Rapidly Expanding Access to Care for hiv in Tanah Papua (reach) 2012 2016

2.8 Component 4: Operational research

Overarching objective: Generate and apply knowledge that contributes to achieving the program goal of increasing access to high quality HIV-related treatment in Tanah Papua

Sub-component objective - Operational research: Undertake research that generates reliable knowledge with the potential to contribute to the program goal.

Outcome: International standard operational research conducted and published

Sub-component objective - Local engagement: Increase the use of operational research data to achieve program goals

Outcome: Results of operational research incorporated into policy and delivery of HIV-related services in Tanah Papua, and in supply chain management nationally.

Why are we doing this?

REACH is essentially about technical assistance for service provision, not research. However, an operational research component is needed to generate reliable evidence on how to deliver these services most effectively in a geographically and politically fluid landscape where governance is weak and human resources are at a premium.

Operational research is also often referred to as implementation research because it focuses on how interventions work when they are implemented in the real world, rather than in ideal trial conditions. The lessons the program will generate about implementation and service delivery will be of use to service providers in fields other than HIV, and in contexts other than Tanah Papua. This learning can be generated at relatively low additional cost because the vast majority of the data that will be used in systematic studies under this program will be collected as part of the robust monitoring system implemented by REACH, (see section 3.4).

It is one thing to generate rigorous operational research; it is another thing to ensure that it results in rapid changes in policies and procedures, so that service delivery improves. REACH will design and conduct its operational research in ways that will maximise the likelihood that the results translate into better care for more Papuans during the life of this program.

How will we do it?

The first step in REACH's operational research activity, and the one most likely to ensure that results are used rapidly to improve program outcomes, is an inclusive process to set the operational research agenda.³⁸ The early inclusion of the MoH’s AIDS Sub Directorate, PHOs, DHOs, CoE, puskesmas and CHAI program managers will ensure that research questions are program-relevant and that those most likely to use the results know (and, we hope, care) about the research from the start. Where appropriate, members of communities among whom research will be conducted will be included in the agenda setting and research design discussions. AusAID will also be asked to contribute to discussions as an important end-user of research data.

This process means that it is not possible to determine at the design stage what questions will be answered in operational research. However an indicative list of the sorts of issues that may be addressed is included in Box 1 below. REACH is not planning to conduct randomised controlled trials. Study designs that maximise the use of routine data (including rigorous relationship mapping) will be favoured. The design of the program lends itself well to implementation research: regional autonomy has led to different approaches to health financing and service provision. By working through these varied systems REACH will be able to compare costs and effectiveness of different models and approaches to implementation. And because we are starting from scratch and working in a virtual vacuum in so many areas, changes over time are more easily attributable to interventions than might otherwise be the case.

Box 1: Indicative research questions

What are the major reasons for loss to follow-up among those started on ARV treatment?

Is adherence to cotrimoxazole a good indicator of adherence to ARVs, and if so, over what time period?

How do media campaigns, SMS broadcasts and public events compare in increasing the uptake of HIV testing?*

What contribution can cell-phones make to promoting adherence to ARVs in areas of universal cell phone ownership but patchy electricity and signal coverage?

Does provider-initiated testing increase gender differences in access to treatment?

Can outreach at football matches and other male gathering places reduce the gender disparity in access to HIV-related care?

Can we estimate the extent of duplicate reporting of the same HIV infections, and reduce them through a unique identifier system?

Is provision of HIV-related CST through a decentralised model of care, involving puskesmas linked to district hospitals, a cost-effective model in the Papuan provinces?

Are there differences in cost-effectiveness of HIV-related CST services between REACH sites that are have adapted the model in different ways?

Can innovative models of co-financing with existing district services (kabupaten induk) increase service provision in newly-established districts (kabupaten pemekaran)?

How representative are reported Health Facility Survey data of actual service availability?

Can church burial records be used to estimate changes in age-specific mortality in Papua and if so has young adult mortality increased in HIV-affected areas?

Does attendance by key policy-makers at regular analysis meetings affect the speed or direction of changes in program implementation?

Does treatment literacy training for church volunteers decrease loss to follow up in ARV treatment?

Does immediate ARV treatment of infected sex workers reduce ongoing transmission of HIV?*

Is distribution of resources for HIV more rational in districts where AIPD is active than in other districts?*

Are there fewer stock outs of drugs and reagents in a province where distribution is decentralised, compared to one with central distribution?

Is advocacy at the national, provincial or district level most likely to lead to changes in policy and practice around delivery of effective HIV-related treatment?

What proportion of infections in a highland area can be traced to commercial sex?*

* Examples of questions addressed by research that might be externally funded or co-funded with other programs, but embedded in the REACH program

The REACH operational research and monitoring manager will have primary responsibility for translating the research agenda into a program of high quality research studies. This role will include active guidance on research design, co-ordination of protocol development, ensuring approval by relevant ethics committees and quality control for research implementation. S/he will also be responsible for appropriate dissemination of results. The REACH provincial program analysts will oversee much of the day-to-day operational of research, and will be centrally involved in data analysis and interpretation, although CoE program analysts may take the lead on specific studies.

Operational research will be undertaken in partnership with other organisations as appropriate. MoH, PHOs, DHOs and the CoE will be the core partners in most studies. REACH will also partner with AIPD to answer questions around budgetary planning and its effect on service delivery in a decentralised health system. HCPI will be likely to play a key role in research related to adherence support and demand creation for services, in particular HIV testing and STI treatment. REACH will seek opportunities to embed within its knowledge generation program research funded by other sources (for example planned National Institutes of Health (NIH)/WHO research on antiretroviral therapy as an HIV prevention intervention). Although we do not propose to drive basic science research ourselves, we may form partnerships with other institutions in this area. REACH expects, for example, to collaborate with the Jayapura office of the MoH's NIHRD to genotype HIV samples collected during routine service provision in the Papuan highlands, providing potentially important information about the relationship between geography, ethnicity, risk and the spread of HIV. Phylogenetic analysis may, for example, provide an indication of the proportion of infections originating in commercial sex. This information should help to inform prevention programming, and may lead to better targeting of HIV testing. Embedded research will leverage REACH's contribution to knowledge creation at minimal cost to AusAID.

As appropriate, operational research data will be disaggregated by sex and analysed to determine if there are any gender based differences. Strategies to deal with gender-related findings will be developed.

Operational research is more likely to be discovered and taken seriously both locally and internationally if it is published in a peer reviewed journal. Planning for publication also often improves the quality of research in the field. Yet this type of research often falls at the publication hurdle simply because people focused primarily on service delivery do not have the time (and sometimes do not have the skills) to write papers for journals. From the third year of the program, when publishable results can realistically be expected, REACH plans to include on its team a professional medical editor to support and manage the publication process. Since REACH is committed to maximising data use, all research will be published in open access journals/formats.

Will it work?

In the past, operational research has been undervalued; academics have not considered it rigorous enough to be ’real science’, while program implementers think it takes up time that would be better spent delivering services. It's now apparent that research around what we need to do (basic science) is of little value if we do not know how to do it. In other words, implementation research is vital. But the second objection -- the opportunity cost of doing and publishing research -- remains a real one in the context of REACH. We believe that we will minimise that risk by:

• 
  Appointing a strong research manager. Finding an appropriate person to fill this post will be critical to the success of the operational research program
  
• 
  Having the program analysts responsible for research working with program implementers on a daily basis at the provincial and CoE levels.
  
• 
  Using simple research designs based on existing, properly staffed data collection systems.
  

The other major risk is that research will get done, but its results will not be used to improve HIV-related treatment in Papua. Since the most important users are going to be defining the questions they want answered and will be involved in ongoing discussion of the implications of emerging findings every six months at a minimum, we believe this likelihood is low.

Both the monitoring and the governance structures are designed to maximise the rapid use of emerging data in program implementation at local and national levels. The regular ‘live analysis’ meetings described in the monitoring section (see 3.4.2) are designed to ensure that the findings and implications of operational research will be quickly taken up in the form of program refinements and adjustments. These live analysis meetings are linked to an annual meeting of the two REACH Provincial Steering Committees (see the governance section – 3.3.1) which provides a feedback loop to PHOs and other provincial level partners. Findings will also be shared with health donor implementing agencies through the proposed Coordination Groups (see section 3.3.2). At the national level, operations research and analysis from the monitoring system will be shared with the AIPH Partnership Coordination Committee which is co-chaired by the National AIDS Commission, (see section 3.3.1). This will ensure that findings are shared beyond just the health sector. For example, given the Global Fund’s push for national investment approaches, information on the costs and effectiveness of decentralised treatment delivery and other key areas of operational research will be of interest to a broad range of partners. Where operations research and monitoring data has policy implications, this will be fed back to the Policy Secretariat in the MoH’s AIDS Sub Directorate.

How will we know?

On a national level, we would expect to see research results incorporated into the work undertaken under component 3 (Policy). Locally, we can track whether results of operational research have been presented and discussed at the regular program analysis meetings described in section 3.4.

Since every analysis meeting will start with a report on actions taken (or not taken) as a consequence of discussions at the previous meeting, the minutes of these meetings should provide a built-in measure of the success of operational research in generating knowledge and changing practice.

Will the benefits last?

Newly generated knowledge is valuable until it is superseded by newer, more valuable knowledge. The extent to which the ability to continue to generate new knowledge outlasts the REACH program will depend in large part on the success of larger transition plans discussed under programmatic needs post 2016 (above), as well as on the career choices made after the end of the program by program analysis staff.