researchonline.lshtm.ac.ukresearchonline.lshtm.ac.uk/1918305/1/tougher_edited_lw.docx · web...

“TOUGHER_July_8_2014.docx” 7/8; lw 7/14; lw to jf 7/14; jf to ms 7/15; ms 7/16; jf to lw 7/16; lw to au 7/16;

Supranational subsidies to iImproveing Access in the private sector: Evidence

from a To Malaria

Medicine Through Private-

Sector subsidy Subsidies program In Seven7

African Countries

Abstract

Improving access to quality-assured artemisinin combination therapies (ACTs)

(QAACTs) is an important component of malaria control in lower- and middle-income

countries. In 2010, the Global Fund to Fight AIDS, Tuberculosis, and Malaria launched

tThe Affordable Medicines Facility— – malaria (AMFm) project was launched as eight

national scale pilot programs in seven African countries. The goal of the project was in

2010. The program aimed to decrease malaria morbidity and delay the development of

drug resistance by increasing QAACT the use of ACTsartemisinin combination

therapies, primarily such therapies while decreasing use of other treatmentsthrough

subsidies intended to reduce costs. We collected data on price and retail markups on

antimalarial medicines from 19,625 private for-profit retail outlets before and 6–-15

months after the program's implementation. We found that in six of the AMFm's pilot

programs, QAACT prices for quality-assured ACTsartemisinin combination therapies

decreased by US$1.28–-4.34 and absolute retail markups on these therapiesQAACTs

decreased by US$0.31–-1.03, in six of the eight pilot programs. Prices and markups on

other antimalarial classes of antimalarials also changed duringover the evaluation

period, but not to the same extent. In all but two of the pilot programs, we found

1

Ferris, Jeanne, 07/15/14,

INSTRUCTIONS TO AUTHORS:Health Affairs’ editors are committed to working collaboratively with authors. Use this Word document for all revisions, comments, and corrections and return to Health Affairs.Please leave all Health Affairs edits visible in track change mode. Please also leave all Health Affairs balloon comments.Please ensure all author-entered edits are visible in track change mode.Please respond to all comment balloons by typing “OK” or providing an appropriate response in the comment balloon or within the brackets. In the Notes and exhibits sections, author queries may appear within brackets (“<< >>”).


AU: Below in the text you say "There were substantial decreases during the study period in median quality-assured ACT price in six of the eight pilots, ranging from US$1.28 to US$4.82 per adult equivalent treatment doses." There we pointed out that Exhibit 1 shows a decrease in Uganda of $0.83, which is smaller than $1.28. Please be sure that abstract, text, and exhibit are consistent.


AU: Information in the abstract needs to be "abstracted" from the text. We did not find this number in the text, so please add it there to allow it to remain in the abstract.


evidence that there wasfor further scope for price reductions. Thus, confirming that

;concerns may be warranted that about distributorsers are capturing subsidies instead

of passing them on to consumers may be warranted. These findings demonstrate that

supranational subsidies can dramatically reduce retail prices of health commodities, but

concerns about the ability of distribution chain actors to capture subsidies may

warranted. Recommended retail prices (RRPsand that recommended retail prices )

communicated to a wide audience may be an effective mechanism for controlling

market power.

Introduction

In many low- and middle-income countries, patients frequently seek treatment for

serious illnesses like such as malaria, diarrhea, and tuberculosis in the private for-profit

sector, where treatment quality may vary and treatment is more expensive than in the

public sector are key constraints. Subsidies for high-quality medicines for certain

diseases have been a prominent global health strategy for improving treatment

outcomes in the private for-profit sector in these countries. A key concern with this type

of intervention is that subsidies are not passed alongthrough to patients in the form ofas

lower prices for preferred medications in the private sector butand are instead are

"‘captured"’ before they can get to patients by wholesalers and retailers, who use the

subsidies to earn "supernormal" profits.

To determine whether suppliers were capturing the subsidy intended to benefit

patients, weThis paper analyzedexamines the Affordable Medicines Facility— – malaria

(AMFm), the largest initiative of this kindsubsidy program to date for the private sector

2

lw, 15/07/14,

AU: Market power of who or what?

lw, 07/15/14,

AU: What does “further scope” mean?


of low- and middle-income countries, the Affordable Medicines Facility – malaria

(AMFm), to examine whether there was evidence that suppliers were capturing the

subsidy intended for the benefit of patients so that suppliersthey could to earn super-

normal profits. This articleIt examines whether there iswas potential for further

reductions in the prices of subsidized medicines to be reduced further through greater

adherence to recommended retail prices (RRPs) and through further reductions in retail

markups.

Background On Malaria Treatment And AMFm

Malaria is the fifth leading cause of years of life lost from premature mortality (1).

[1] Improving access to first-line treatments, known as artemisinin combination

therapies (ACTs), is critical to malaria control.

Even though ACTs are provided for free in the public sectors of many low-income

countries, ACT use of these therapies remains low, because ofdue to unreliable public-

sector supplies and the long distances patients have tomust travel to facilities to obtain

the therapiessupplies. In private for-profit outlets, where patients frequently seek

malaria treatment is frequently sought, less-effective medicines are widely available and

substantially cheaper than ACTs (2, 3).[2,3]

The Uuse of antimalarials other than ACTs has two important implications for

public health. First, mortality increases when medicines are used that are no longer

effective because ofdue to drug resistance are used. Second, the use of antimalarials

other than ACTs such as artemisinin or other monotherapies likely contributes to the

spread and intensification of drug resistance. Drug resistance is less likely to develop

when combination therapies with two independent modes of action are used, because a

3

msaunders, 07/15/14,

AUS: is this in developing countries? In the world? Please clarify


malaria parasite is less likely to spontaneously mutate sosuch that it becomesis

resistant to both drugs simultaneously (4).[4]

In order tTo increase the appropriate use of quality-assured ACTs (QAACTs) and

decrease the use of other antimalarials, the Global Fund to Fight AIDS, Tuberculosis,

and Malaria launched AMFm in 2010. AMFm operated as eight national-scale pilots in

seven countries— (Ghana, Kenya, Madagascar, Niger, Nigeria, Tanzania (mainland

and Zanzibar, a semi-autonomous part of Tanzania) and Uganda—) until 2012. It, and

was then incorporated into the Global Fund’s regular financing stream (5).[5]

AMFm had three components: price negotiations with manufacturers of

QAACTsquality-assured ACTsartemisinin combination therapies to obtain lower- cost

supplies of these medicines; the provision of subsidies in the form of co-payments from

the Global Fund to manufacturers for purchases of these therapies by approved

importers; and demand- and supply-side interventions to facilitate program

implementation and the rational use of QAACT such therapiesuse, such as through

RRPsrecommended retail prices and communications campaigns (6).[6] Further details

on the operation of AMFm can be found elsewhere (7-9).[7–9]

An independent evaluation of AMFm in 2012 assessed changes in Qquality-

assured ACTartemisinin combination therapy QAACT prices, availability, market share,

and use against pre-defined success benchmarks of success after 6–-15 months afterof

the program's implementation. The evaluation reported large improvements—

particularly in the private for-profit sector— in the QAACT availability of such therapies,

with lower prices and increased market share for these products in most pilots (7, 8).

[7,8] The few pilots with appropriately timed household survey data showed some

4


AU: Reworded to be consistent with the abstract.


AUS: was the evaluation in 2012? If not, please correct


AU: Here you discuss the evaluation as if it had nothing to do with you, but in the "Study Data And Methods" section you discuss it as if it were your study. Please reword to clarify your relationship with the evaluation.


limited evidence of improved use of QAACTsquality-assured ACTsartemisinin

combination therapies for the treatment of fever (10-12).[10–12]

In spite of observed improvements, there is still concern that the AMFm subsidy

was captured by wholesalers and retailers. Maximizing price reductions is critical to

subsidy interventions like AMFm, because poorer households would may not benefit

from the intervention and patients may continue to purchase non-recommended

treatments if prices of the preferred drugs remain high (13).[13]

Study Data And Methods

The AMFm Iindependent Eevaluation used a pre-post study design with thorough

documentation of the implementation process and context, in accordance with current

guidance for evaluating complex interventions (14).[14]

Price and markup data were collected using nationally representative surveys of

private for-profit retail outletsantimalarial outlets. In most pilots, baseline data collection

took place in the period August–-December 2010., while eExisting surveys conducted

by ACTwatch were used as baselines in Nigeria (September–-November 2009) and

Madagascar (April–-June 2010)(15) were used as baselines.[15] End-line data

collection in all pilots took place in the period October 2011–-January 2012 in all pilots.

The Ssurveys were timed to coincide with malaria transmission seasons, and

baselines data collection waswere intended to take place before the arrival of co-

paidsubsidized ACTs. However, although baseline data collection in Kenya began one

month after the first imports of drugs subsidized by AMFm co-paid drugs.

Study Sample

5


AUS: previously you discuss subsidies; here you mention co-paid ACTs; we need to keep the terminology the same so the reader won’t become confused; I think subsidized is the more appropriate term here; ok?


AUS: please explain/define “endline” as used here.

lw, 07/15/14,

AU: Confirm accuracy.


Outlet surveys used a cluster sampling approach. Clusters were administrative

areas with approximately 10,000–-15,000 inhabitants (for example,e.g. sub-district

level). Sampling frames were stratified by urban orand rural domain, and clusters were

selected using probability proportional to population size sampling.

Samples were drawn independently at baseline and end line. Sample sizes were

calculated to detect a 20- percentage- point change in the availability of quality-assured

ACTsartemisinin combination therapy QAACT availability in each domain, with 80%

percent power and 5% percent significance. Sample size calculations at baseline

assumed a design effect of four and baseline availability of 40 percent%., while

sSample size requirements at end line were calculated using baseline results for

availability and design effect.

A full census was conducted throughout Zanzibar, due to because of its small

population. Further details on the samples of each pilot may be found in the onlineweb

Aappendix (16).[16]

BecauseAs complete lists of antimalarial stockists were not available, all outlets

with the potential to sell antimalarials within a cluster were listedenumerated.

BecauseAs few public health facilities and pharmacies were likely to be found in each

cluster, they were oversampled by listingenumerating all such providers within a larger

administrative area (for example, ae.g. district) in which a selected cluster was located

(15).[15]

Screening questions were administered at each listedenumerated outlet. Outlets

that had antimalarials in stock or that had stocked them in the previous three months

were eligible for interview. A questionnaire was administered at eligible outlets following

6

lw, 07/15/14,

AU: Who or what is “they?”


AU: Please reword to clarify the meaning. Is the point something like "All of Zanzibar was surveyed"?


AU: Please reword to clarify the meaning, using nontechnical language.


the obtaining of informed oral consent. The questionnaire covered provider

characteristics and information on each antimalarial in stock, including its retail selling

price and wholesale purchase price.

Ethics approval was obtained from national ethics committees in each country

and the Institutional Review Boards of ICF International and the London School of

Hygiene and Tropical Medicine.

Statistical Analysis

Analysis was conducted using Stata v11.0 and R v2.14.2.Antimalarial medicines

were classified into four categories: non-artemisinin therapies, artemisinin

monotherapies, non-quality- assured ACTs, and QAACTsquality-assured

ACTsartemisinin combination therapies. The Global Fund’s quality-assurance policy

was used to identify QAACTsquality-assured ACTsartemisinin combination therapies

(17),[17] while the presence of the AMFm logo at end line was used to

identifydistinguish AMFm- co-paidsubsidized QAACTstherapies.

We present Rresults are presented in the main text for QAACTs two categories:

(quality-assured ACTsartemisinin combination therapies and non-artemisinin therapies

only), but (results for other antimalarial categories may be found in the web aAppendix

(16).[16] We focused on these two antimalarial categories because AMFm sought to

lower the prices of QAACT quality-assured ACTsartemisinin combination therapies

prices so that they were competitive with non-artemisinin therapies. Statistical Aanalysis

was conducted using Stata, version 11.0, and R, version 2.14.2.

7

lw, 07/15/14,

AU: “presence of the AMFm logo” where?


AUS: This sentence moved further down in paragraph


AU: In Exhibit 4 notes you refer to the "importer's price." Is that the same as the "wholesale purchase price"? If so, please use one term consistently throughout your paper to avoid confusing the reader.


QAACTsQuality-assured ACTsartemisinin combination therapies , which were

targeted for subsidization under AMFm, are exclusively tablet formulations., while

oOther antimalarial categories also have oral and injectable formulations. BecauseAs

oral liquids and injections have different price distributions and tend to be more

expensive than tablets (15),[15] analysis was restricted to tablet formulations. Thisto

ensured that price and markup data were comparable across antimalarial categories.

Price data were collected in country currencies. Baseline data for Nigeria and

end-line data for all pilots were converted into 2010 dollars using national consumer

price indices (18).[18] Prices were then converted to US dollars using the average

interbank rate for 2010 (19).[19]

Absolute retail markups were calculated for each product as retail selling price

minus wholesale purchase price. Relative retail markups were calculated by dividing a

product’s absolute retail markup by its wholesale purchase price. Total markups, which

capture the cumulative markup added by importers, other distributors, and retailers,

were calculated by subtracting a particular product’s mean importer price from its retail

selling price. Mean importer prices were available for co-paidsubsidized

QAACTsquality-assured ACTsartemisinin combination therapies only., and They were

calculated by pilot for each QAACT brand, age-range band, and package size using

routine data collected by the Global Fund. Markups therefore cover both costs to the

provider and profit margins.

Retail selling prices, absolute retail markups, and total markups are reported in

terms of adult equivalent treatment doses, which are (AETDs), defined as the amount of

active pharmaceutical ingredient required to treat a sixty60-kilogramkg adult (20).[20]

8


AU: Please see questions below about your use of the term "adult pack" in some places. If that term must stay in the paper, please refer to it here.


AU: As asked above, is this the same as the "wholesale purchase price," which you also refer to in this paragraph? If so, please use just one term consistently.


ForWithin each pilot we present the median and interquartile range (IQR) for all

price and markup indicators. Point estimates were weighted using survey weights, and

standard errors accounted for clustering and stratification. We show the difference

between baseline and end line in median retail price and absolute retail markup, and the

p- value from the Wilcoxon rank- sum test of the hypothesis of no difference between

baseline and end-line distributions.

Results are shown for private for-profit outlets only. Antimalarials were generally

free or subsidized in public- health sectors of the AMFm pilots., while In addition, the

numbers of community health workers and private not-for- profit facilities were very

small (8).[8]

Limitations

It was not possible to restrict the intervention geographically within participating

countries to create control areas, nor was it feasible to use countries that were not

participating in AMFm as comparators. The lack of formal comparators limits the degree

to which changes over time in prices and markups may be attributed to AMFm. We

addressed this limitation by systematically collecting data on the implementation

process of [please provide] and context through key informant interviews and document

reviews in each pilot (14).[14] Further justification for the non-experimental design of the

evaluation is available elsewhere (8, 21).[8,21]

We intendedaimed to conduct baseline surveys no more than two months before

the first arrival of the first co-paidsubsidized ACTs, and end-line surveys were timed to

maximize the length of program implementation. In practice, the midpoint of baseline

9


AU: Please reword to clarify the point. Is the meaning something like "to maximize the length of time that the program was in effect"?


AU: Please briefly explain in the text—whose implementation process?


AU: Please add some wording to the text to explain the role of community health workers here.


AU: Please reword to clarify how a "pilot" can have a "public-health sector." Is the meaning something like "in public-health sectors of the countries where AMFm pilots took place"?


data collection took place 4.5–-15.0 months beforeprior to the arrival of co-

paidsubsidized ACTs in five of the eight pilots. This was (because of the use of existing

surveys for baseline data and delays in program implementation).

End-line data collection took place after a fairly short period of AMFm

implementation. By the midpoint of this collection, Hhalf of the pilots had had co-

paidsubsidized ACTs available in their -country for more than aone year; by the mid-

point of endline data collection, while the other halfremaining pilots had had themco-

paidsubsidized ACTs for 6.5– to 9.5 months. All pilots had supporting interventions in

place for less than aone year (8).[8]

It is unclear what impact a longer implementation of AMFm would have had on

prices and markups, especially in light of order rationing by the Global Fund because

ofdue to financial constraints as of mid-2011. Indeed, small-scale price tracking surveys

that were conducted in six of the AMFm pilots found evidence that prices of Qquality-

assured ACTsartemisinin combination therapy QAACT prices were increasing in

Ghana, Kenya, and Uganda in 2012 (22).[22]

Reporting bias might affect price and markup data collected through recall. For

example, outlets might report lower prices than they actually charged or might conceal

antimalarials that they are not permitted to stock. Non-response on wholesale purchase

price is a potential source of bias. Retail selling price was reported for 91.2– percent%

to 99.3 percent% of antimalarial observations in private for-profit outlets. However, but

markup data wereare only available for only 45.4– percent% to 94.0% percent of

observations because ofdue to low reporting of wholesale purchase prices. High unit

10


AU: Please reword to clarify the point—they literally might hide the drugs? Or they might not reveal information about the drugs?


AUS: please clarify- recall of persons interviewed who work in outlets?


AU: As above, please reword to clarify the meaning. Is the point that the implementation process could have been longer, or that more time could have elapsed after implementation?


AU: Please reword to clarify the point. Is the meaning something like "End-line data collection took place a fairly short time after AMFm was implemented"? Or it "took place a fairly short time after the beginning of AMFm implementation"?


non-response for wholesale purchase price mightmay reflect concerns that the

information couldmay be shared with regulatory authorities, or cases in whichwhere the

respondent was not involved in purchasing stock and therefore was unaware of

wholesale prices.

Study Results

Retail Prices

At baseline, QAACTsquality-assured ACTsartemisinin combination therapies

were more expensive than non-artemisinin therapies in all pilots except Madagascar

(Exhibit 1). In all other pilots, median price per adult equivalent treatment doses AETD

was US$2.47– to US$5.99 for QAACTsquality-assured ACTsartemisinin combination

therapies, but was only US$0.14 i. In Madagascar the price was only US$0.14,

reflecting the presence of a pre-existing national ACT subsidy program. Baseline

median price per adult equivalent treatment doses AETD of non-artemisinin therapies

ranged from US$0.31 to US$1.39 in all pilots at baseline, making them the least

expensive treatment category everywhere but Madagascar.

There were substantial falls decreases duringover the studytime period studied in

median QAACT quality-assured ACTartemisinin combination therapy price in six of the

eight pilots, ranging from US$1.28 to US$4.82 per adult equivalent treatment doses

AETD (p < 0.0001) (Exhibit 1). In Madagascar there was a price increase of US$0.46 (p

= 0.0009). In Uganda, we could not reject the null hypothesis that end-line and baseline

quality-assured ACTartemisinin combination therapy QAACT prices were the same (p =

0.2647). At end line, the median quality-assured ACTartemisinin combination therapy

QAACT price was US$0.58– to US$1.96 across the eight pilots, which was equivalent

11


AU: Exhibit 1 shows Uganda QAACT base as $2.79 and QAACT end as $1.96. Doesn't that mean that the decrease during the study period in Uganda was $2.79 - $1.96, or $0.83, which is smaller than $1.28 (in the text)? Please reword to avoid this apparent inconsistency.


AU: Please reword to clarify the meaning of "unit" in this context.


to less than a quarter of their baseline levels in three pilots and less than half of their

baseline level in three otheradditional pilots.

We also observed price changes in the prices of drugs in other antimalarial

categories. However, the magnitude of the reductions in these prices of other

antimalarial categories was much smaller than that of the declines in prices of quality-

assured ACTsartemisinin combination therapy QAACT prices everywhere exceptbut

Niger, where the price for non-artemisin therapy increased by US$0.17 (p = 0.0022)

(Exhibit 1). Prices of non-artemisinin therapies decreased by US$0.03–-0.13 in three

pilots and increased by US$0.17 in one pilot Niger(p < 0.0001 to p=0.0022). In the

remaining four pilots, there were some apparent decreases in non-artemisinin therapy

prices, but there was no statistical evidence that baseline and end-line prices differed (p

= 0.4414– to p=0.9323).

Non-artemisinin therapies were the cheapest treatments in all pilots at end line

except in Tanzania mainland and Kenya. In those two pilots, where median Qquality-

assured ACTartemisinin combination therapy QAACT prices were equal to those of

non-artemisinin therapies.

[Exhibit 1]

In addition to the change in overall prices for quality-assured ACTsartemisinin

combination therapy QAACT prices overall, there were substantial decreases in prices

in rural and urban areas in all pilots except Madagascar and Uganda (for further details,

see the web aAppendix for further details (16)).[16] In all pilots Aat baseline,

QAACTsquality-assured ACTsartemisinin combination therapies were cheaper in rural

areas compared to urban areas in all pilots. In six pilots, decreases in QAACT prices

12


AU: Does this p value go with the increase in Niger, and the other value with the decreases below? If not, please reword as needed to clarify which of the two values goes with which dollar amounts.


were larger in urban areas compared to rural areas. At end line, in all pilots except

Nigeria and Uganda, QAACTsquality-assured ACTsartemisinin combination therapies

were still less expensive in rural areas compared to urban areas in all pilots other than

Nigeria and Uganda.

Exhibit 2 examines QAACT prices in relation to RRPsrecommended retail prices

for AMFm co-paidsubsidized QAACTsquality-assured artemisinin combination therapies

that were in place in all pilots but Madagascar. RRPsRecommended retail prices were

set at different levels for each age bandrange, and were promoted through

communications campaigns (9).[9] RRPsRecommended retail prices for adult packs of

malaria pills ranged from US$0.43 in Nigeria to US$0.93 in Ghana (2010 US$

equivalent). In Ghana, Kenya, Tanzania, mainland and Zanzibar, the median and 25thth

percentile prices for adult quality-assured ACTartemisinin combination therapies

QAACT formulations wereas equal to or just above the recommended retail pricesRRP

(Exhibit 2). In theall other three pilots, the median prices exceeded the recommended

retail pricesRRP, markedly so in Uganda.

[Exhibit 2]

13


AU: Exhibit 2 title referred to adult packs but the exhibit's notes referred to adult treatment. Please see the questions about "adult packs" in the exhibits and be sure that the wording here is consistent with the final version in Exhibit 2. (In "Absolute Retail Markups" you refer again to adult equivalent treatment doses.)


Relative Retail Markups

At baseline, non-artemisinin therapies had the highest relative retail markups in

all pilots, with median markups ofranging from 40.0– percent%-81.8 percent% (for more

details, web see the aAppendix for more details (16)).[16] Median relative retail markups

for quality-assured ACTs were 33.3–% percent -50.0% percent for QAACTsquality-

assured artemisinin combination therapies.

Median relative markups foron QAACTsquality-assured ACTsartemisinin

combination therapies increased in all pilots (p < 0.0001 to p = 0.0220) except Niger

and Kenya, where there was no statistical evidence of change (p = 0.2624 and p =

0.4302, respectively). The largest increases were observed in Zanzibar and Uganda,

where the median increased by 77.3 percentage points and 83.1 percentage points,

respectively. The median increased by less than 16.7 percentage points in the

remaining pilots.

Median relative markups for non-artemisinin therapies increased by 4.4

percentage points in Niger (p = 0.0252) and 12.5 percentage points in Nigeria and

Nigeria, respectively (p=0.0252 and p = 0.0271)., and tThere was no statistical evidence

of change in the other pilots (p = 0.4836– to p=0.9005).

At end line, QAACTsquality-assured ACTsartemisinin combination therapies had

the highest relative markups in three pilots (Nigeria, Uganda, and Zanzibar)., while

nNon-artemisinin therapies still had the highest relative markups in the other five.

Absolute Retail Markups

Although nNon-artemisinin therapies had the largest relative markups at

baseline. However, they had the smallest absolute retail markup per adult equivalent

14


AU: Correct? If not, please clarify which p value goes with which country.


treatment doses AETD everywhere exceptbut Madagascar. Baseline median absolute

markups on non-artemisinin therapies ranged from US$0.14 to US$0.35 in all pilots but

Uganda, where they were US$0.61 (Exhibit 3). Baseline absolute markups on

QAACTsquality-assured ACTsartemisinin combination therapies ranged from US$0.68

to US$1.41 everywhere but Madagascar, where they were US$0.06.

There were large falls decreases in median absolute markups on

QAACTsquality-assured ACTsartemisinin combination therapies in six pilots, ranging

from US$0.31 to US$1.03 per adult equivalent treatment dosesAETD (p < 0.0001)

(Exhibit 3). In Madagascar there was, an increase of US$0.11 in Madagascar (p =

0.0006). and In Uganda there was no substantial change in Uganda (p = 0.798).

Changes in median absolute markups foron other categories of malaria drugs

were also observed in some pilots. However, but decreases were smaller in both

absolute and relative terms when compared to QAACTsquality-assured

ACTs,artemisinin combination therapies with few exceptions (for more details, see the

web aAppendix for more details (16)).[16] For non-artemisinin therapies, there were

small changes of US$0.01– to US$0.02 in absolute markups in four pilots (p < 0.0001 to

p = 0.0103), an increase of US$0.09 in one pilot (p < 0.0001), and no substantial

changes in three pilots (p = 0.6208– to p=0.7639) (Exhibit 3).

In spite of the large reductions in absolute retail markups on QAACTsQquality-

assured ACTsartemisinin combination therapies in six of the eight pilots, Kenya and

Tanzania were the only two pilots in whichwhere QAACTsthese therapies had the

lowest absolute retail markups at end line. In the other pilots, non-artemisinin therapies

had the lowest absolute markups at endline. In these six pilots, median absolute retail

15


AU: Some condensing to streamline the argument.


markups on QAACTsQquality-assured ACTsartemisinin combination therapies were

US$0.06– to US$0.32 higher than those on non-artemisinin therapies.

[Exhibit 3]

Total Markups

Total markups— from the importer's price to the retail selling price— of co-

paidsubsidized ACTs were lowest in Kenya and Madagascar (US$0.40 and US$0.45),

followed by Tanzania mainland and Ghana (Exhibit 4US$0.84 and US$0.87). Total

markups were above US$1.00 in the remaining four pilots, and were highest in Uganda

(US$1.83) (Exhibit 4).

Retail markups accounted for less than 50.0 percent% of total markups on

subsidizedco-paid QAACTsquality-assured ACTsartemisinin combination therapies

everywhere exceptbut Uganda, where the retail markup was 53.6 percent% of the total

markup from importer price to retail selling price. Retail markups were the lowest

relative to total markup in Niger (30.3 percent), followed by Nigeria (35.5 percent),

where they were 30.3 percent% and 35.5 percent% of total markups, respectively.

[Exhibit 4]

Discussion

This study presents nationally representative antimalarial price and markup data

from the private for-profit sector across all AMFm pilots in seven African countries

before and 6–-15 months after the arrival of AMFm- co-paidsubsidized ACTs. We

foundOur study showed substantial reductions in quality-assured ACTartemisinin

combination therapyQAACT prices and absolute markups in six of the eight pilots.

16


AU: Some condensing here and below, to avoid repeating in the text too much information from the accompanying exhibit.


AU: Shouldn't this be "quality-assured ACTs" here, to be consistent with the next paragraph?


AU: Data for Exhibit 3 has endline QAACT price of 17 cents and endline nAT price of 13 cents for Madagascar, a difference of 4 cents instead of the 6 cents you mention in the text. Please reword to avoid this apparent inconsistency.


No decrease in the price of QAACTsquality-assured ACTsartemisinin

combination therapies was observed in Madagascar, where baseline prices were much

lower than in other pilots, due to because of the presence of a highly subsidizedsed

papediatric quality-assured ACTartemisinin combination therapy that was QAACT

distributed through a national social marketing program. Nevertheless, Madagascar had

the second lowest QAACT prices at end line. In Uganda, the lack of overall decrease in

QAACT price may be attributed to very low levels of awareness of AMFm and

RRPsRrecommended retail prices, a result of thedue to substantial delays in the

implementation of communication interventions (8, 9).[8,9]

In addition to changes in quality-assured ACTartemisinin combination

therapyQAACT prices and markups, we observed decreases in retail prices and

absolute retail markups of other types of antimalarials in most pilots. There are two

possible explanations for the reductions observed for non-subsidizsed antimalarials.

First, QAACTsquality-assured ACTsartemisinin combination therapies are

substitute goods for other types of antimalarial types. Thus, so prices for other

antimalarial categories may be lowered in response to decreased demand for non-

subsidizedsed medicines.

Second, there could be declines in the prices of antimalarial medicines for

reasons unrelated to AMFm, such as reductions in marginal costs. Even if there wereas

a downward trend in prices of antimalarial medicines, it is likely that AMFm was still

responsible for a large share of the reductions in quality-assured ACT QAACT prices,

since the changes in thesequality-assured artemisinin combination therapyQAACT

17


prices in relative and absolute terms were typically much larger than the changes we

observed for other antimalarial categories.

Opportunities For

Is there scope for further reduction in QAACT prices?Further Price Reductions

Although tThere were substantial reductions in quality-assured ACTartemisinin

combination therapyQAACT prices in several pilots. However, it is important to assess

whether there is scope for further reductions in retail prices in order to maximizese the

potential public health impact of the intervention.

Our comparison of RRPsrecommended retail prices and end-line QAACT prices

provides some insight into this issue. RRPsRecommended retail prices appeared to act

as a price floor. The 25thth percentile for prices of adult packs of QAACTsquality-

assured ACTsartemisinin combination therapies did not fall below their

RRPsrecommended retail prices in any pilot.

One factor thatwhich appears to help explain whether QAACT prices were close

to their recommended levels was the extent of supporting communication and public

awareness campaigns (8).[8] In the four pilots with high RRP awareness of

recommended retail prices and consistent implementation of communications

campaigns (Ghana, Kenya, Tanzania, mainland and Zanzibar), the interquartile range

for adult QAACT prices was very narrow, with most observations equal or close to the

recommended retail pricesRRP. The other three pilots with RRPsrecommended retail

prices (Niger, Nigeria, and Uganda) experienced various difficulties with implementing

these supporting interventions, and QAACT prices there were far more variable and

exceeded RRPsrecommended retail prices (9).[9] This suggests that if

18


AU: If you replace "adult packs" above in response to questions there, please replace the term here, too.


RRPsrecommended retail prices are usedutiliszed, both their level and their sustained

promotion about them are critical.

RRPsRecommended retail prices were set through a consultative process in

each country with the aim of reflecting the costs of distributing medicines and

reasonable markups at different levels of distribution. The persistence of high QAACT

prices relative to RRPsrecommended retail prices in Niger, Nigeria, and particularly

Uganda could therefore imply that actors in the distribution chain actors may have been

earning super-normal profits on these medicines, if the RRPsrecommended retail prices

there wereare a true reflection of costs. However, even where QAACT prices were

equal to RRPsrecommended retail prices, there may still be scope for further price

reductions, becauseas private- sector actors may have exaggerated their costs during

the consultation processes.

A second way to ascertain whether there are opportunities to reduce prices paid

by patients for subsidizedsed medicines is to compare end-line QAACT retail prices and

markups for the preferred therapy with those for non-artemisinin therapies. Non-

artemisinin therapies are appropriate for this comparison, because they were widely and

cheaply available in the private for-profit sectors of the AMFm pilots and elsewhere (8,

15),[8,15] which indicates that retailers and wholesalers are willing to sell antimalarial

medicines at these prices.

At end line, median quality-assured ACT QAACT prices were less than or equal

to median prices for non-artemisinin therapies in only two pilots, Kenya and Tanzania

mainland. Relative and absolute retail markups on QAACTsquality-assured

ACTsartemisinin combination therapies were also lower than those on non-artemisinin

19


therapies in these pilots. This suggests that there may not be much further scope for

price reduction in these settings, but that further reductions may be possible in other

pilots.

However, in all pilots but Zanzibar and Uganda, relative retail markups on

QAACTsquality-assured ACTsartemisinin combination therapies were within the range

of regulated private retail markups permitted in countries in the African region that

regulate pharmaceutical markups (23).[23] On this basis, relative retail markups on

subsidizedco-paid ACTs may be considered reasonable everywhere except inbut

Uganda and Zanzibar.

The wholesale prices faced by retailers are a limiting factor to further price

reductions. Even though absolute retail markups on QAACTsquality-assured

ACTsartemisinin combination therapies exceeded those on non-artemisinin therapies

everywhere but Kenya and Tanzania mainland, if retailers were to reduce absolute

markups on QAACTssuch therapies to the same levels as those on non-artemisinin

therapies, QAACT prices for quality-assured ACTs would still exceed the prices foron

non-artemisinin therapies. Indeed, even in the highly improbable circumstance of

retailers charging no markups on QAACTsquality-assured ACTsartemisinin combination

therapies, QAACTssuch therapies would still be more expensive than non-artemisinin

therapies everywhere except inbut Kenya, Tanzania, mainland and Uganda. This

means that for end users' prices for quality-assured ACTs to reach the levels of non-

artemisinin therapy prices, private- sector wholesalers and distributors operating at

higher levels of the distribution chain would need to lower their selling prices in order for

20


the prices paid by end users for QAACTsquality-assured artemisinin combination

therapies to reach the levels of non-artemisinin therapies.

Possible Explanations For High Wholesale Prices

While iIt is not possible to determine definitively why wholesale prices for

subsidizedsed medicines remained high in some settings. However, the literature

suggests several hypotheses. Limited pass-through from wholesalers might reflect the

high costs of distributing antimalarials to some areas. It could also reflect the nature of

the supply chain. One; a six- country study found that antimalarials can move across as

many asup to four to six levels between manufacturer and retailer, with markups being

added at each level (24).[24]

Economic theory further suggests that the total value of a subsidy will not be

transmitted to the end -user unless there is perfect competition. Moreover, even under

perfect competition, the reduction of end -users' prices is dependent on the relative

elasticities of supply and demand (25).[25]

There are several reasons why antimalarial markets might be imperfectly

competitive. For example, barriers to entry are created by regulatory requirements for

opening and running a pharmaceutical retail or wholesale business, antimalarials are

differentiated goods, and there may be a limited number of wholesalers or retailers

supplying certain areas. Since 2012, reduced supplies of subsidized ACTs resulting

from order rationing by the Global Fund may have increased opportunities for suppliers

to exert monopoly power in some of the AMFm pilots (7).[7]

Policy Implications And Conclusion

21


AUS: please define “differentiated goods” here for our readers


AUS: Please define or explain “pass-through” for our readers


AU: Each section, such as this one ("Discussion") can have either no subsections or multiple subsections. There was only one subsection ("Opportunities For Further Price Reductions"). Thus we added another subheading here, to create a second subsection. If you prefer to divide the "Discussion" section differently, please provide new subheading(s).


OurThese findings demonstrate that it is possible to achieve a large reduction in

retail prices in a relatively short time period through the application of a supranational

subsidy ofto commodities essential to improving public health. Concerns about the

distributorsers capturing subsidies may be warranted ability of distribution chain actors

to capture the subsidy seem to have been warranted in some pilots, but not in others.

For RRPsrecommended retail prices to be an effective mechanism for controlling

market power, great care must be taken in setting the RRP price level and

communicating it to a wide audience. Pricing and markup decisions throughout the

supply chain will also influence final prices, and limit the transmission of the subsidy to

the end -users—the - patients.

The findings that subsidies administered in the private sector may not reduce

prices to the extent expected could be relevant for other health commodities that

havewith similar characteristics similar similar to malaria treatment—that is, , in terms of

being-commodities that are highly valued by consumers and frequently purchased in the

retail sector in low- and middle-income countries without consultingation with a health

worker.. Examples of products with these characteristics that are common in low- and

middle-income countries include These commodities include(such as treatments for

pneumonia, diarrhea,diarrhoea or sexually transmitted infections;; , condoms; and

insecticide- treated bed nets).

These products have all been previously been subsidizedsed in the private

sector through various mechanisms in the past (26, 27).[26,27] Given the role of the

private sector in their provision of these commodities and the importance of increasing

access to themse commodities and their low coverage, subsidies administered through

22


AUs: expected by policymakers? By whom? Please clarify and further explain this sentence – in the first sentence of the previous paragraph, the conclusion is that the program achieves great success; in this sentence, you seem to be saying that the program is not as successful as anticipated.


AUS: Is the added wording ok? If not, please revise


the private sector are a potential viable strategy for improving access to these priority

public health interventions.

ReferencesNotes

1. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. Global

and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a

systematic analysis for the Global Burden of Disease Study 2010. Lancet.

2012;380(9859):2095–-128.

2. Sabot OJ, Mwita A, Cohen JM, Ipuge Y, Gordon M, Bishop D, et al. Piloting the

global subsidy: the impact of subsidized artemisinin-based combination therapies

distributed through private drug shops in rural Tanzania. PloS oOne. 2009;4(9):e6857.

3. Rao VB, Schellenberg D, Ghani AC. Overcoming health systems barriers to

successful malaria treatment. Trends in pParasitology. 2013;29(4):164–-80.

4. White NJ, Nosten F, Looareesuwan S, Watkins WM, Marsh K, Snow RW, et al.

Averting a malaria disaster. Lancet. 1999;353(9168):1965–-7.

5. Global Fund to Fight AIDS, Tuberculosis, and Malaria. Use of a private sector co-

payment mechanism to improve access to ACTs in the new funding model: Information

note. Geneva: Global Fund; to Fight AIDS, Tuberculosis and Malaria, 2013.

23


AU: The Notes have been edited to conform to Health Affairs' standards and style. The accuracy of the information in the Notes is the responsibility of the authors. Please verify that all the information in the Notes is correct. Whatever changes you need to make in the Notes must be visible in track change, within our edited Notes. If you need to add or reorder notes and are unable to do so with track change on, please either give clear instructions about the necessary addition or reordering in balloon comments in the manuscript or contact a Health Affairs copy editor, rather than creating a new notes list.


6. Adeyi O, Atun R. Universal access to malaria medicines: innovation in financing

and delivery. Lancet. 2010;376(9755):1869–-71.

7. AMFm Independent Evaluation Team. Independent evaluation of phase 1 of the

Affordable Medicines Faciity– - malaria (AMFm) phase 1: Mmulti-country indpendent

evaluation report [Internet]. Calverton (MD): ICF International,d and London School of

Hygiene and Tropical Medicine;, 2012 Sep 28 [cited 2014 Jul 14]. Available from:

http://aphrc.sprintwebhosts.com/wp-content/uploads/2013/11/Independent-Evaluation-

of-Phase-1-of-the-Affordable-Medicines-Facility-malaria-AMFm_Multi-Country-

Independent-Evaluation-Report.pdf

8. Tougher S, ACTwatch Group, Ye Y, Amuasi JH, Kourgueni IA, Thomson R, et al.

Effect of the Affordable Medicines Facility—--malaria (AMFm) on the availability, price,

and market share of quality-assured artemisinin-based combination therapies in seven

countries: a before-and-after analysis of outlet survey data. Lancet.

2012;380(9857):1916–-26.

9. Willey BA, Tougher S, Ye Y, ACTwatch Group A, Mann AG, Thomson R, et al.

Communicating the AMFm message: exploring the effect of communications and

training interventions on private for-profit awareness and knowledge related to a multi-

country antimalarial subsidy intervention. Malaria Jjournal. 2014;13(46).

24


PMG: No page number.


PMG: Fourth author is "ACTwatch Group." Authors' names must appear in the PDF as they appear here.


AU: Please disregard this and subsequent comments for our formatter. PMG: Second author is "ACTwatch Group." Authors' names must appear in the PDF as they appear here.


AU: Author, title per the item we found through a Google search.


10. AMFm Independent Evaluation Team. Independent evaluation of Pphase 1 of the

Affordable Medicines Facility— – malaria (AMFm): multi-country independent evaluation

report: final report: Ssupplementary report on ACT use based on household surveys.

Calverton (MD): ICF International and London School of Hygiene and Tropical

Medicine;, 2012 Nov 10.

11. Clinton Health Access InitiativeCHAI. Price subsidies increase the use of private

sector ACTs: evidence from a systematic review [Internet]. New York (NY): CHAIClinton

Health Access Initiative,; 2012 [cited 2014 Jul 14]. Available from:

http://www.clintonhealthaccess.org/files/ACT_Usage_1.1.pdf

12. Cohen JL, Yadav P, Moucheraud C, Alphs S, Larson PS, Arkedis J, et al. Do

price subsidies on artemisinin combination therapy for malaria increase household

use?: eEvidence from a repeated cross-sectional study in remote regions of Tanzania.

PloS oOne. 2013;8(7):e70713.

13. Arrow KJ, Panosian C, Gelband H, editors. Saving lives, buying time: economics

of malaria drugs in an age of resistance. Arrow KJ, Panosian C, Gelband H, editors:

Washington (DC): National Academies Press; 2004.

14. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M, et al.

Developing and evaluating complex interventions: the new Medical Research Council

guidance. BMJmj. 2008;337:a1655.

25


PMG: No issue number.


AU: Changes per http://www.google.com/url?sa=t&rct=j&q=&esrc=s&frm=1&source=web&cd=2&ved=0CCMQFjAB&url=http%3A%2F%2Fwww.theglobalfund.org%2Fdocuments%2Famfm%2FAMFm_Phase1IndependentEvaluationSuppReportHouseholdSurveysNovember2012_Report_en%2F&ei=GSnEU7P6HMeKyASz1oHABw&usg=AFQjCNH9WgbGGRfbCZ3qh4VmUnGfYOzmiQ&bvm=bv.70810081,d.aWw


15. O'Connell KA, Gatakaa H, Poyer S, Njogu J, Evance I, Munroe E, et al. Got

ACTs? Availability, price, market share and provider knowledge of anti-malarial

medicines in public and private sector outlets in six malaria-endemic countries. Malaria

Jjournal. 2011;10:326.

16. To access the Appendix, click on the Appendix link in the box to the right of the

article online.

17. Global Fund to Fight AIDS, Tuberculosis, and Malaria. Global fund quality

assurance policy. Geneva: Global Fund to Fight AIDS, Tuberculosis and Malaria,; 2010.

18. International Monetary Fund. International financial statistics. Washington, (DC):

International Monetary Fund,; 2012.

19. OANDA Corporation. Historical exchange rates [Internet]. New York (NY):

OANDA Corporation; [cited 2014 Jul 14]2013. Available from:

http://www.oanda.com/currency/historical-rates/.

20. Shewchuk T, O'Connell KA, Goodman C, Hanson K, Chapman S, Chavasse D.

The ACTwatch project: methods to describe anti-malarial markets in seven countries.

Malaria Jjournal. 2011;10:325.

21. Tougher S, Ye Y, Goodman C, Arnold F, Hanson K. Evaluation of the Affordable

Medicines Facility—--malaria— - Aauthors' reply. Lancet. 2013;381(9872):1095–-6.

26




AU: The Appendix will be prepared for posting online after the article’s preparation for publication has been completed. Please be sure that the appendix you have submitted is complete, contains components that correspond to those you have specifically called out in the text, shows no change tracking, and is formatted the way you want it to be. Health Affairs editors do not edit or format appendices. All we do is add a citation to the article and make PDFs for online posting. Any errors or formatting problems are solely the responsibility of the authors. Please verify that you understand this by posting a comment here.




22. Health Action International. Retail prices of ACTs co-paid by the AMFm and other

antimalarial medicines: Ghana, Kenya, Madagascar, Nigeria, Tanzania and Uganda.

[please provide city and country]: Health Action International;, 2012.

23. Ballell D. The regulation of mark-ups in the pharmaceutical supply chain

[Internet]. Geneva: World Health Organization and Health Action International;, 2011

May [cited 2014 Jul 14]Contract No.: (Working Paper No. 3). Available from:

http://www.haiweb.org/medicineprices/05062011/Mark-ups%20final%20May2011.pdf

24. Palafox B, Patouillard E, Tougher S, Goodman C, Hanson K, Klienschmidt I, et

al. Understanding private sector antimalarial distribution chains: a cross-sectional mixed

methods study in six malaria-endemic countries. PloS oOne. 2014;9(4):e93763.

25. Kotlikoff LJ, LH. Summers LH. Tax incidence. In: Auerbachj AJ, Feldstein M,

editors. Handbook of Ppublic Eeconomics. Vol. 2. Amsterdam: Elsevier Science; 19878.

2:1043–92.

26. Taylor TA, Yadav P. Subsidizsing the distribution channel: Ddonor funding to

improve the availability of products with positive externalities [Internet]. Unpublished

paper.Berkeley, CA.: Haas School of Business, University of California Berkeley, 2011

Jan [cited 2014 Jul 14]. Available from:

http://haas.berkeley.edu/faculty/papers/taylor_subsidizing.pdf

27


AU: Changes per Library of Congress online catalog and http://www.kotlikoff.net/sites/default/files/Tax%20Incidence.pdf


AU: Author's name per item we found online through a Google search.


AU: Please provide city and country for Health Action International. If this item is online, please also provide the URL.


27. Schäaferhoff M, Yamey G. Estimating benchmarks of success in the Affordable

Medicines Facility— - malaria (AMFm) Pphase 1 [Internet]. San Francisco (CA):

University of California, San Francisco;, 2011 Jan 14 [cited 2014 Jul 14]. Available for

download from: http://www.seekdevelopment.org/en/publications

Exhibit Llist

Exhibit 1 (figure)Caption:Title: Median Price Per Adult Equivalent Treatment Dose (AETD) Of Antimalarial Treatments In Private For-Profit Outlets At Baseline And End lLine, (2010 US Dollars$ Equivalent)

Source/Notes: SOURCEource: aAuthors’ aanalysis.NOTESotes: nAT= non-artemisinin

therapies. QAACT=quality-assured artemisinin-based combination therapy. The

whiskers show the interquartile range (IQR) for price. The asterisks denote the p- value

from a two-sided Wilcoxon Rrank-s Sum test of no difference between the baseline and

end-line distributions for each antimalarial category. *=p-value <0.05, **=p-value<0.01

and ***=p-value<0.0001. pP- values are not presented for Zanzibar (a semi-

autonomous part of Tanzania) because a complete census of outlets there was

conducteddone. Tanzania includes results are for the mainland only. nAT is= non--

artemisinin therapies. QAACT is =quality-assured artemisinin-based combination

therapy. **p < 0.05 ***p < 0.01 ****p < 0.001

*=p-value <0.05, **=p-value<0.01 and ***=p-value<0.0001.Exhibit 2 (figure)CaptionTitle: Median Price In Relation ToAnd IQR (Y-Axis) Plotted Against Theinter-

Quartile Ranges Compared With Recommended Retail Selling Price (X-Axis) Per Adult

28


AU: Added to be consistent with Exhibit 3.


AU: HA uses the standard asterisks system shown below. We will change your single asterisks to double asterisks, your double asterisks to triple asterisks, and your triple asterisks to quadruple asterisks.* p < 0.10 ** p < 0.05 *** p < 0.01 **** p < 0.001We will probably change "base" to "baseline" and "end" to "end line" in the key. We will probably label the vertical axis "2010 US dollars."


AU: Please also explain in the exhibit notes why there are no p values for Uganda.


AU: We need at least this as source information. If possible, please add to the source information what you analyzed. For example, if you analyzed something that is already listed in an endnote, add the number of the note. If you analyzed an online or print item, please provide here all the information that would be required for a note (see the edited notes as models).


AU: All of the exhibits are figures and will be available for your review with your galley proofs. Please respond now to any comments here about the figures.


Equivalent Treatment For An Adult Pack Of Co-PaidsSubsidized Quality-Assured

Artemisinin-Based Combination Therapies (ACTcts), (2010 US Dollars$ Equivalent)

Source: authors’ analysis

Source/Notes: SOURCE Authors’ analysis. NOTESotes: The dotted line demarcates

the line of equality (dotted line) is where the median price is equal to the recommended

retail prices. The whiskers show the interquartile range forof the price offor an adult

equivalent treatment of quality-assured ACTQAACT treatment. In cases wWhere a

whisker is not visible (Kenya, Zanzibar, Tanzania, and Ghana), the median is equal to

the 25thth and/or 75thth percentile or bothis equal to the median. Zanzibar is a semi-

autonomous part of Tanzania. Tanzania results are for the mainland only. Madagascar

is not shown since it did not have a recommended retail price for AMFm co-paid

antimalarials that received subsidies from the Affordable Medicines Facility—malaria

initiative.

Exhibit 3 (figure)Caption:Title: Median Absolute Markup Between Retail Purchase Price And Retail Selling Price Per Adult Equivalent Treatment DoseAETD Of Antimalarial Treatments In Private For-Profit Outlets At Baseline And End lLine, (2010 US Dollars$ Equivalent)


Source/Notes: SOURCE Authors’ analysis. NOTESotes: nAT= non-artemisinin

therapies. QAACT=quality-assured artemisinin-based combination therapy. The

whiskers show the interquartile range (IQR) for absolute retail markups. The asterisks

denote the p- value from a two-sided Wilcoxon Rrank-s Sum test of no difference

29


AU: As explained above, we need at least this as source information. If possible, please add to the source information what you analyzed. For example, if you analyzed something that is already listed in an endnote, add the number of the note. If you analyzed an online or print item, please provide here all the information that would be required for a note (see the edited notes as models).


AU: Please provide a new version of the figure that clearly indicates which bars the asterisks now floating well above the bars belong to, and that includes country names along the horizontal axis (Ghana, Kenya, Madagascar, Niger, Nigeria, Tanzania, Uganda, and Zanzibar in that order, based on the data you provided).


AU: Please verify that the edited version of the exhibit notes is correct, or reword as needed to clarify the point.


AU: If the final version of the figure uses something other than a dotted line, we'll reword here as needed. We'll omit "mainland" after "Tanzania" in the figure because we added the explanation about Zanzibar and Tanzania to the exhibit notes.




AU: Please add here some time information for these prices. For example, "At Baseline."


AU: Reworded to be consistent with Exhibit 1. If "adult pack" must be restored here and in the exhibit notes, please briefly explain in those notes the relationship between the dose and the pack.


between the baseline and end-line distributions for each antimalarial category. p values

are not presented for Zanzibar (a semi-autonomous part of Tanzania) because a

complete census of outlets there was conducted. Tanzania results are for the mainland

only. nAT is non-artemisinin therapies. QAACT is quality-assured artemisinin-based

combination therapy. **p < 0.05 ***p < 0.01 ****p < 0.001

*=p-value <0.05, **=p-value<0.01 and ***=p-value<0.0001. P-values are not presented for Zanzibar because a complete census of outlets was done. Tanzania includes results for the mainland only.

Exhibit 4 (figure)Caption:Title: Total Markup From Importer Price To Retail Selling Price Per AETD And Median Absolute Retail Markup sPer Adult Equivalent Treatment Dose Per AETD For Amfm Co-Paidsubsidized QaactsqQuality-Assured Artemisinin Combination Therapies (ACTs) Subsidized By The Affordable Medicines Facility—Malaria Initiative In Private For-Profit Outlets At End lLine, (2010 US Dollars$ Equivalent)


Source/Notes: SOURCE Authors’ analysis. NOTESotes: Total markup is the markup

between the mean importer's price and the retail selling price. Median absolute markup

is the markup between retail purchase price and retail selling price. The Wwhiskers

show the interquartile range (IQR). Mean importer prices, which was used to calculate

total markups, wereas missing for 35% percent of co-paid QAACT observations of

subsidized quality-assured ACT sales in Niger; 3% percent of observations in Kenya,

Tanzania mainland, and Zanzibar; and less than 1% percent of observations in Ghana,

Madagascar, Nigeria, and Uganda. Zanzibar is a semi-autonomous part of Tanzania.

Tanzania results are for the mainland only.

Acknowledgment

30


AU: If any of the authors is a government employee, please verify that no organizational disclaimer is needed or, if such a disclaimer is needed, provide the required wording.


AU: Please verify that the edited version of the exhibit notes is correct, or reword as needed to clarify the point.


AU: We'll omit "mainland" after "Tanzania" in the figure because we added the explanation about Zanzibar and Tanzania to the exhibit notes.




AU: HA uses the standard asterisks system shown below. We will change your single asterisks to double asterisks, your double asterisks to triple asterisks, and your triple asterisks to quadruple asterisks.* p < 0.10 ** p < 0.05 *** p < 0.01 **** p < 0.001We will probably change "base" to "baseline" and "end" to "end line" in the key. We will probably label the vertical axis "2010 US dollars."


AU: Please also explain in the exhibit notes why there are no p values for Uganda.


A previous version of Tthis articlemanuscript was presented at the Private Sector in

Health Symposium at the 9Ninth World Congress on Health Economics, [please provide

city and country], on July 6th, 2013. The AMFm Independent Evaluation was funded by

the Global Fund to Fight AIDS, Tuberculosis, and Malaria (Grant No. 058992), with

support from the Bill & Melinda Gates Foundation for ACTwatch Central and ACTwatch

surveys in three countries (#058992). The authors thank the large number of people

who assisted with data analysis, data processing, and primary data collection. The

authors also thank the members of the ACTwatch Group: [please provide]. The

evaluation was funded by the Global Fund to Fight AIDS, Tuberculosis and Malaria with

the support from the Bill & Melinda Gates Foundation for ACTwatch Central and

ACTwatch surveys in three countries (#058992).

Bios for 2014-0104_Tougher

Bio 1: Sarah Tougher ([email protected]) is a research fellow in the Department of Global Health and Development, London School of Hygiene and Tropical Medicine, in the United Kingdom.Bio 2: Andrea G. Mann is a lecturer in the Department of Infectious Disease Epidemiology, London School of Hygiene and Tropical Medicine.Bio 3: The ACTwatch Group is recognized in an acknowledgment at the end of the article.Bio 4: Yazoume Ye is a senior infectious diseases specialist at ICF International, in Rockville, Maryland.Bio 5: Idrissa A. Kourgueni is general director of the Centre International d'Etudes et de Recherches sur les Populations Africaines and of the Institut National de la Statistique, both in Niamey, Nigeria.Bio 6: Rebecca Thomson is a research fellow at the Ifakara Health Institute, in Dar es Salaam, Tanzania.Bio 7: John H. Amuasi is a doctoral candidate in the School of Public Health, University of Minnesota, in Minneapolis, and a researcher at the Komfo Anokye Teaching Hospital (KATH), in Kumasi, Ghana.Bio 8: Ruilin Ren is a senior technical specialist at ICF International.Bio 9: Barbara A. Willey is a lecturer in the Department of Infectious Disease Epidemiology, London School of Hygiene and Tropical Medicine.

31


AU: This is how we have handled previous groups as authors. See related request in the acknowledgments.


AU: To ensure accuracy, Health Affairs requests that you carefully check all author bio information that will appear on p. 1. The accuracy of this information is the responsibility of the authors. Please verify that all the information is correct and consistent. If revisions are necessary, please provide.


AU: The omitted sentence seemed only to repeat what you say above.


AU: This is how we've acknowledged previous groups as authors. Please list the names of the group members where requested, unless there are dozens of members—in which case omit ": [please provide]."


AU: If this isn't what 058992 is, or if the number isn't related to the Global Fund, please revise as needed to clarify the point.


AU: Please add city and country name for the symposium where requested.


Bio 10: Daniel Ansong is a senior lecturer in the Department of Child Health, School of Medical Sciences, Kwame Nkrumah University of Science and Technology, and deputy director of the Research and Development Unit, KATH.Bio 11: Katia Bruxvoort is a research fellow at the Ifakara Health Institute and at the London School of Hygiene and Tropical Medicine.Bio 12: Graciela Diap is a medical coordinator at the Drugs for Neglected Diseases initiative, in Geneva, Switzerland.Bio 13: Charles Festo is a researcher at the Ifakara Health Institute.Bio 14: Boniface Johanes is a researcher at the Ifakara Health Institute.Bio 15: Admirabilis Kalolella a researcher at the Ifakara Health Institute.

32

researchonline.lshtm.ac.ukresearchonline.lshtm.ac.uk/1918305/1/tougher_edited_lw.docx · web...

Documents