may change or items maybe added or deleted, depending on the geographicarea under consideration. For instance, schistosomia-sis, to which a high priority was assigned, does notoccur in many areas of the developing world. Our ~

In selecting the health problems that should receivethe highest priorities for prevention and treatment,four factors should be assessed for each disease:prevalence, morbidity, mortality and feasibility of

control (including efficacy and cost).Table 2 illcorporates these factors into an analysis

of three representative illnesses of the less developedworld. The newly discovered Lassa fever was as-sociated with a 30 to 66 per cent mortality rate in thefew limited outbreaks recorded in Nigeria, Liberiaand Sierra Leone. Those who survived recovered fullyafter an illness lasting seven to 21 days. Although thisfatality rate seems to suggest giving Lassa fever highpriority in a major health program, the prevalence ofovert disease appears to be low. Furthermore, the onlytreatment available is injections of serum frompatients who have recovered. Since its mode of trans-mission is unknown and there is no vaccine, Lassafever is impossible to control at present.' Therefore,concentration on preventing Lassa fever would be

neither efficient nor efficacious.Ascaris, the giant intestinal roundworm, causes the

most prevalent infection of man, with one billion casesthroughout the world.7 Yet disability appears to beminor and death relatively rare.)" Treatment. howev-er, requires periodic chemotherapy for an indefiniteperiod.).'" Control may ultimately require massive,

Table 1. Prevalence. Mortality and Morbidity of the Major Infectious Diseases of Africa. Asia and Latin America. 1977-1978..

0 (THO~"'OS Of

c..ES/Ya)

Diarrheas 3-5,CXXJ,CXXJ

Respiratory inrcctionsMalaria 1 SO,CXXJMeasles 80,CXXJSchist~omiasis 20,CXXJWhooping cough 2O,CXXJTuberculosis 7CXXJ 3NeonatalteLanus 120-180 IDiphtheria 700-900 3Hookworm 1500 4South American trypan- 1200 2

osomiasisOnchocerciasis

Skin disease ':CJ.~:- ---3

3O,CXXJRiver blindness 1-2

Meningitis I SO 1Amebiasis 400,CXXJ 3Ascariasis 800,OOO-I,CXXJ,OOO 3Poliomyelitis 80,000 2Typhoid 1000 2leishmaniasis 12,000 I 3Arrican trypanosomiasis 1000 Ileprosy 2-:Trichuriasis 500,000 ,Filariasis 2SO,OOO 3Giardiasis 200,000 JDengue 3~ 2

Malnutrition 5-800,000-

°S.oed on c",mol« from .he World Hc.ilh Orl.no.."on .nd it, Spc.-,.1 PrOlr...mc lor Rcscorch .nd fr.,n,nl on fropoc.. 0, confirmed or modifIed byc.,r.pot..from publ.'hcd cp,dcm,olo"c ..ud... perlormcd ;n well dcr.ncd popul."on, (!CO ,clc,cnc«) f'IU'« du nol .1.." m..ch ,h.". ofT.:i.lly rC\'Ortcd. b..: ."'"rrcpo"'"

,rc..'I dcno,c, bcdr,ddcn. 2 .blc 10 I.nc"on on own '° "'.., c"C.' ..,mo.t.,or) " ..m,no,

D.."'.(THOVS"'OS/V,)

A VE""'E ~O. OfD.n Of LJfE losT

(PE. ~E)

3-S5-73-S

I~I~6(»-1 (XX)

21-28200-400

7-107-10

100600

RounO£PtaSOMAl.

Dt...aIUTTf

22-322

3-42

IN'ECTION l"nc-no1'S(THO"..,,"osfYa)

5-10,0004-SOOO

1200900

500-1 0002s0..450

400100-150

~S(}-6()

60

!-5,OOO,000

800,00085,000

200,00070,000

1,000.000120-18040,000

7-900,00012,000

Low

20-50303020

10-2025

5

2 -)\XX) 3M

J<XX!7-107-107-10

J<XX!+14-28

100-200ISO

5OQ-J<XX!7-10

10005-75-7

200-500ISO

1500

I<XX>2<XX>500

12. <XX>10

12.<XX>100

2- )<XX)

5001-2<XX>

Very lowLowLow

Very low0.1

2(XXJ

Vol 301 No 18 DISEASE CONTROL IN DEVELOPING COUNTRIES -WALSH AND WARREN969

Table 2. An Approach to the Establishment of Priorities for Disease Control, Based on Prevalence, Mortality, Morbidity andFeasibility of Control of Three Representative Infections.

Po.,V4UNCOcno.. MOOT.un Mo'.,Oln FE.SI8tUTVOf CONTaOL

Extremely poorat present

~IOOITY

La... rever Unknown(lhoughllobelow)

Exlrcmcly high(lhoughllo .ffectI billion pcoplc)

High (3(}..66%) Moderate

(bedridden7-21 days)

Low (minor dIs-ability &.orten asymp-tomatic)

High (severe,many compli-~Iions, ortenrecurrcnt)

Low: prcvllCDcelow, fcasibilityof coDlrol poor

Low: morulily&I. morbidilYlow, fcasibililyof CODlrol poor

High: prcvalcncehigh, morbidityhigh. fcasibililyof CODlrol load

Extrcmcly low

(approximately0.001%)

Poor (conlinuousdrug lrcalmen!required)

Malaria High (morc than300 million in-fcctcd annually)

Low (approxi-matelyO.I%)

Good (chcmoprophy-laxis avaibble;

regular sprayingprograms (orvectors pratticol)

A medium or low priority was assigned if controlmeasures were inadequate. For example, there is noacceptable therapy for chronic Chagas' disease.}"Only toxic drugs and procedures of unknown efficacy,such as nodulectomy, are available for treatment ofonchocerciasis.}" Leprosy and tuberculosis requireyears of drug therapy and even longer follow-up

Table 3. Priorities for Disease Control in the DevelopingWorld. Based on Prevalence. Mortality. Morbidity and Feasi-

bility of Control.

PR'ORrnGROUP

REASONS ro. ASS'GN"E'" TO THIS CARGO.T

High prcvalence, high mortalily or highmorbidilY, clTcx:tivc control

suits and rationale for the proposed hierarchy are list-ed in Table 3.

Group I contains the infections causing the greatest'amount of most readily preventable illness and death:~diarrheal diseases, malaria, measles, whooping cough,!schistosomiasis and neonatal tetanus. With the excep-tion of schistosomiasis, all the infections receiving~~hest priority for health-care planning affect young:hildren more than adults. 10-14 Together with respira-tory infections and malnutrition, they account for

of the morbidity and mortality among infantsand young children.II.ls-11 Members of this age group(five years old or less) have a death rate many timesgreater than that of their counterparts in Westerncountries -accounting for 40 to 60 per cent of allmortality in most less dc:veloped countries.II,J1-1' If in-fant and child deaths from these infections arereduced, a large declinc: in the overall death rate will:result. Such a situation would bc: an optimal outcome

-a selective disease-control program.Groups II and III contain health problems that are

~ither less important or more difficult to control. Res-Jlratory infections, a major cause of disability and~th, are not listed in Group I because of the dif-

[Iculties involved in preventing and managing them. AMae variety of viruses and bacteria are associatedwith pulmonary infections, and no specific causativeigent has been found in most patients.I,,20 As in thelI1dustrialized world, where pneumonia is frequentlyne tenninal episode in elderly patients weakened by;ancer or cardiova~lar disease, lower-respiratory-:Tact infections affect children in developing countriesWho are already afflicted with chronic malnutritionInd parasitic infections. I' Pneumococcal and in-luenza vaccines prevent only a small percentage of:ases, and influenza immunization must be givenlmost yearly because the virus changes antigenical-y. When penicillin injections were given to allhildren with clinical signs of pneumonia in the

~,arangwal Project in India, the mortality rate~~reased by SO per cent,21 but this method must be~aJuate? ~ore extensivel.y before it. can be re~arded!~~ ffiaJor Improvement In prevention of respiratory~sc:.

I High

DiarThcal diseascsMcaslcsMalaria

Whooping coughSchistosomiasisNconalaltCUnus

I( MediumRespir2tory infections

Poliomyelitis

Tuberculosis

Onchocerciasis

Meningitis

Typhoid

Hookwonn

Malnutrition

H~h prevalence, high monalilY, noe/f..:tive control

Higb prevakn<X, low monality,e/f..:tive control

High preva1coce, high mortality,<:antra! difrocult

Medium prnalen<x, high morbidity, lowmortalitY, control difrlCUll

Medium prnaJen<x, higb mortality,control difficull

Medium prnalen<x, high monality,conlrol difficult

High prnalen<x,low mortality,conlrol difficult

High prevalen<x, high morbidity,control complex

r

r

III lowSouth American trypanoso-

miasis (Chagas' disease)Arrican trypanosomiasis

leprosyAscariasis

DiphtheriaAmebiasisleishmaniasisGiardiasisFilariasis

Dengue

Control difficult

low prevalcncc. control difficultContro! difficultlow mortality, low morbidity,

rontrol difficultLow mortality, low morbidityControl difficultControl difficultControl difficultControl difficultControl difficult

THE NEW ENGLAND JOURNAL OF MEDICINE970 Nov. 1979

pcriods to cnsure cure.4.22.23 In~tcad of attempting im-mcdiate, large-scale trcatment programs for these in-fections, the most efficient approach may be to investin rcsearch and developmcnt of lcss costly and morcefficacious means of prevention and thcrapy. Toreiterate, the most important factor in e~tablishingpriorities for endemic infections, even when evaluatingdiscases with high case rates, is a knowlcdge of whichdiseases contribute most to the burden of illness in anarea and which arc rcasonably controllablc.

poor in developing countries by the year 2000 will b6$5.4 to $9.3 billion (in 1975 prices).26 This investmentjwhich includes only initial capital investment andtraining costs, would provide one community healt

~worker or aux.iliary nurse-midwife for every 1500 t2000 people and one health facility for every 8000 t12,000 people or every 10 km2, whichever is g~ater. In!the model area in Africa, the World Bank estimateulthat supplying the minimum ca~ offe~d by buildi~one health post with one vehicle per 10,000 people andtrain.ing 125 auxiliary nurse-midwives and 250 comlmumty health workers would cost $2,500,000, or $Sper capita. To this figure must be added the recurren~costs of salaries, drugs, supplies and maintenance;Other costs not included are for training facilities;continuing education, expansion of referral servicesand development of communication, transportationand administrative networks to supply and managethe health facilities. Furthermo~, the effectiveness ofthis model program for averting deaths or applyingsuch preventive measures as education in sanitationand nutrition has not been clearly established.

The pilot projects for providing basic health-careservices that have been evaluated vary in their effec-,tiveness in improving the general level of health care.!For example, an outside evaluation of primary health!service in Ghana revealed that a third to half the pop.4ulation of the districts lived outside the effective reach!of health units providing primary care. Only abou~one fifth of the births were supervised by traine~midwives; only one fifth of the children under the a

~ ., of five years had been seen in a child.,health clinic, an .

two thirds of the population lacked environmentsanitation services. Furthermore, the services weoften of poor quality, notably in the crucial area oBchild care.21.2. i

The cost and effectiveness of several experiment~programs providing primary health care in localize

areas have been compared in Imesi, Nigeria2'Etimesgut, Turkey]O,)I; Narangwal, India21;jamkhedJIndian.)); Guatemalan villages"; Hanover, jamai~ca)5-)1; and Kavar, Iran.» The estimated cost pedcapita varied widely among the programs, partiCUl

~ly because they were initiated at different times ovethe past 15 to 20 years and furnished different servicto their communities. In general, however, the cosper capita ranged between 1 and 2 per cent of thCJnational per capita income of the particular countrvJThe cost for infant deaths averted were difficult rcompare because of the paucity of control groups an"inconsistency of the population groups monitored~Figures ranged from $144 to $20,000, with a media

~1

of 1700. The only precise calculations for the costs peinfant death averted ($ 144) or child death averte(1988 per one to three-year-old child) were formedical-care and nutrition-supplementation projec~in Narangwal, India.21 The estimates were muc

~higher for deaths averted by nutrition supplements.Under some circumstances, programs of basi

primary health care have been successful, but the cos

EVALUATING AND SELEcTING MEDICAL INTERVENTIONS

Once diseases al'e selected for prevention and treat-ment, the next step is to devise intervention programsof reasonable cost and practicability. The interven-tions relevant to the world's developing areas that areconsidered below are comprehensive primary healthcare (which includes general development as well asall systems of disease control), basic primary healthcare, multiple disease-control measures (e.g., insec-ticides, water supplies), selective primary health care,and research. Below is a discussion of each approach,with emphasis on the relative cost involved in undel'-taking and maintaining these programs and on thebendits that have accrued.

This section of our analysis relies on reportedresults from individual studies conducted in variousparts of the world. In addition, we have examined es-timates of cost and effectiveness in terms of expecteddeaths averted by each intervention for a model areain Africa. The model area is an agricultural, rural por-tion of Sub-Saharan tropical Africa with a populationof about 500,000 (100,000 are five years old or less).For reference purposes, the average figures for Sub-Saharan Africa will be used: the birth rate is 46 perthousand total population, the crude de3th rate 19 perthousand total population, and the infant mortalityrate 147 per thousand live births.24.%S

Comprehensive versus Basic Primary Health Care

Comprehensive primary health care for everyone isthe best available means of conquering global dis-ease, the humane and noble goal declared at AlmaAta. As defined by the World Health Organization,this system encompasses development of all segmentsof the economy, ready and universal access to curativecare, prevention of endemic disease, proper sanitationand safe water supplies, immunization, nutrition,health education, maternal and child care and familyplanning. Since resources available for health pro-grams are usually limited, the provision of compre-hensive primary health care to everyone in the near fu-ture remains unlikely.

Basic'primary health-care systems are far more cir-cumscribed in their goals, which are to provide healthworkers and establish clinics for treating all illnesseswithin a population. Nevertheless, this approach is farfrom inexpensive. The World Bank has estimated thattht: cost of furnishing basic health services to all the

Vol 301 No. 18 DISEASE CONTROL IN DEVELOPING COUNTRIES -WALSH AND WARREN

971

about $3.70 and good r~sults hav~ b~en r~port~d: th~preval~nc~ of th~ inf~ction has d~cr~as~d from 45 to 35p~r c~nt in adults and from 21 to 4 p~r c~nt inchildr~n. D~spit~ th~s~ h~art~ning figur~s, ~radicationof th~ v~ctor cannot b~ consid~r~d on th~ horizon.Schistosomiasis is a long-t~rm, chronic inf~ction andth~ d~ath rate will not b~gin to d~clin~ until manyy~ars aft~r continuous mollusk control.

and the degree of improvement in community healthhave varied markedly enough that refinements in theapproach are still needed.

Multiple Dlsease-Control Measures

These interventions, which include vector control,water and sanitation programs and nutrition sup-plementation, are more specific and easily managedthan primary health-care programs, and they controlmany similarly transmitted diseases simultaneously.They can decrease mortality and morbidity and haveserved as interim strategies for health care in lessdeveloped countries.

Vector Control

Vcctor control is directcd at managing thc insectsand mollusks that carry human disease. 111is ap-proach has thc advantagc of bcing comparativcly in-cxpcnsive, but it must be continucd indcfinitcly andmay be cphemcral sincc thc vcctors tcnd to bccomercsistant. Thc cxamplcs bclow rcvcal somc of thc com-plcxities of maintaining vector control.

Thc control of malaria transmission through inscc-,ticides has becn highly cffectivc. In thc tropical'regions and savannas of Africa, twicc-ycarly spraying~ dccrcased thc crude dcath ratc by approximatcly~ pcr ccnt and infant mortality by 50 pcr ccnt.)"""I11c World Hcalth Organization has cstimatcd thatthc avcrage cost for housc-to-house spraying withchlorophenothanc (DDT) is $2 pcr capita annual-ly.' Thcrcforc, thc cost pcr adult and infant dcathavcrtcd is $250, and the cost pcr infant dcathavcrtcd is $600. Unfortunatcly, cradication of malariawith insccticides is bccoming morc difficult to ac-complish. Bccausc mosquitocs can bc cxpcctcd tobccomc rcsistant to DDT within a few years, othcr,:much more expcnsivc pcsticidcs must bc substituted;ithc usc of propoxur or fcnctrithion will raisc the cost:Of thc chemicals fivc to 10 tim~s.' Furthcrmorc, th~rcis no way of knowing how long thcs~ ins~cticidcs will:remain 'toxic to thc mosquitocs. Among th~ mos-:quitocs in which widcsprcad rcsistance to insecticidcs~ dcvclop~d arc Culex pipims fatigans, th~ major yec-~or of urban filariasis, and Aedes aegypti, thc vector ofyellow fev~r and d~ngue.s

Two othcr vector-control programs illustrate thcDrolong~d maintenance required by this type of hcalthmtcrvcntion. Onchocerciasis, a potentially blindingilelminth infection affecting 30 million people in\Inca, is bcing managed in the Volta River Basin!Tough a 20-y~ar larvicide operation to control theIlackfly vcctor. The program is cstimated to cost $18~r capita for thc cntire 20-year period or $.90 pcrapita pcr year.2' Disability will be prcvcnted, andconomic activity in the arca may increase if theIrogram is successful, but continuous, indefiniteIpplications of insecticide will be nec~ssary. Since965, St. Lucia has had a program to control theuail-transmitted helminth infcction schistosomiasisLrough molluscicides. An annual cost p~r capita of

Water and Sanitation Programs

Proper sanitation and clean water make a substan-tial difference in the amount of disease in an area, butthe financial investment involved is enormous. Thesuccess of such projects also depends on rigorousmaintenance and alteration of engrained culturalhabits.

With the installation of community water suppliesand sanitation in developing areas, deaths fromtyphoid can be expected to decrease 60 to 80 percent,') deaths from cholera 0 to 70 per cent,')'" fromother diarrheas 0 to 5 per cent,"-SI from ascaris andother intestinal helminths 0 to 50 per cent'.10.S2-S. andfrom schistosomiasis 50 per cent'2.S2 (after 15 to 20years). The World Bank has estimated that the cost ofproviding community water supplies and sanitation toall those in need by the year 2000 will be $135 to $260billion.26.ss Construction of a rural community stand-pipe costs 120 to 126 per capita, and rural sanitationcosts $4 to $5 per capita. In urban areas the costs are$31 and $23, respectively. In our model area of Sub-Saharan Africa the initial investment would be $12 toS 15 million. If amortization and annual maintenancecosts are only 10 per cent of this sum, the annual costper deaths averted will be $2400 to $2900, and the costper infant and child deaths averted will be $3600 to14300.

What must be realized is that the above sums arelargely for public standpipes, which are not highly ef-fective in reducing morbidity and mortality fromwater-related diseases. It is well documented that con-nections inside the house are necessary to encouragethe hygienic use of water. so For example, shigella-

caused diarrheas decreased 5 per cent with outsidehouse connections but fell 50 per cent when sanitationand washing facilities were available within thehome.sl

All these estimates depend on exclusive use ofprotectcd sanitation and water supplies, without con-tinuing use of environmental sources. In Bangladesh,for example, there was no reduction in cholera inareas supplied with tube wells, primarily because ofthe use of contaminated surface water as well as theprotected water supply." In St. Lucia, contact withsurface water could not be discouraged until house-hold water supplies and then swimming pools andlaundry units were installed, and an intensive health-education campaign was instituted.'2 In other words,changing peoples' habits in excretion and water usagetakes more than introducing an adequate, dependableand convenient new source. Realistically speaking, a

THE NEW ENGLAND JOURNAL Of MEDICINE972 Nov. I. 1979.

pervasive and effective health-education campaigns"s,is required.

Nutrition Supplementation

Nutrition programs have been advocated as amongthe most efficient means of decreasing morbidity andmortality in children, but supplementation alone hashad no notable effect. Malnutrition is an underlyingor associated factor in many deaths from infections inchildren; in a group of Latin American children, itwas associated in 50 per cent of the cases.s' Poor nutri-tion may also increase susceptibility to disease orpredispose an infccted child to more sevcre illness.60-42Conversely, infection may be a prominent cause ofpoor nutrition".')'" since less food is ingested and ab-sorbed by a sick child. Therefore, if infections could becontrolled it is probable that the nutritional status ofchildren would improve greatly. There have beensome situations, however, in which malnutrition hasbeen reported to protect against certain infections,e.g., the Sahel famine was thought to suppressmalaria, and iron deficiency was reported to protectagainst bacterial infections."-'o

In view of these findings, it is not surprising that fewnutrition-supplementation programs alone have ef-fected a major decrease in the death rate. TheNarangwal Project is one of these few, but even in thatprogram the cost per death averted in infants was1213. In children one to three years old the cost was$3000 -1.5 to three times higher than the cost ofmedical care alonc.21

Selective Primary Health Care

The selective approach to controlling endemic dis-ease in the developing countries is potentially the mostcost-effective 'type of medical intervention. On thebasis of high morbidity and mortality and offeasibility of control, a circumscribed number of dis-eases are selected for prevention in a clearly definedpopulation. Since few programs based on this selectivemodel of prevention and treatment have been at-tempted, the following approach is proposed. Theprincipal recipients of care would be children up tothree years old and women in the childbearing years.The care provided would be measles and diph-theria-pertussis-tetanus (DPT) vaccination for chil-dren over six months old, tetanus toxoid to all womenof childbearing age, encouragement of long-termbreast feeding, provision of chloroquine for episodes offever in children under three years old in areas wheremalaria is prevalent and, finally, oral rehydrationpackets and instruction.

If even 50 per cent of the children and their mothersand 50 per cent of the pregnant women in a com-munity were contacted, deaths from measles would beexpected to decrease at least 50 per cent,',.'2 deathsfrom whooping cough 30 per cent, ') from neonataltetanus 45 per cent, '4 from diarrhea 25 to 30 percent's." and from malaria 25 per cent.' Oral rehydra-tion has been used successfully in hospitals,"." in out-

patient clinics" and recently in the home's." to treatdiarrheas of numerous causes.

These services could be provided by fixed units orby mobile teams visiting once every four to six mont~in areas where resources were more limited. Mobil~units have been successfully used in immunizationprograms for smallpox and measles,la.11 in treatmentservices directed against African trypanosomiasis and;meningitis"2 and in provision of child care in rural!areas.'}-8S

The cost of fixed units would be similar to that 0\;basic primary health care, although efficienc;:y ShoUld

]be much greater. Cost estimates for a mobile healthunit used in the model area in Africa for malaria con.1trol and water and sanitation programs were based OnJan extensive study of the Botswana health services byGish and Walker."s They estimated $1.26 as the cost

!per patient contact in 1974, on a sample 306-km tripthat reached 753 patients; the estimated cost per in.'fant and child death averted was $200 to $250.1Medications accounted for 30 to 50 per cent of th,cost, but this figure could be decreased with contrib~jtions of drugs from abroad or their manufacture!within the country. j

Whether the system is fixed or mobile, flexibility is;necessary. The care package can be modified at any;time according to the patterns of mortality and mor-:bidity in the area served. Chemotherapy for intestinal;helminths, treatment of schistosomiasis and sup-plementation with new vaccines or treatments as they!become available are all types of selective primaryhealth care that could be added or subtracted to t~core of basic preventive care. It is important, however,~for the service to concentrate on a minimum number,of severe problems that affect iarge numbers of people;and for which interventions of established efficacy canbe provided at low cost.

Research

For a number of prevalent infections, treatment orpreventive measures are expensive, difficult to ad-minister, toxic or ineffective. These infections, whichinclude Chagas' disease, African trypanosomiasis,leprosy and tuberculosis, may better be dealt withthrough an investment in research. In terms of thepotential benefits, the cost of research is low. Indeed,the total amount now being spent on research in alltropical diseases is approximately $60 million, ex-ceedingly small in relation to the number of people in-fected. As Table 4 shows, expenditures for research onsome of the major diseases in the developing worldhave by far the lowest per-capita cost of all medica! in-terventions discussed."

The estimated cost for the research and develop-ment leading to the pneumococcal vaccine licensed inthe United States in 1978 was $3 to $4 million(Austrian R: personal communication). Death anddisability in developing countries would be reducedby heat-stable vaccines for measles, malaria, leprosyand rota virus and Eschenchia coli-induced diarrheas.

DISEASE CONTROL IN DEVELOPING COUNTRIESVol. 301 No. 18 WALSH AND WARREN

adult population of the area covered by the service. Asthe table suggests. selective primary health care maybe a cost-effective interim intervention for many lessdeveloped areas.

by improved chemotherapy for leprosy, tuberculosis,American and African trypanosomiasis, onchocercia-sis and filariasis and by depot drugs for malaria andintestinal helminths.

CONCLUSIONS

Until comprehensive primary health care can bemade available to all, services aimed at the few mostimportant diseases (selective primary health care)may be the most effective means of improving thehealth of the greatest number of people. The crucial:point is how to measure the effectiveness of medical'interventions. In all the foregoing calculations, we~ ased ?ur analysis of cost effective,ness on changes in

ortallty or deaths averted. We did not measure the

ness and disability that would be prevented. Nother benefits for which intervention may have been

:responsible were measured because they are muchimore difficult to quantify. The inadequacy of avail-pble data makes it impossible to measure distinct and!undeniable secondary benefits. For example, water:supplies close by would save time for the women whoprry water. and increased amounts could irrigate afome garden.r Accordingly, Table 5 summarizes the estimated~ ts per capita and per death averted for the various

ealth interventions considered. The per capita costs" calculated in tenns of the entire infant, child and

l-

f-5. Estimated Annual Costs of Different Systems of

Health Intervention.

EMT1ON PIa CAPITA COST PE. INFANT ANO/O.COST (S) CHIlD DEAn. AvE.no. (S}

040-7.50 144-20,<XX> (I)

2.00 700

2.00 600(1)

090 Few inr.nl 81. child

dcalhs3.70 Few jnr.nl 81. child

dcalhs3600-4300 (I,C)

213(1))(XX) (C)

200-250 (I,C)--tDch...cd by ..11... health work.r.

30-54

175

025

is-

~ primary health caret~~g.Median

~osquilo control (or malaria~ocerciasis control program

ioUulk control (or-..schistosomiasis;GImunilY water supplies &I.., Ianltatlonfnngwal nutrition

~ supplementation

~'C primary health caret-

;nr.nl " C dlildCUe, dcl;vcn:d by mob,l. unO'5

REFERENCES

I. World Health Organization: Declaration or Alma Ata (Repon onthe International Conrerence on Primary Health Care, Alma Ata,USSR, September 6-12, 1978). Geneva, World Health Organiution.1978

2 McNamara RS: Address to the Board or Governon orthe World Bank.Washington, DC, World Bank, 1978

3. Geographic Medicine ror the Practitioner: Algorithms in the diagnosisand management or exotic diseases. Edited by KS Warren, AAFMahmoud. Chicago, Univenity or Chicago Press. 1978

~. Tropical Medicine. Edited by GW Hunter III, JC Swanzwelder, DFClyde. Firth edition. Philadelphia, WB Saunden Company, 1976

5 Resistance or Vecton and reservoirs or disease to pesticides: twenty-sec-ond repon or the WHO Expen Committee on Insecticides. WHO TechRep Ser 585:1-88, 1976

6. Yirallnrections or Humans: Epidemiology and control. Edited by ASEvans. New York, Plenum Medical Book Company, 1976

7. Peters W: Medical aspecu -comments and discussion II, The Rele-vance or Parasitology to Human Welrare Today (Symposia or theBritish Society ror Parasitology. Yol 16). Edited by ERA Taylor, R

Muller. Oxrord. Blackwell Scientific Publications, 1978, pp 2~18. Arfaa F, S.hba GH, Farahmandian I: Eovaluation or the effect of dif-

ferent methods of control of soil-transmitted helminths in Khuzestan,southwest Iran. Am J Trop Med HY8 26:2J{}.233, 1971

9 WHO Expen Committee on Malaria: sixteenth report. WHO Tech Rep-Ser 549:1-89, 197~

10. Preston SH, Keyfiu N, Schoen R: Causes or D'tath: Ljrc tables rornational populations. New York, Seminar Press, 1972

II. Wyon JB, Gordon JE: The Khana Study: Population problems in therural Puojab. Cambridge, M chusetts, Harvard University Press,1971

12. Ongom YL, Bradley DJ The epidemiology and consequences orSc/rutosoma "'ansoni inrection in West Nile, Uganda. I. Field studies ora community at Panyogoro. Trans R Soc Trop Med Hyg 66835-851,1972

13 Farooq M, Samoan SA. Nielsen T Assessment or severity or diseasecaused by Schurosoma hat_tobium and S. ",anso,,; in the Egypt-49project area. Bull WHO 35:389-404, 1966

I~. Siongok TKA, Mahmoud AAF, Ouma JH, et al: Morbidity inSchutosomiasis "'an.rOlli in relation to intensity or inrection: study of acommunity in Machakos. Kcnya. Am J Trop Med Hyg 25:273-284,1916

15. Hull TH, Rohde JE: Prospecu for Rapid Decline or MonaJity Rates inJava: A study or causes or death and tbe reasibility or poliey interven-tions ror monality control. Yogyakarta, Indonesia, Population In-stitute, Gadjah Mada Univcrsity, 1978

16. Bulla A, Hitze KL Acute respiratory inrections: a review. Bull WHO56:~81-498, 1978

17. Dyson T: Levels, t~nd.r, differentials and causes of child mortality -asurvey. World Health Stat Rep 30:282-311,1977

18. Preston SH: Mortality Patterns in National Populations: With specialrererence to recorded causes of death. New York, Academic Press. 1976

19 United Nations Demographic Yearbook 197~. New York, United Na-tions, 1975

20. Sobeslavsky 0, Sebikari SRK, Harland PSEG, et al: The viral etiologyor acute respiratory infections in children in Uganda. Bull WHO55:625-631, 1977

21. Taylor CEo Kielmann AA, Parker RL. et al: Malnutrition, Inrection,Growth and Development: The Narangwal experience: final reportWashington, DC, World Bank, 1978

22. Fox W, Mitchison DA Shon course chemotherapy ror pulmonary tu-berculosis. Am Rev Respir Dis II 1:8~5-848; 329-352, 1975

23. WHO Expert Committee on Leprosy: firth report. WHO Tech Rep Ser601: 1-48, 1977

2~ Kane TT, Myers PF: 1918 World Population Data Sheet. Washington,DC, Population Reference Bureau, 1918

25 United Nations Demographic Yearbook 1976. Geneva, World HealthOrganization. 1977

26. Burki S1, Voorhoevc 1JC, Laylon R, et al: Global Estimates rorMeeting Basic Needs: Background paper (Basic Needs Paper No. I)Washington, DC. World Bank, 1977

27. Instil ute of Development Studies Research Repon.r: Health Needs andHealth Services in Rural Ghana Brighton, England, University orsus-sex, 1978

28 Primary care in Ghana Lancet 21085, 1978

r

.

HE NEW ENGLAND JOURNAL OF MEDICINE Noy 1979

29. Cunningham NJ The under fives clinic -what difference docs il make.J Trop Pediatr (in press)

30 Fisek NH: An Account of the Activities of the Etimesgut Rural HealthDistrict 1967, 1968, and 1969 Ankara, Hacettepe Press and HacettepeUnivenity School of Medicine, Institute of Communily Medicine,1970

31. Id_m An Account of the Activities of the Etimcsgut Rural Health Dis.trict 1970-1974 Ankara, Ayyildiz Matbaasi and Hacettepe UniversitySchool of Medicine, Institute of Community Medicine, 1975

32. Arole M, Arole R: A comprchensive rural health project in Jamkhed(India). Health by the People. Edited by KW Newell. Geneva, WorldHealth Organizalion, 1975, pp 70-90

33 Gwatkin DR, Wilcox JR, Wray JD: Can Intervention Make a Dif.ference?: The policy implications of field experiment experience: areport to lhe World Bank. Washington, DC, World Bank, 1978

34 Working Group on Rural Medical Care: Delivery of primary care bymedical auxiliaries: lcchniques of use and analysis of benefits achievedin some rural villages in Guatemala, Medical Auxiliaries: Pr~dingsofa symposium held during the lwclflh mccling oflhe PAHO AdvisoryCommiltcc on Medical Research, June 25, 1973. Washington, DC, PanAmerican Hcallh Organizalion, 1973, pp 24-40

35 Alderman MH, Husled J, uvy B, et AI: A young-child nulrilionprogramme in rural Jamaica lancel 1:1166-1169, 1973

36 Aldcrman MH, Cadien DS, Haughton PBH, et al: A sludent ruralhcallh project in Jamaica. Wesl Indian Mcd J 21(1):20-24, 1972

37. Alderman MH, Wise PH, Ferguson RP, el al: Reduction of young childmalnulrition and mortality in rural Jamaica. J Trop Pedialr 24:7-11,1978

38. Ronaghy HA: Kavar villagc hcalth worker projcct. J Trop Pedialr2413-60, 1978

39. Kouznetsov Rl: Malaria conlrol by applicalion of indoor spraying ofresidual insecticides In lropical Africa and its impact on communilyhcallh. Trop DOCI 7:81-91, 1977

40. Paync D, Grab B, Fontaine RE, el II: Impact of conlrol measures onmalaria lransmission and general mortality. Bull WHO 54:369-377,1976

4 I. Fonlaine RE, Pull JH, Payne D, el al: Evalualion of fenitrithion for lhcconlrol of malaria. Bull WHO 56:445-452, 1978

42. Jordan P: Schislosomiasis -research 10 conlrol Am J Trop Med Hyg26:877-886,1977

43 Zahccr M, Prasad BG, Govil KK, cl al: A nolo on urban waler supplyin Uttar Pradesh. J Indian Med Assoc 38:17-82, 1962

44 Azurin JC, Alvcro M: Field evaluation of environmcntal sanitationmcasures againsl cholera. Bull WHO 51:19-26, 1974

45. Wolff Hl, Van Zijl WJ: Houseflies, the availability of wIler, and diar-rhoeal dj,ease Bull WHO 41:952-959, 1969

46. Briscoe J: Thc' role of water supply in improvin8 hcalth in poorcounlries (with special refcrcncc 10 Bangladesh). Am J Clin Nl1Ir31:2100-2113,1978

47. Sommer A, Woodward WE: Thc influence of prolected walcr supplieson the sprcad of ciassical-inaba and EI Tor-Ogawa cholera in EastBcngal. Lancet 2:985-987, 1972

48. levinc RJ, Khan MR, D'Souza S, ct al: Failure of sanilary wells toprotect againsl cholcra and olhcr diarrhoeas in Bangladcsh. Lancet2:86-89, 1976

49 Schneider RE, ShifTman M, Faigenblum J: The potential cfTect of watcron gastrointestinal infections prevalcnl in developing oountries. Am JClin Nutr 312089-2099, 1978

SO. Feachem R, Burn E, Caimcross S, ct al: Watcr, Health and Dcvelop-mcnt london, Tri.Mcd Books, 1978

51. Hollister AC Jr, Beck MD, GiWesohn AM, et al: Influence of wIleravailability on Shig~IIQ prevalence in children of farm labor familiesAm J Public Health 45354-362, 1955

52. Khalil M: The relalion betwccn sanitation and parasitic infections inthe tropics J R Sanit Insl 47210-215, 1926

53. Chandler AC A comparison of helminthic and prolozoan infections intwo Egyptian villages two years after the inslallation of sanitaryimprovements in one of them Am J Trop Med Hyg 359-73, 1954

54 Schlie,smann OJ, Atchley FO, Wilcomb MJ Jr, et al Relation of Envi-ronmental Factors to the Occurrence of Enteric Diseases in Areas ofEaslern Kenlucky (PHS Publication No 591). Washington, DC, Gov.ernment Printing Office, 1958, pp 1.35

55 Appropriate Technology for Waste Supply and Waste Disposal inDcvelopin8 Countries Washington, DC, World Bank, 1977

56 White GF, Bradley DJ, White AU: Drawers of Water: Domestic wateruse in Easl Africa Chicago, University of Chicago Press, 1972

57 Wolman A Environmenlal sanitation in urban and rural areas its im.portancc in the conlrol of enteric infections Bull Pan Am Health Organ9157.159, 1975

58. Gordon JE, Behar M, Scrimshaw NS: Acute diarrhoeal disease in lessde.eloped countri.. 3 Melhods for prevention and control Bull WHO3121-28, 1964

59 Puffcr RR, Scrrano CV Pattcrns of Mortality in Childhood.Washington, DC, Pan Amcrian Health Organization, 1973

60 Mata U: The Children of Santa Mari' Cauque A prospective rtcldstudy of health and growth. Cambridge, Massachusetts. MIT Press,1978

61. ld_m The malnutrition-infection complex and its environmentalfactors Presented at the Symposium on Protein-Energy Malnutri.tion .ponsored by The Nutrition Foundation, London, September,1978

62. Mata U, Kronmal RA, Garcia B: Breast-feeding, weaning and tbediarrhoeal .yndrome in a Guatemalan Indian village. Ciba FoundSymp 42:311.338,1976

63 Condon.Paoloni D, CrayiOlO J, Johnston FE, el al: Morbidity andgrowlh of infant. and young children in a rural Mexian villa8e. Am JPublic Health 67:651~56, 1977

64. Martorell R, Habicht J-P, Yarbrough C, et al: Acute morbidity andphy.ical growth in rural Guatemalan children. Am J Dis Child1291296-1301,1975

65. Whitchead RG: Some quantitative considcrations of importance to theimproycmcnt o( thc nutritional status of rural childrcn. Proc R SocLond (Bioi) 199:49-60, 1977

66. Rowland MGM, Colc TJ, Whitchead RG: A quantitatiyc .tudy intothe rolc of infcction in detcrmining nutritional status in Gambian viI.lage childrcn. Br J Nutr 37:441-450, 1977

67. Scrimshaw NS, Taylor CEo Gordon JE: Intcraclions of nutrition andinfection. Am J Med Sci 237:367-403, 1959

68. Murray MJ, Murray AB. Murray NJ, et aI: Refeeding -malaria andhypcrfcrracmia. Lancet 1:653~54, 1975

69. Murray MJ, Murray AB, Murray MB, et al: Thc adyCfSC effect of ironrcplction on tho course of certain infections. Br Med J 2: 1113.1115,1978

70. Murray J, Murray A, Murray M, et al: Thc biological .uppressioo ofmalaria: an ccological and nutritional intcrrclationship of a host andtwo parasitcs. Am J Clin Nutr 31:1363-1366, 1978

71. Clinial trial of liyc mcaslcs yaccinc given alonc and liyc vaccineprcccded by killed yaccinc: fourth rcport to tho Medical RcscarchCouncil by the Maslcs Sub-committee of tho Committee on Deyelop-ment of Vaccincs and Immunisation Proeedurcs Lancet 2:571-575,1977

72 Ministry of Health of Kcnya and the World Hcalth Organizationmaslcs immunity in tho fi~t yar aftcr birth and tho optimum ago (oryaecination in Kenyan children. Bull WHO 55:21-30. 1977 j

73. Mahicu JM, Mullcr AS. Yoorhocyc AM, et al: Pertussis in a rurala~:of Kcnya: cpidcmiology and a prcliminary rcport on yaccinc trial. BullWHO 56:773-780, 1978

74 Kielmann AA, Vohra S Control of tClanusneonatorum in rural com.'munitics -immunization cfTccts of high.oose calcium phosphate ad.sorbed tetanus toxoid. Indian J Med Rcs 66:906-916, 1977

75 Kielmann AA, McCord C: Home treatment of childhood diarrhea inPunjab Yillagcs. J Trop Pediatc 23:197.201. 1977

76. Rohde IE; ~ring (or tbe next round: coDvalcscent care after acuteinfection. Am J Clin Nutr 31:2258.2268, 1978

77. Nalin DR. Levine !.tM, Mala L, et al: Comparison o( .ucrose withglucose in oral therapy of infant diarrhoea. Lancet 2:277.279,1978

78. Chatterjee A, Mahalanabis D, Ialan KN, et aI: Oral rchydration in in.fantilc diarrhoca: controlled trial of a low sodium glucose electrolytesolution. Arch Dis Child 53:284-289, 1978

79. Mahalanbis D, Choudhuri AB, Bagchi NG, et al: Oral fluid therapy Dfcholera among Bangladcsh refugees Iohns Hopkins Med J 132:197.205, 1973

80. Focge WH: Eyaluation of Smallpox EradiationlMeaslcs ControlProgram -Thc Gambia Atlanta, National Communiable DiseaseCenter, 1968

81. ld_m Measles Yaccinalion in Afria: Proocedings -InternationalConference on the Appliation of Vaccincs against Viral, Rickettsial,and Bacterial Diseases of Man Washington, DC, Pan Amerian HealthOrganization. 1971, pp 207-221

82 Gonzalez CL: Ma,s Campaigns and Gcneral Haith Services Geneya,World Haith Organization, 1965

83 Van Der Mci J, Belcher DW Comparing under.fiye programmcs in ahospilal-based clinic and in satellite mobile clinics. Trop Geogr Med26449-456. 1974

84 Wilkinson JL. Smith H, Smith 01: The organization and economics ofa mobile child welfare tam in Sierra Leone. J Trop Med Hyg 70: 14-18,1967

85 Gish 0, Walker G Mobile Hcallh Services London, Tn.Med Books,1977

86 World Hcalth Organization Report of the Meeting of TechnicalReyiew Group III. Geneva, 28 Aug -I Sept 1978: UNDP/WorldBank/WHO Special Programme for Research and Training in TropI-cal Diseases Geneya. World Haith Organiulion, 1978