A

Bs(iMaamctCpe©

K

I

up2duS

hB

n

0d

Available online at www.sciencedirect.com

International Journal of Drug Policy 19S (2008) S74–S79

Short report

Managing information: Using systematic data collection to estimateprocess and impact indicators related to harm reduction

services in Myanmar

Simon Baldwin a,∗, Neil Boisen b, Robert Power a

a Burnet Institute, Centre for Harm Reduction, Melbourne, Australiab Consultant Bangkok, Thailand

Received 11 September 2007; received in revised form 3 December 2007; accepted 8 December 2007

bstract

ackground: Discussion about coverage has primarily focused on answering the question: what level of coverage is required to reduce thepread of HIV among people who inject drugs? This paper documents the process involved in designing a Monitoring Information SystemMIS) that provides a tool to estimate coverage, frequency of contacts as well as provides a mechanism for correlating these data with changesn risk behaviour among the surveyed population.

ethods: The system uses paper and pencil data collection forms to record information about the type and location of a contact. Informationbout the content of the contact such as the services, equipment or education that is delivered is also collected. This data is then entered intocomputer program that manages the information and allows for simple standardised reports to be generated. The reports provide a simpleechanism for analysing process indicators such as the number and frequency of contacts, where the contact occurred as well as what the

ontact consisted of (i.e. education content or distribution of equipment). The system also allows correlations to be made between exposureo services and changes in behaviour thus providing a mechanism for assessing impact indicators.

onclusion and discussion: We present a brief description of the Monitoring Information System, its structure and functions and encourageractitioners to consider the importance of adopting standardised monitoring systems to measure coverage. We also explore some potentialthical limitations around using the system.

2008 Elsevier B.V. All rights reserved.

WcatnciU

eywords: Monitoring and evaluation; Coverage; Harm reduction

ntroduction/background

Considerable attention has focused on the need to “scale-p” HIV prevention services to reach a greater number ofeople who inject drugs (Bedell, 2007; Vickerman et al.,006). Empirical evidence exists to show that when serviceso not reach (or “cover”) the required proportion of a pop-lation, HIV continues to spread (de la Fuente et al., 2006;harma et al., 2007; Vickerman et al., 2006).

Defining appropriate and effective coverage remains aotly contested issue within the field (Bluthenthal et al., 2004;urrows, 2006; Sharma et al., 2007; Stimson et al., 2005).

∗ Corresponding author. Tel.: +61 39282 2169; fax: +61 39282 2144.E-mail addresses: simonb@burnet.edu.au (S. Baldwin),

boisen@yahoo.com (N. Boisen), Robert@burnet.edu.au (R. Power).

ent(dbo

955-3959/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.drugpo.2007.12.005

hile definitions of coverage differ, the basic components ofoverage include contacts with the client group (Aceijas etl., 2007; Irwin et al., 2006) as well as a method to measurehe efficacy of the contact (Sharma et al., 2007). Up untilow, the coverage debate has largely focused on how oftenlients need to be contacted, and with what types of services,n order to reduce HIV transmission (Sharma et al., 2007;NAIDS, 2006; Zheluk & Burrows, 2006).Alongside interventions and activities to improve the cov-

rage of harm reduction targeting injecting drug use, weeed to develop information systems that can monitor andrack progress. Building on the work of Gray and Burrows

Gray, 2006) in Central Asia this paper will describe andiscuss a Management Information System (MIS) that haseen designed to measure, among other things, the numberf clients accessing harm reduction services. Other key func-

urnal of

titWdwo

ttca(tbb

ocfttMsao

M

P

W

thi

tftcninassa

airu

e

tltplu

iPtCAOeoppMfifi

actsmttMtowl

tos

C

omptesaiw

A

S. Baldwin et al. / International Jo

ions of the MIS include systematic monitoring process andmpact indicators, the ability to easily merge data in a cen-ralise database and to produce informative reports easily.

hile this paper is based on a process of consultation andevelopment in Myanmar, the authors present these ideasith the goal of encouraging similar systems to develop inther parts of the world.

The theoretical underpinnings of the MIS suggest that ifhe target population size is known, then a monitoring systemhat can track individual contacts can be used to estimate bothoverage of clients contacted as a percent of the population,s well as frequency of contact made within a populationBurrows, 2005; Gray, 2006). Further, when this informa-ion is correlated with other data (behavioural as well asiological measures) the efficacy of client contacts can alsoe estimated.

This paper aims to discuss the process involved in devel-ping the MIS. Consideration will be given to the operatingontext within Myanmar and potential ethical implicationsor the proposed data collection system. We will also presenthe major functions of the MIS developed for Myanmar andhe steps taken to design the system. It is our belief that the

IS will have broader applicability in monitoring and mea-uring the coverage and scale-up of harm reduction services,s well as supplying descriptive data on the range and qualityf services.

ethod

rocess of consultation

hy the MIS was developedThe setting in Myanmar reflects a context common

hroughout the world, with several partners implementingarm reduction programs in the absence of a coordinatedntegrated monitoring and evaluation system.

An initial review of monitoring and evaluation (M&E) sys-ems in Myanmar conducted by the Burnet Institute’s Centreor Harm Reduction revealed that the quality of M&E sys-ems across harm reduction project sites varied and that dataollection was not standardised. It also noted that many part-ers did not use the data they were collecting to evaluate andmprove their services. The review also noted that there waso apparent mechanism for consolidating data collected atnational level. It was also observed that the existing M&E

ystems focused on basic process indicators and did not mea-ure coverage of services, frequency of contact, nor identifynd correct for duplicate contacts.

Informed by this review, the Burnet Institute commencedn extensive collaborative process of consultation with lead-ng local implementation partners to identify the optimal

equirements and functionalities of a monitoring and eval-ation system for harm reduction.

This paper presents a case study, using Myanmar as thexample, in which several partner agencies worked together

rie

Drug Policy 19S (2008) S74–S79 S75

o design a coordinated system capable of measuring the col-ective coverage of harm reduction services. The MIS followshe principles of the UNAIDS Three Ones Policy, in that itrovides a standardised monitoring system at the nationalevel, while also providing the flexibility at the site level fornique data to be collected and analysed.

A number of local and international agencies were activen providing harm reduction services in Myanmar at the time.artner agencies collaborating in the MIS design included

he Asian Harm Reduction Network, Medicines Du Monde,ARE, Myanmar Anti Narcotics Association, AusAID’ssia Regional HIV Project Sites and the United Nationsffice on Drugs and Crime. An initial workshop raised inter-

st in, and support for, the project and established consensusn the scope of the MIS through a review of current “bestractices” and a mapping of existing local harm reductionolicies and procedures. Subsequent workshops with local&E staff from collaborating organisations refined the MIS

unctional requirements, data collection instruments and usernterface design. Finally, prototypes and a beta version wereeld tested before final revision and initial release of the MIS.

Through this process, partners identified a series of char-cteristics that were considered essential in the new dataollection system. Partners required that the MIS be ableo generate reports to assist programs to better understandervice delivery; that the MIS was flexible enough to accom-odate the diverse needs of multiple end-user organisation;

hat the system collects data for a constellation of harm reduc-ion services, including Outreach, Drop in Centre attendance,

edical services, Needle and syringe programs, and Educa-ion and behaviour change interventions. A final requirementf the MIS was that it would be easy to use by local staffith limited computer expertise and varying levels of English

anguage skills.A number of guiding principles based on systems design

heory (Briggs and Nunamaker, 2004) informed the devel-pment process for the MIS and these are described in theub-sections below.

larity of vision and goalIdeally, an MIS fulfils several functions: (1) feedback on

perational processes and inputs for daily project manage-ent, (2) monitoring of program deliverables and monitoring

rocess indicators and (3) evaluation of programme effec-iveness and impact indicators. These functions are not ofqual importance and it is necessary to prioritise them. Con-equently, primary emphasis should be put upon monitoringnd evaluation, with project management information lim-ted to those data that would otherwise be difficult to trackithout systematic data collection over time.

void needless complexity

As system complexity increases, user friendliness and data

eliability decreases. Functions and parameters that do notncrease the analytic capabilities of the system should beliminated.

S urnal of

T

ijs

rma

B

twbr

R

fMChoitaansiao

S

ksIaefoiu

UpkIura

gd

awbimfct

M

m

cceLcw

mCa

tObd

tCtatmc

tpf

S

dddr

76 S. Baldwin et al. / International Jo

he “Minimalist Approach”A guiding rule should be to only collect data that will

mpact on program quality. If the collection of data cannot beustified as to exactly how and why it would be used, then ithould not be incorporated into the system.

Limit data collection and reporting to the minimal dataequirements to fulfil the three MIS functions (i.e. processonitoring, impact evaluation and operational management)

s specified in the program’s M&E framework.

egin with output analysisOne way to achieve the “Minimalist” goal is to start from

he desired result and work backwards. Begin with what youant to know – your outputs or analytic reports – and workack from there to determine the minimal data collectionequirements needed to be generated.

esults

In the case of the MIS developed for Myanmar, several keyeatures were identified as necessary to achieving the stated

IS goals. These included using a Unique Identification (ID)ode that would allow data to be linked to an individual, andence would allow calculations to be made about frequencyf contact, as well as behaviour change over time, expressedn relation to exposure to services. Also, it was decided thathe MIS needed to be configured to allow optimal flexibilityt the site level, thus providing the flexibility for individualgencies to collect information that was not required at aational consolidated level. Finally, due to varying computerkills of the end users, it was decided that a bilingual usernterface was required along with a simple yet robust datanalysis and reporting function based on a pre-formatted setf modifiable report templates.

tandardised unique client identification codes

A client code that uniquely identifies an individual is theey to measuring frequency of client contact, while at theame time maintaining client confidentiality and anonymity.n addition, when the unique client ID system is standardisednd adopted by all implementing partners, it can be used toliminate double counting of individuals contacted by dif-erent organisations in a consolidated data set. In the casef Myanmar (as elsewhere) several programmes operatedn close proximity, so universal adoption of an agreed uponnique ID code was essential for accurate regional reporting.

Developing an appropriate coding scheme to serve as anique ID code remains a challenge for many monitoringrograms (Burrows, 2005; Gray, 2006). There are severaley features that need to be satisfied in choosing a unique

D code (Burrows et al., 2006). These include the need forniqueness, confidentiality and ease of recall. Programs thatequire a Unique ID code generally follow one of two mainpproaches: first they provide clients with an automatically

ts

a

Drug Policy 19S (2008) S74–S79

enerated (and usually sequential) ID code; or second theyevise a code based on a series of key questions.

The first approach is easier to implement, but has the dis-dvantage of being program-specific. The latter approach,hile more difficult to design, has the marked advantage ofeing universal or non-programme specific, with the samendividual contacted by different organisations being auto-

atically assigned the same Client ID. The MIS developedor Myanmar permits the use of either or both types of clientodes, thus allowing each organisation to gradually transitiono a system based on personal data.

odular MIS structure

Fig. 1 shows the major components of the Myanmar MISodular design and its corresponding functions.Data collection begins with basic pencil and paper data

ollection forms. Separate data collection forms are used toollect data. The information collected from these forms isntered into the Site Level Database. The role of the Siteevel Database is to store and manage all information that isollected at the site level. The Site Level Database is set upith internal checks to improve data integrity.The Site Level Database is organised into modules. These

odules (Registration, Contact, Services and Needle/Syringeollection) constitute the shell or structure of the databasend share a common screen design and functionality.

All sites in Myanmar have a stand alone Site Level MISo handle the data entry and analysis relevant to that site.nce data is entered into the Site Level MIS, reports cane generated using the report function (described in greateretail below).

The MIS Configuration Editor determines the informa-ion that can be entered into the Site Level Database. Theonfiguration Editor is managed at a national level and main-

ains a standard list of parameters which site-level users canctivate or deactivate depending on the services they offer atheir particular site. This allows flexibility between sites while

aintaining the standardisation required for meaningful dataonsolidation at a regional or national level.

Deployed at the regional/program and national levels ishe data consolidation module of the MIS. This is the focaloint where all data collected at the Local Site Level is sentor analysis at the regional, program or national level.

tandardisation and flexibility

As one of the goals of the system is the consolidation ofata at a central level, the standardisation of parameters andata field values was essential. This requirement for stan-ardisation needed to be balanced against the conflictingequirement for flexibility at the site level to accommodate

he varying data collection and operational practices of eachite.

Balance between these competing requirements waschieved in three ways. First, all partners collaborated in

S. Baldwin et al. / International Journal of Drug Policy 19S (2008) S74–S79 S77

betwee

dtboTmmal

S

efpdTm

lCN

tdcttmtt

Fig. 1. Information flow

etermining the standardised list of field values (List Items)o be incorporated in the system. Second, functionality wasuilt into each module allowing List Items to be turned on orff at the Local Site level in accordance with each site’s needs.hird, several additional “local” fields were included in theodules, the values of which can be independently deter-ined by each individual site in accordance with its needs,

nd which are not consolidated at the regional or nationalevel. Each of these features will be explained below.

tandardisationAs mentioned above, standardisation is important to

nsure that data collected at the local level can be meaning-ully consolidated at the central level. Through a collaborative

rocess involving all partners, consensus was establishedefining the essential data to be collected and centralised.his information was then organised into data collectionodules containing standardised Lists and Items.

F

te

n the MIS components.

The modules of the MIS are launched from a centralaunch window. The modules include Client Registration,lient Contacts, Medical and Social Services Module andeedle and Syringe Collection.Standardised Lists were then developed that covered all

he types of information the collaborating organisationseemed essential. For example, the Client Contacts Module isoncerned with measuring contacts with clients. Informationhat is managed in this module includes such factors as theype of contact, the products or information education com-

unication (IEC) material distributed during the contact andhe educational themes discussed during the contact. Each ofhese Lists is further broken down into a series of Items.

lexibilityFlexibility was build into the MIS in two ways. Firstly,

hrough the ability of Local Sites to activate and deactivateither entire modules or individual Items in Lists. Secondly,

S urnal of

bd

afioWrl

enntrg

inswawd

B

be

D

qapEAcai

C

pltsmvabi

iatistuaAic

giWiaMttst

tsipaihstf

fpctccidbtccoariI

78 S. Baldwin et al. / International Jo

y creating fields that allowed Local Sites to enter uniqueata that is not consolidated.

This level of flexibility allows individual sites to activateny item that is contained in the configuration list (see con-guration file). Further, only the activated items will appearn the corresponding screens of the Local Site Level MIS.hen a site decides to introduce another one of the IEC mate-

ials within the configuration, it simply activates it on theist.

The additional benefit of this configuration system is thease of updating the various lists in the MIS. If a new itemeeds to be added, an updated configuration file is dissemi-ated to each site and the new item automatically appears inhe local MIS interface. In this way, the MIS can ‘grow’ toeflect changes in the services that are being delivered on theround without the need for reprogramming.

An additional level of flexibility was achieved by provid-ng “local” fields that are not consolidated and, therefore, doot need to be standardised between all sites. For example,ome sites were interested in collecting field notes associatedith medical and social services within the MIS database

nd others were not. Each site can chose which type of data itishes to enter in the “local” fields without fear of impedingata consolidation.

ilingual interface

The MIS user interface is bilingual and can be switchedetween English and a single local language at any time fromach module’s main menu.

ata analysis and reporting

A major goal of the MIS was to enable each Local Site touickly and easily generate reports on its own data. This waschieved in the Myanmar MIS through the report function thatrovides easy access to a comprehensive set of pre-formattedxcel PivotTable reports and charts linked to the database.ny user with even an elementary knowledge of PivotTables

an modify the pre-formatted reports to perform additionalnalysis as needed. In addition, the MIS data can be importednto other statistical programs if needed.

onclusion

The Management Information System discussed in thisaper has been developed to standardise and simplify the col-ection, storage, management and analysis of key data relatingo harm reduction interventions. It has been designed to mea-ure process indicators relating to service delivery and as aechanism to measure the coverage of harm reduction ser-

ices, both in terms of the number of individuals reachednd frequency of contacts. When this data is correlated withehavioural and biological data, the MIS will assist in assess-ng impact indicators related to harm reduction programming.

ipdu

Drug Policy 19S (2008) S74–S79

The role of the MIS is not to replace other forms of mon-toring and evaluation. On the contrary, MIS should be seens a tool to complement a suite of monitoring and evalua-ion activities. As always, it is important to collect qualitativenformation about service delivery, such as client satisfactionurveys. We also need to continually access changes withinhe broader community, such as changes in patterns in drugse, community attitude and policy. Similarly, an MIS is notreplacement for research studies with randomised samples.ny findings based on data generated from an MIS must be

nterpreted cautiously and without generalising beyond thelient group.

Despite these words of caution, a well-designed MIS cano a long way in simplifying the monitoring of impact-levelndicators that are common to many harm reduction services.

ith a sufficient level of built-in flexibility, an MIS can evolven pace with the evolution of the programmes and be easilydapted to different settings. A well planned and designedIS will monitor and measure the coverage of harm reduc-

ion services, as well as supplying descriptive data on bothhe quantity of service delivered, as well as the impact theseervices have on influencing change in the population theyarget.

Considerable attention was given to the ethical implica-ions of using the MIS in Myanmar. While it is beyond thecope of this paper to assess drug use and HIV related harmn the country (Reid et al., 2006), HIV prevalence amongeople who use drugs is high (Aceijas et al., 2004) and over-ll coverage of services is poor. Further, drug use remainsllegal and punishable by 5 years imprisonment. Despite this,arm reduction services continue to improve in the country,mall-scale methadone maintenance therapy is available, andhere is a commitment within the Myanmar’s Strategic Planor HIV to reach more people who inject drugs.

A major concern in developing the MIS was the potentialor the information collected by the MIS to be used inap-ropriately by the authorities. A number of issues whereonsidered. Firstly, the potential for a client’s confidentialityo be broken and their records to be disclosed. However, it wasoncluded that given the non-identifiable nature of the uniqueode it was unlikely that it could be used to identify any onendividual. A second point was raised about the potential forata to be misused or interpreted. This position was counteredy the notion that current data collection focuses around theotal number of needles distributed, and that this figure isonsistently rolled out as an index of harm reduction suc-ess. However, and especially to the uninformed, the statisticften runs counter to the popular notion of reducing harmnd is more often associated with the perception that harmeduction services are encouraging drug use. A final ethicalssue was ensuring that the adoption of the MIS and uniqueD code did not provide barriers to clients entering or access-

ng services. While this was a major initial concern, severalartners conducted field trials with clients and found that itid not cause major issues. In fact, most organisations alreadysed some form of ID code. So for many, shifting to the new

urnal of

sn

tpwrdat

eos

C

A

aaToh

a

R

A

A

B

B

B

BB

B

d

G

I

R

S

S

U

V

Deficiency Syndrome, 42, 355–361.

S. Baldwin et al. / International Jo

ystem was a matter of adopting a new code, rather than aew protocol.

Finally, it was agreed by all harm reduction organisationshat the overriding principle for the MIS was that it should notrevent clients from accessing services. Indeed, the priorityas to provide people who use drugs with the services they

equire and not to collect data from them. A major focus of theata collection training completed by staff using the systemnd the accompany policy and procedures guide focused onhis principle.

It is hoped that the ideas presented in this paper willncourage others to collaborate on the design and devel-pments of monitoring systems that encourage increasedtandardisation and measurement of coverage.

onflict of interest

All authors confirm that there is no conflict of interest.

cknowledgements

The authors would like to thank staff from the partnergencies who have been involved in the design, field testingnd implementation for the Monitoring Information System.he authors would especially like to acknowledge the effortsf Dr. Hla Htay, Yin Yin Min and Dr. Bronwyn Wells whoave worked tirelessly on the MIS for the past few years.

The Monitoring Information System was funded throughgrant from the Fund for HIV AIDS Myanmar.

eferences

ceijas, C., Hickman, M., Donoghoe, M. C., Burrows, D., & Stuikyte,R. (2007). Access and coverage of needle and syringe programmes(NSP) in Central and Eastern Europe and Central Asia. Addiction, 102,1244–1250.

ceijas, C., Stimson, G., Hickman, M., Rhodes, T., & United NationsReference Group on HIV/AIDS Prevention and Care among IDU inDeveloping and Transitional Countries. (2004). Global overview ofinjecting drug use and HIV infection among injecting drug users. AIDS,18, 2295–2303.

Z

Drug Policy 19S (2008) S74–S79 S79

edell, R. (2007). The art and the science of scaling-up needle and syringeprogrammes. Addiction, 102(8), 1179–1180.

luthenthal, R. N., Malik, M. R., Grau, L. E., Singer, M., Marshall, P.,Heimer, R., et al. (2004). Sterile syringe access conditions and varia-tions in HIV risk among drug injectors in three cities. Addiction, 99,1136–1146.

riggs, R. O., & Nunamaker, J. (2004). Information systems design—Theoryand methodology. Journal of Management Information Systems, 20(Spe-cial Issue, 4), 5–8.

urrows, D. (2005). Unique identifier code. Kyrgyzstan: USAID.urrows, D. (2006). Rethinking coverage of needle exchange programs.

Substance Use & Misuse, 41(6), 1045–1048.urrows, D., Zheluk, A., Karmanova, G., Gray, R., Magdiyev, Y., Dairov, A.,

Abidjanova, N., Magkoev, V., Bolotbaeva, A., Ajoldasov, A., Bozrikova,T., Korneo, O., Alaxandrov, T., Turaev, B., et al. (2006). Unique identifiercode. Kyrgyzstan: USAID.

e la Fuente, L., Bravo, M. J., Toro, C., Brugal, M. T., Barrio, G., Soriano,V., Vallejo, F., Ballesta, R., & Project Itinere, G. (2006). Injecting andHIV prevalence among young heroin users in three Spanish cities andtheir association with the delayed implementation of harm reductionprogrammes. Journal of Epidemiology and Community Health, 60(6),537–542.

ray, R. (2006). Scaling up in Central Asia—Recent advances and somethoughts. In Proceedings of the APMG Satellite Session on coverage ofharm reduction programs at the 17th International conference on thereduction of drug related harm.

rwin, K., Karchevsky, E., Heimer, R., & Badrieva, L. (2006). Secondarysyringe exchange as a model for HIV Prevention Programs in the RussianFederation. Substance Use & Misuse, 41(6), 979–999.

eid, G., Devaney, M. L., Baldwin, S., et al. (2006). Drug production, traf-ficking and trade in Asia and Pacific Island countries. Drug and AlcoholReview, 25(6), 647–650.

harma, M., Burrows, D., & Bluthenthal, R. (2007). Coverage of HIV pre-vention programmes for injection drug users: Confusions, aspirations,definitions and ways forward. International Journal of Drug Policy,18(2),92–98.

timson, G. V., Hickman, M., Rhodes, T., Bastos, F., & Saidel, T. (2005).Methods for assessing HIV and HIV risk among IDUs and for evaluatinginterventions. International Journal of Drug Policy, 16(Suppl. 1), 7–20.

NAIDS. (2006). High coverage sites: HIV prevention among injecting drugusers in transitional and developing countries. Geneva: UNAIDS.

ickerman, P., Hickman, M., Rhodes, T., & Watts, C. (2006). Model projec-tions on the required coverage of syringe distribution to prevent HIVepidemics among injecting drug users. Journal of Acquired Immune

heluk, A., & Burrows, D. (2006). A spectrum of services approach. In Pro-ceedings of the APMG Satellite Session on coverage of harm reductionprograms at the 17th International conference on the reduction of drugrelated harm.