Remote monitoring of CMAM programmes coverage: SQUEAC lessons in Mali and Mauritania

By Jose Luis Alvarez Moran, Brian Mac Domhnaill and Saul Guerrero

Jose Luis Alvarez Moran is a Medical Doctor with a PhD in International and Public Health. He works as an assistant in Rey Juan Carlos University and is currently conducting nutrition surveys for Action Against Hunger.

Brian MacDomhnaill is an independent expert in monitoring and evaluation of health programmes. He has worked in Ghana, Brazil, Angola, Mauritania and Djibouti.

Saul Guerrero is the Evaluations, Learning and Accountability (ELA) Advisor at Action Against Hunger (ACF-UK). Prior to joining ACF, he worked for Valid International Ltd. in the research, development and roll-out of CTC/CMAM. He has worked in Afghanistan, Algeria, Chad, DRC, Ethiopia, Indonesia, Kenya, Liberia, Malawi, Mali, Mozambique, Nepal, Niger, Nigeria, Sierra Leone, North & South Sudan and Zambia.

The authors would like to thank Chantal Autotte Bouchard, David Kerespars, Dr. Theophane Traore, INSTAT, and the ACF teams in Mali, Mauritania and Spain (Elisa Dominguez in particular) for their support. To Ernest Guevara (Valid International) and Mark Myatt (Brixton Health) for their valuable comments and to the European Commission Office for Humanitarian Aid & Civil Protection (ECHO) for their financial support.

Action Against Hunger (ACF) currently supports community based management of acute malnutrition programmes (CMAM) programmes in over 20 countries around the world, with a long-standing presence in the Sahel region of West Africa, including Mauritania, Niger, Mali and Chad. Most of these interventions are integrated CMAM programmes, operated by Ministries of Health and local partners with technical and logistical support from ACF teams on the ground. Monitoring the impact of these interventions, and their coverage in particular, is of paramount importance to the organisation. Increasing coverage was instrumental in the shift from inpatient care in the form of therapeutic feeding centres (TFCs) to outpatient models (CMAM) and remains one of the most widely accepted indicators of programme performance and impact. Whilst other indicators (e.g. cure rates, length of stay, average weight gain) provide an insight into the efficacy of treatment, only when combined with coverage do they provides an accurate and reliable indication of the needs met by a programme. Since December 2010, ACF has been increasingly relying on the Semi-Quantitative Evaluation of Access & Coverage (SQUEAC) to measure programme coverage and identify the factors affecting the performance of CMAM programmes¹.

Data collection team measuring MUAC in the village of Hamakouladji, Sony Aliber municipality (Gao, Mali)

According to a recent UNICEF estimation, there are 55 countries currently implementing CMAM in one form or another². The scale of CMAM programming, limited nongovernmental organisation (NGO) resources, and deteriorating security conditions in many regions (including in the north-west and Horn of Africa) is increasingly forcing support organisations such as ACF to operate remotely with limited access to programme areas. The extent of the constraints varies, from limited access to areas within a district (e.g. ACF supported programme in Guidimaka, Mauritania), to limited access to parts of a country (e.g. ACF supported programme in Gao, Mali) to limited access to an entire country (e.g. ACF supported programmes in Somalia). All of these environments present challenges, in particular for the implementation of monitoring and evaluation activities with a strong field component such as SQUEAC.

Monitoring coverage remotely

Experiences in using SQUEAC remotely have been limited, with the most notable experience provided by Valid International and Oxfam-Novib in Somalia³ (see Box 1). Recently, ACF carried out SQUEAC investigations in Mauritania (February 2011) and Mali (July-August 2011). In both cases, lack of security prevented the SQUEAC lead investigators from travelling to the programme areas. In the case of Gao (Mali), the lead investigator was unable to visit the district in which the programme operated but was able to visit a neighbouring district. In the case of Guidimaka (Mauritania), the investigator was able to visit the district but could not travel to most areas outside of the district capital.

The analysis presented here will draw largely from these two experiences. A brief synopsis of the SQUEAC methodology and its key features in more conventional settings is included in Box 2. The article focuses its attention on two general stages of using SQUEAC to monitor programme coverage remotely: planning and implementation. It will conclude with some lessons learned and provide practical suggestions for other practitioners wishing to undertake similar exercises in the future.

ACF’s remote experiences in Mali and Mauritania

CMAM programmes supported by ACF in Gao (Mali) and Guidimaka (Mauritania) are largely inaccessible to expatriate staff due to security threats posed by AQMI (Al-Qaida au Maghreb Islamique) in the region. Security threats do not prevent local teams from implementing programme activities, but monitoring supervision is more difficult since the local teams often need to travel to more accessible areas to meet with technical support and management staff. The decision to evaluate the coverage of both these programmes forced the organisation to explore different means of employing SQUEAC.

Both investigations faced similar accessibility problems and relied on the work of external SQUEAC lead investigators brought into the programme especially to carry out the investigations. The lead investigators had constant remote support from ACF’s Evaluations, Learning & Accountability (ELA) Advisor based in London. In both countries, two teams were formed: a coordination team (including the lead investigator, the ACF Medico-Nutritional Coordinator, and the logistics department at capital level) and a data collection team (composed of the investigator’s assistant and local enumerators recruited for the purposes of SQUEAC).

The type of training received by the lead investigators prior to their respective SQUEACs was different. The lead investigator for Mauritania received a three-day SQUEAC introductory training prior to departure, and remote technical support throughout the investigation period. The lead investigator for Mali received a 5-day, on-the-job training in-country, which included joint analysis of existing programme data and the development of preliminary hypotheses. The availability of previous SQUEAC experience was helpful in planning SQUEAC remotely, particularly for developing a hypothesis about coverage with limited access to the programme.

Key lessons learned

Lead investigator's assistant collecting qualitative data with mothers of the village Monzonga, Ansongo (Mali)

ACF’s experiences in implementing SQUEAC remotely in Mali and Mauritania provided five key lessons:

Advanced planning

When undertaking SQUEAC remotely, forward planning is essential. This is partly due to time constraints. When working remotely, activities take longer, but since the exercise must be completed in a similar timeframe (to remain practical and cost-effective), time must be managed more strategically than in ‘conventional’ environments. ACF’s experience showed that both the coordination and data collection teams must be well coordinated to ensure an optimal use of each team’s time. For example, with advanced planning, the coordination team is able to carry out some parts of the analysis whilst the field team simultaneously collects field data. In that respect, the SQUEAC methodology is appropriate for such environments, as it is not always a linear process (between inputs and outputs) and is flexible enough to allow for multiple activities to be implemented, sometimes in parallel. Advanced planning is also essential to ensure an adequate recruitment process for reliable enumerators that can take significantly longer when undertaken remotely.

Data collation

The first stage of a SQUEAC investigation involves collating/collecting programme data to build a picture of what programme coverage is and where the areas of high and low coverage are likely to be. This process of data collation normally takes place during the SQUEAC investigation period, partly on the assumption that these data are readily available from programme reporting, databases and other information management systems. Collating all this information in remote programmes can be a long process, especially for integrated (MoHled) programmes where information is often held at the Service Delivery Units (e.g. health centres). The experience from Mali shows that collating such information prior to the start of SQUEAC can ensure that Stage One focuses mostly on analysing the data (and requesting additional data) rather than on collating it. In this respect, having a multi-layered team (with coordination and a data collection team in the field) enables some elements of the analysis/collection of (last minute) data to be undertaken in tandem.

The Mali and Mauritania experiences show that some data can and should be collected in advance (see Box 3). Some of this information is consistently collected through routine monitoring data (including admissions, defaulters and deaths) but other atypical and non-routine data require specific mechanisms to collect them. Integrating these last ones in the basic programme monitoring data would facilitate the implementation (and mainstreaming) of coverage investigations.

Multi-layered team

SQUEAC investigators will still need to determine what and how additional qualitative and quantitative data are to be collected, as well as means of analysis. In conventional SQUEAC investigations, these processes generally occur in the same place and are carried out by the same teams (enabling a more real-time, active/reactive process of data collection). In remote investigations, a separation of the two processes may be necessary, employing a multilayered team approach. The model used in Mali replicated the two-tier CMAM implementation approach used by ACF to support the CMAM programme. In other words, most technical/ strategic/analytical processes were carried out remotely by one team (the coordination team) while a second team (the data collection team) had access to the programme area and was in charge of carrying out qualitative and quantitative data collection processes (see Figure 1). For this two-tier arrangement to succeed, regular communication (prior to and during) the investigation was crucial.

Regular communication

Even when a programme area is not equally accessible to all, it is important to bring all the teams working in a SQUEAC to an accessible location for discussion about the activities and processes involved (e.g. calculating weight for height, measuring mid upper arm circumference (MUAC), presence of oedema, etc.). Face-to face communication should occur at least once with the lead investigator. During these meetings it is important to involve everyone in the development of a map of the programme area. The development of a map jointly with the team not only ensures that the spatial dimension of the exercise is understood, but it is a critical step in ensuring that the lead investigator gets an opportunity to discuss and explore questions about the programme area. Working with a map will help in the implementation of SQUEAC and also assist the supervision of the teams.

Once SQUEAC begins, regular communication becomes essential. New technologies, such as internet, emails and mobile telephones, are able to provide a real-time link between those with direct access to the field and the coordination team working remotely. In Mali, other platforms such as radio proved helpful in enabling field teams to notify remote communities of their planned field visits. New technologies allowed for a timely transfer of information between field teams and coordination teams. More importantly perhaps, new technologies enabled both teams to remain in touch and in the process steer the process of data collection and data analysis.

Linking data analysis and data collection and steering the process of data collection is particularly important when it comes to collecting qualitative data. Qualitative data collection in SQUEAC can set out to assess factors that are known to influence coverage (see Box 4) but it must ultimately be an iterative process, adapted to newly emerging information and trends. Communication between those collecting qualitative data, and those responsible for analysing it and identifying new lines of enquiry, is therefore essential, as is the triangulation of qualitative data.

Supervision & motivation

The process of qualitative and quantitative data collection in SQUEAC often merits close supervision to ensure that data are adequately triangulated (by source and method) and to ensure that sampling is comprehensive and exhaustive. In Mali and Mauritania, supervision could not be undertaken directly by the coordination team or the lead investigator due to a lack of access to (most) programme areas. Some of the issues already discussed (e.g. regular communication, advanced planning and recruitment of adequate field teams, etc.) combined with well managed workloads and clear roles and responsibilities can help minimise risks of remote coverage The selection of a strong and reliable assistant(s) is an essential part of remote SQUEAC implementation. Spending sufficient time to transmit the methodology and the processes involved can ensure that the assistant(s) will be able to steer the teams in the right direction. Constant communication is also essential.

Surveyor filling up the questionnaire in the village Zinda, Gabero municipality (Mali)

The experiences from Mali and Mauritania provide some examples of how how proactively to strengthen supervision and motivation. In Mali, teams carried out daily phone conversations at the start of the day to discuss the daily plan of action and at the end of every day to follow up, strengthen the team motivation and address everyday field problems. In Mauritania, data collection teams returned to base whenever possible to debrief, relay data, and discuss challenges.

Proactively investing in recruitment and training cannot always ensure a successful outcome. In Mauritania, the data collected as part of the wide-area survey (Stage 3) was found, upon checking, to be unreliable. A decision was made to send a second team of enumerators to re-verify the data. This was only possible because of the contingency planning developed to accommodate the remote nature of the exercise.

Conclusions

ACF’s experiences in Mali and Mauritania have shown that physical lack of access to programme areas is not an insurmountable barrier to monitoring the performance of the intervention. Implementing remote SQUEAC investigations is feasible and can provide sufficiently reliable data about programme coverage and the factors affecting it. Remote coverage investigations do not require additional time or resources if there is enough advance planning, support from the local base and a contingency plan has been provided. They do require that standard SQUEAC processes be accentuated or strengthened. These include: advanced planning, preparation of data for its analysis, separating data collection and data analysis processes, using new technologies to ensure regular communication between both sets of activities, and addressing the issue of supervision and motivation proactively and reactively as the investigation develops.

Like other aspects of remote technical support, implementing SQUEAC investigations remotely does require a greater degree of reliance on field teams. Trust is of the essence, but CMAM programmes can minimise potential risks by investing time in the selection of these teams and by allocating manageable daily workloads. Although SQUEAC was designed to be implemented by programme staff, the involvement of experienced lead investigators/technical advisors often proves valuable in the process of data analysis, by bringing a measure of objectivity to key processes (e.g. interpretation/weighing of findings when building a prior). In remote SQUEAC, the presence and input of external technical advisors can help bridge the gaps left by lack of access and limited data accessibility. As the experience from Mali showed, such input during the early part of the process (Stage One) was particularly helpful in ensuring that subsequent processes were adequately implemented. Finally robust data collection, always important for SQUEAC, is essential for remotely managed programmes. By introducing local teams to SQUEAC, it becomes easier for programmes to adopt SQUEAC-based monitoring frameworks that can facilitate future SQUEAC investigations and programme monitoring as a whole.

SQUEAC was developed as a way for nutrition programmes to monitor their own performance. For programmes operating in areas with limited access/mobility, the need for reliable self-evaluation tools is particularly pressing. Carrying out SQUEACs in such contexts is possible with only minimal changes to the methodology. The real challenge lies in creating the capacity within these programmes to collect, document, analyse and report routine data in a manner that enables them to carry out future exercises with minimum external support.

For more information, contact Saul Guerrero, email: s.guerrero@actionagainsthunger.org.uk

¹For more on the SQUEAC method and its use see Myatt, M. SQUEAC: Low resource method to evaluate access and coverage of programmes. (Field Exchange, Emergency Nutrition Network, Issue 33, June 2008, p.3) & Schofield, L. et.al. (2010) SQUEAC in routine monitoring of CMAM programme coverage in Ethiopia (Field Exchange, Issue 38, April 2010, Emergency Nutrition Network, p.35).

²UNICEF & Valid International (2011) Global Mapping Review of Community-based Management of Acute Malnutrition with a focus on Severe Acute Malnutrition (Nutrition Section, Nutrition in Emergency Unit, UNICEF HQ-NY and Valid International, March 2011)

³Valid International. Personal Communication.

⁴In Bayesian inference, the prior is a probabilistic representation of available knowledge about a quantity. In SQUEAC, the prior is a probabilistic representation of knowledge relating to program coverage. SQUEAC uses a Beta distributed prior.

⁵In Bayesian inference, the information provided by new evidence. The likelihood is use to modify the prior to arrive at the posterior. In SQUEAC, this is the information provided by a survey (the likelihood survey).

⁶In Bayesian inference, the posterior is the result of modifying prior belief using new evidence