Addressing inequities in access to fortified sunflower oil: Costs of small, medium & large-scale fortification in Tanzania

Dylan Walters1, Edna Ndau2, Elina Maseta3, Nadira Saleh4, Dr. Theobald Mosha5, and Dr. Susan Horton6

INTRODUCTIONVitamin ADeficiency (VAD) isattributed with 157,000 child deaths (6-59months) globally eachyear (1). InTanzania, VAD is considered aseverepublic health problem responsible forahighburden of thechildmortality andmorbidity. TheTanzania DHS2010estimated that the rateofVAD is 38% inchildren (2). Hunger, lackof fooddiversity, limited consumption ofvitamin-A richfoods andpovertyaremaindeterminants ofVAD. InTanzania, Vitamin A is foundnaturally insome foods butconsumption is less thanoptimal. InManyaraandShinyanga, Only51% and85%children (6-23months) consumed vitamin-A rich in thepast24hours respectively (3). 37%and27%ofchildren (6-59months) received Asupplement in the6months prior (i.e.Aug.2015-Feb.2016) tosurvey inManyara andShinyanga (3). Fortifying staple foodswith vitaminscanbeaneffectivevehicle for improving dietary consumption ofdeficientmicronutrients butisoften limited to foodsproduced through large foodmanufacturers, which arenotaccessibletopoorer and ruralhouseholds.

TheMASAVAproject,in theManyaraand Shinyanga regions ofTanzania, works with smallandmedium-sized enterprise (SME)sunflower oil processors and retailers to fortifyanddistributesunflower oil with Vitamin-A tohouseholds vulnerable toVAD– particularly households withlactatingwomen andyoung children.Thegoal oftheproject is to reduce theprevalence ofVAD in these regions, 48%and42% in2010 respectively (2)byimplementing asustainablebusiness model forvitamin A fortification in these regions between 2014and2017.

KEYHOUSEHOLDSURVEYRESULTSBaselinesurveyMajorityofhouseholds (60.4%)usecookingoilwithbelowguidelinelevels of20—40ug/mgofretinol.HouseholdsinShinyangahavemoreadequatelyfortifiedoil(35%)comparedtoManyara(18%).Thiss tudyfoundVADprevalencerates of72.1%and71.7%inManyaraandShinyangarespectively.Therewas nos ignificantdifference,testedus ingindependentsamplest-testwithap-value=0.05,inmeanRBPlevelbetweenShinyanga(u=14.87,87,SD=4.99)andManyara(u=14.80,SD=5.88)norbetweenintervention(u=14.87,SD=5.37)andcontrols ites(u=14.79,SD=5.23).VADprevalencewas higherinchildrenwithlowestdietarydiversityscore(DDS).VADprevalencewashigherandmeanRBPlevels ignificantlyhigherinchildrenwhowerebreastfed(u=15.31,SD=5.50)comparedtochildrenwhowerenot(u=14.03,SD=4.90),F(1,525)=8.037,p=0.005).Childrenwholivedinhouseholds us ingoilwithretinollevels belowWHOfortificationguidelinelevels hadhigherVADprevalence(73.6%)comparedtothosewithadequate(69.5%)andaboverecommendedlevels(65.8%).Endlinesurvey

1.Startup offortification production with SMEs2.Distribution of fortified oil vegetableoil

3.Behaviour changecommunicationcampaign

4.E-vouchers5.QualityTesting ofoil

38%47. 60%

41. 50%

72. 10% 71. 70%

0%

50%

100%

Nat ional Manyar a Shinyanga

Pr evalence of Vit am in A Def iciency

DHS 2010 MASAVA 2015

• TheMASAVAproject aimstoreduceVADthroughsunfloweroilfortificationwithvitaminAbySMEoilprocessorsandretailers.

• VADratesinchildrenarehighinManyara(72.1%)andShinyanga(71.7%)regionsofTanzania- higherthanpreviouslyreported.

• Theincrementalcostoffortificationisestimatedtorangefrom0.6%forlarge-scale(or$0.04percapita/year)toahighof22.1%(or$1.51percapita/year)forsmall-scaleenterprisesproducingoilin20Lcontainerscomparedtothelowestpriceofunfortifiedoil.

• Low-costandavailableoptionsforfortificationequipmentandpackaginginformatspreferred cansubstantiallyreducetheincrementalcostofSMEfortificationfrom$0.51to$0.10perlitre,ifgovernmentapprovestheiruseandremovesanewVATinordertoincreaseaccessandaffordabilityinruralandpoorerregions.

• ImprovingaccesstofortifiedsunfloweroilinruralandpoorerregionsofTanzaniarequiresinnovationbySMEsandregulatorstoincorporatelow-costoptionsthatsuitneedsofthemostvulnerablehouseholdswithchildrenandmothersatriskofVAD.

REFERENCES1. BlackRE,VictoraCG,WalkerSP,BhuttaZA,ChristianP,deOnisM,etal.2013.Maternalandchildundernutritionand

overweightinlow-incomeandmiddle- incomecountries.Lancet.382(9890):427-451.2. NationalBureauof Statistics,ICFMacro.TanzaniaDemographicandHealthSurvey2010.2011.3. Ministryof Health,CommunityDevelopment,Gender,ElderlyandChildren(MoHCDGEC)[TanzaniaMainland],Ministryof

Health(MoH)[Zanzibar] ,NationalBureauofStatistics(NBS) ,Officeof theChiefGovernmentStatistician(OCGS),andICF.2016.TanzaniaDemographicandHealthSurveyandMalariaIndicatorSurvey(TDHS-MIS) 2015-16.DaresSalaam,Tanzania,andRockville,Maryland,USA:MoHCDGEC,MoH,NBS,OCGS,andICF.SMESurveyandotherreport

4. Fiedler,JLandR.Af idra.2010.VitaminAfortificationinUganda:Comparingthefeasibility,coverage,costs,andcost-effectivenessof fortifyingvegetableoilandsugar.

5. MarketAxisLimited.2016.SunflowerOilSMEIndustryReport:InputintoavitaminAstrategyforvitaminAfortification.6. Randall,P.2016.Fortif icationofSmallScaleColdPressedEdibleOilProductionReport.7. TanzaniaMinistryof Industry,TradeandInvestment.2016.Tanzaniasunf lower sectordevelopmentstrategy2016-2020.

FUNDINGSUPPORTThis project is undertaken with the financial support of the International Development ResearchCentre (IDRC)and theGovernment ofCanadaprovided through Global AffairsCanada (GAC).AUTHORAFFILIATION1. PhDcandidate, Canadian Centre forHealthEconomics, IHPME,University ofToronto, Canada2. PhDstudent, SokoineUniversity ofAgriculture, Tanzania3. PhDstudent, SokoineUniversity ofAgriculture, Tanzania4. Project Manager, MEDA, Canada5. Professor, SokoineUniversity ofAgriculture, Tanzania6. Professor andCIGIChair, Global Health Economics, University ofWaterloo, Canada

SUMMARYRESEARCH QUESTIONWhat are theprivate sector incremental costs (IC)and cost drivers of locally-produced ediblesunflower oil fortified with Vitamin AbySMEs compared to large-scale producers?METHODS• Fortification trialwas aquasi-experimental non-equivalent control group researchstudydesign. Therewere3 intervention districts and1control district ineachof the tworegions ofstudy.Datacollected using baseline and endline household and retailer surveys.

• Using thevitaminA fortification costing framework fromFiedler andAfrida 2010 (4), thiscosting analysis model used aningredients approach toestimate thehypothetical costs offortification foreachofahypothetical small, medium and large-scale enterprise – a low-cost-small-scale scenario - using costdatacollectedduring the retailer and SMEsurveys (5,6, 7)anddata from literature andproject reports.

Type/scale ofsunfloweroilproduction

ICperlitre

ICpercapita/yr

Avg.HouseholdCost/Yr

%Increase

Unfortified(min.) $34.72Large-scale fortif ied(20L) $0.01 $0.04 $34.95 0.66%

Medium-scale fortif ied(20L) $0.16 $0.87 $39.15 12.78%Medium-scale fortif ied (1L) $0.39 $2.11 $45.47 30.97%Small-scale fortif ied(20L) $0.28 $1.51 $42.43 22.21%Small-scale fortif ied(1L) $0.51 $2.75 $48.74 40.40%Low-costmedium-scalefortif ied(hypothetical) $0.12 $0.62 $37.90 9.17%Low-costsmall-scalefortif ied(hypothetical) $0.10 $0.53 $37.41 7.75%

DISCUSSIONThe incremental costof fortification isminimal for large-scale producers,similar totheexperience inUganda(4). Knowing that increasing access toSME-produced oil is likely necessarytobridge theurban/rural divide inaccess to fortified oil, SMEsneed toinnovate to reduce thehighequipment andpackagingcosts.Removal ofthenew 18%VAT forSMEsmayalsohelp makefortified oilaffordable.Reducing costs maybeessential forcreating the incentive toenter the fortification market,produce affordable product forpoorer and rural households thatarekeyto reducing VAD.

ConclusionResults ofcosting analysis modelemphasize theneed forgovernment regulators andSMEs in thesunflower oil industry toconsider innovating with low-cost orcost-savingactions to increaseequity inaccess to fortified oiland, ultimately, sustainable reductions inVAD in children andmothers. Endline baseline datawill help estimate the impactofthe intervention reducing VAD inchildren andmothers.

RESULTS• Thebaseline dataon retinol levels inhousehold cooking oil showed: 1)Theretinol content was lower inManyaracompared toShinyanaga and2)within eachregion, rural districts had lower content thanurban districts (butonly significant forManyara.)

• As well, inManyara, only8%ofhouseholdsreported purchasing aknown fortified brandofoil that, whereas inShinyanga 76%ofrural and96%ofurbanhouseholds did.

RESULTS(Continued)• Themodelled incremental cost (IC)of fortification is estimated at$0.01,$0.16and$0.28per litre soldin20Lcontainers for large,mediumandsmall-scale entreprisescompared tounfortified oil.

• This represents arangeof0.66%increase in large-scale oil, or$0.04percapitaperyear, to22.21%increase for small-scale enterpriseoil, or$1.51percapitaperyear,compared tounfortified.

• The largest annuitized costcomponent of fortification for large-scale enterprises producing oilin20L is thepremixcost (60%)whereas formedium-scale it ispremix (31%)andequipment(21%), and for small-scale it is mainlyequipment (48%).

Low-cost small-scale scenario• Thepackagingcost for1Lcontainers for small- andmedium-scale producers itself canaddup to

$0.23centsper litre, or18% increased costper litre forsmall-scale producers. Using20Lcontainers is the largest factor in increasing costs forSMEscurrently.

• Allowing SMEs todistribute sunflower oil in250gand500gsachets would reduce the ICperlitre by$0.16, andsafer for ruralhouseholds who typicallypurchase oil by the“scoop”.

• Approving SMEusageof low-cost plastic ormild steel tank formixing and removing the18%VATonoil would reduce incremental costby$0.04and$0.05per litre.

• All threemeasures combined could reduce the incremental costofmedium-scale production toonly9% higher than lowest priceofoil (270TSHper litre)

Limitations• This costingmodel is preliminary and relies oncostdataandassumptions about theproduction

based onproject reports and literature. Workunderway oncostscomponents andprices ofoilandequipment/materials would improve theaccuracy.Thisdoes not include public sectorcostsof fortification foradvocacy,social marketing, QAand MNE,nor fluctuations in seasonal costs.

SUMMARY

05101520253035

Manyara Shinyanga

Meanretinol contentofhouseholdcookingoil (u/kg)

Rural Urban

1.65

1.61

1.6

1.49

1.42

1.36

$1.00 $1.25 $1.50 $1.75

Current medium-scale fortification scenario (1L containers)

Use of low-cost plastic mixing tanks

Remove VAT (18%) from sales of SME edible oil

Approve the use of low-cost 250/500g oil sachets

Use 20L container size

Combining use of 20L containers, low-cost equipment and removal

of VAT.

Sensitivity analysis: Cost of medium-scale sunflower oil per litre compared to unfortified oil by varying key

cost dr ivers (US$)