Can a community health worker and a trained traditional birth attendant work as a team to deliver child health interventions in rural Zambia?

The study was conducted in Lufwanyama District in the Copperbelt Province of Zambia. Lufwanyama is a large, rural, undeveloped district with an estimated 2011 population of 87,592 [22], with the majority belonging to the Lamba ethnic group. Despite its location in the comparatively urban, industrialized Copperbelt, the district lacks physical infrastructure, and many roads are impassible during the rainy season. It has 12 health centers, five health posts and a newly opened district hospital. The DHMT operated for many years outside the district, but is currently housed at the new district hospital. Many basic health services, including treatment of minor illnesses, health education, antenatal care, family planning services, follow-up of patients with chronic illnesses and referrals, are provided through several categories of minimally trained community workers ‘trained TBAs, CHWs, male motivators, safe motherhood agents, family planning agents, disease surveillance agents, malaria agents, tuberculosis agents, HIV/AIDS agents, and untrained TBAs. The Lufwanyama DHMT, with support from non-governmental partners operating in the district, have trained and deployed CHWs and trained TBAs for over 30 years. CHWs and trained TBAs spend some days in a week working with health center staff at facilities. The health centers provide them with drugs and supplies and health workers supervise their work.

This prospective study assessed the level of teamwork and taskwork among community-based CHW-TBA teams supported by NHC members. We used an assessment tool developed through formative research with community leaders, health workers, CHWs and trained TBAs [23]. We carried out the assessment every six months from June 2011 to March 2013.

A CHW and trained TBA working in the same community formed the CHW-TBA team. We did not create teams for communities which had only a CHW or trained TBA. The CHW-TBA team plus two NHC support members were trained in teaming concepts prior to deployment. The training addressed both specific tasks (Table 1) that the teams would undertake as well as the skills and competencies to maintain a functioning team. The teamwork skills and competencies included i) good communication; ii) respectful dialogue and action; iii) each helping the other, mutual support, and working hand in hand; iv) assess, make decisions and manage conflicts; v) trust and confidentiality of care-seekers/community members; vi) together monitor team task and team maintenance abilities; vii) evaluate successes and failures; viii) asking for feedback; and ix) motivate and encourage each other. The specific tasks and skills required for successful community teams were identified during earlier formative research [23].The training emphasized the importance of performing the joint tasks and the need to document tasks performed. They practiced and demonstrated how to perform these tasks. The training utilized several methods including exercises, practice, demonstrations, role play, experience sharing brainstorming and real-life scenarios for the teams to acquire the necessary knowledge and skills of teamwork competences for maintaining functioning teams. The training also clarified roles of the NHCs as identified by the MOH guidelines [3].

Prior to training, we collected baseline information from team members, including age, gender, education, ethnic group, marital status, religion, membership of a social group (e.g. faith-based fellowships, parent-teacher associations, corporative societies, etc.), length of service, other occupation, and walking time from each other.

An independent, non-LINCHPIN data collector visited the core team members (CHW and trained TBA) and administered a three-part team measurement tool. The first assessment started two months after the teaming training. Part A was administered to both members together and assessed taskwork, i.e., whether the team had jointly performed any of seven agreed specific tasks in the previous three months: 1) meeting with NHCs to discuss work and performance, 2) conducting behavior change communications sessions targeting women on newborn and child care, 3) problem-solving for newborn or child care, 4) participating in outreach services, 5) supporting referral of a pregnant woman or sick child, 6) conducting intra-team referral, and 7) conducting postnatal care visits to a mother with a newborn aged 6-8 weeks. The team scored ‘0’ if a function was not performed, ‘1’ if performed but without documented evidence, and ‘2’ if there was documented evidence of performance. Part B was administered separately to the CHW and trained TBA. It assessed 27 characteristics from eight teamwork processes identified during the formative research [23]: 1) mutual performance monitoring, 2) mutual trust, 3) decision making/planning, 4) team cohesion, 5) team motivation, 6) goals and objectives, 7) communication, and 8) conflict resolution/management. Data were collected from each member about whether, in his/her opinion, the characteristic was present in the team over the previous six months. They scored ‘1’ if a member reported that the characteristic was not or hardly present in the team; ‘2’ if it was present sometimes; and ‘3’ if present all the time. The score for the team was the average score of the two members. Part C - also administered separately to each individual team member - collected information on perceived factors that may influence teamwork such as supervision, refresher training, availability of supplies, incentives, and ownership of bicycle or cell phone.

The score for the taskwork of each team at each assessment was the sum of the scores of the seven functions. The overall taskwork score for the teams was the mean score of the four assessments. For teamwork, the score for the team at each assessment was the average score of the two members from the twenty seven indicators (expressed as a percent). The overall teamwork score was also the mean score of the four assessments.

A team was categorized ‘inactive’ if unavailable for an assessment and the local NHC confirmed its inactivity and break-up. We categorized the remaining teams ‘high’ if the mean score on the taskwork scale was ≥7 of a possible 14, and the mean score on the teamwork scale was <90%; and ‘low’ if the taskwork score was <7 or teamwork score was <90%. We decided on the cut-offs prior to data collection, but we modified the categorization based on the distribution of the teamwork scores. In order to evaluate factors that may influence the level of teaming, frequency and proportions were compared with chi-square test; odds ratios (OR) and 95% confidence intervals (CI) were calculated for each characteristic. All data analysis was conducted in EpiInfo software package [24].

Ethical approval was obtained from the Boston University Institutional Review Board (BU-IRB) and a local Zambian ethical review committee (ERES CONVERGE). Informed consent was obtained from all study participants with a consent form developed in accordance with guidelines of the BU-IRB and the local ethical review committee and translated into Bemba, the language of common communication in the district. During consenting, study personnel explained the purpose and rationale of the study, informed the participants that they were not obliged to participant in the research, and assured them of the confidentiality of the information collected from them.

The project created and trained 47. There were 74 CHWs operating in the district but some CHWs did not have a trained TBA operating in their communities. The CHWs were predominantly male (80.9%), and the trained TBAs were all female (Table 2). CHWs were younger than the trained TBAs (average age of 44 years vs 53 years). Most CHWs had more schooling than the trained TBAs. Half the trained TBAs were of the local Lamba ethnic group while only a third of the CHWs were Lamba. CHWs were more likely to be currently married than the trained TBAs. Only about a fifth of the CHWs and the trained TBAs reported that being a CHW or TBA was their main occupation.

We categorized 21 (44.7%) teams as high, 12 (25.5%) as low, and 14 (29.8%) as inactive. Three teams became inactive after the first assessment, four after the second, and the remaining seven after the third. CHW departure, usually to find a new job, was responsible for most of the inactive teams (71.4%) (Table 3). Two CHWs were employed as casual laborers to work at rural health centers, two CHWs stopped because they became frustrated with the work and one trained TBA was forced to stop because some members of the community believed she was not representing community values.

All team members reported the presence of mutual trust within their teams during all four assessments (Table 4). Many team members reported comprehension of team goals and objectives and team cohesion as present most of the time. On the other hand, decision making/planning and mutual performance monitoring were reported lacking in most cases. The teams reported only six conflicts in the four assessments, all of which were satisfactorily resolved or managed. Team motivation and communication were reported to have improved over time while mutual performance monitoring and decision making/planning after initial improvement, declined during the last assessment (Figure 1).

Table 5 shows reported and documented joint activities. The most common documented joint activity was making a home visit to a mother with a young infant aged about 6-8 weeks where the trained TBA ‘handed over’ the child to the CHW (55.3%), followed by meeting with NHCs to discuss work and performance (36.5%). Less commonly reported joint activities were intra-team referral (e.g., the CHW referring a pregnant woman to the trained TBA or the trained TBA referring a mother with a sick child to the CHW) and joint problem solving (15.6 and 21.6%, respectively). The most common joint activity (documented plus undocumented) was participation in outreach services, including immunization conducted by the supervising rural health center staff, and BCC sessions targeting women to educate them about newborn and child care (Figure 2). The least common activities by these criteria were intra-team referral and supporting referral to health facilities.

Teams with members residing within one hour’s walking distance were more likely to score high (OR = 5.80; 95% CI: 1.52-22.1; p = 0.007). Teams whose members were jointly supervised were also more likely to score high (OR = 3.2; 95% CI: 0.83-12.74; p = 0.05), barely achieving statistical significance (Table 6). Teams whose members were of the same sex and with at least one member receiving some form of incentives (e.g. payment in-kind or cash from the community for services rendered) were likely to score high, but these differences were not significant.

The study has limitations. The assessment consisted mainly of participants’ subjective reports of satisfaction, attitudes, and opinions, and they may have over-rated themselves. The small sample size may have precluded identifying other factors influencing teaming. Another limitation was that the assessment tool was not validated.