Community Health Worker Performance
Conditions
Keywords
Community Health Worker, Health worker supervision, Mali, Randomized controlled trial
Brief summary
Countries across sub-Saharan Africa are scaling up Community Health Worker (CHW) programmes, yet there remains little high-quality research assessing strategies for CHW supervision and performance improvement. This randomised controlled trial aims to determine the effect of a personalised performance dashboard used as a supervision tool on the quantity, speed, and quality of CHW care. This study is a randomised controlled trial in a large health catchment area in peri-urban Mali. One hundred forty-eight CHWs conducting proactive case-finding home visits were randomly allocated to receive individual monthly supervision with or without the CHW Performance Dashboard from January to June 2016. Randomisation was stratified by CHW supervisor, level of CHW experience, and CHW baseline performance for monthly quantity of care (number of household visits). With regression analysis, we used a difference-in-difference model to estimate the effect of the intervention on monthly quantity, timeliness (percentage of children under five treated within 24 hours of symptom onset), and quality (percentage of children under five treated without protocol error) of care over a six-month post-intervention period relative to a three-month pre-intervention period.
Interventions
During the study period, all CHWs, regardless of treatment arm, performed proactive case detection, the process of conducting at least two hours per day of door-to-door home visits to proactively identify - through health history inquiry and/or disease diagnostics - patients who need care. For all patients identified, CHWs provided doorstep counselling, evaluation, diagnostics, treatment, referral to appropriate health facilities, and follow-up. CHWs provided care in the community without user fees, and were able to refer patients to the reinforced government primary health centres for care without user fees as well. CHWs were residents of the communities they served, and they were required to be available at home or by phone for consultation at any time.
CHWs in both study arms received monthly individual supervisory sessions and weekly group supervisory sessions from their dedicated CHW supervisor. An individual monthly session of 360 Supervision included: (i) solicitation of patient perspectives of CHW care; (ii) direct observation of CHW doorstep care; and (iii) a one-on-one feedback discussion with or without the CHW Performance Dashboard depending on treatment arm.
The CHW Performance Dashboard was a graphic display of a CHW's performance along three indicators defined as follows: (i) Quantity of care: the number of homes visited during the month; (ii) Timeliness of care: the percentage of sick children under five treated within 24 hours of symptom onset during the month; (iii) Quality of care: the percentage of sick children under five treated without protocol error among 23 potential errors during the month. The Dashboard displayed an individual CHW's quantity, timeliness, and quality of care indicators from the previous month, using absolute numbers, percentages, and visual graphics, alongside those of the highest performing CHW. During the individual supervisory feedback session, this personalised and relative (to the highest performer) quantitative performance feedback helped orient the discussion of strengths and weaknesses, and allowed the CHW to see quantitatively and visually how his/her performance fared the previous month.
Sponsors
Study design
Eligibility
Inclusion criteria
* To be a CHW in the study site at the time of enrolment (n=148)
Exclusion criteria
* CHW who pretested the Dashboard tool (n=2)
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Quantity of care | 9 months | The number of proactive case-finding home visits during the month |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Timeliness of care | 9 months | The percentage of sick children under five treated within 24 hours of symptom onset during the month |
| Quality of care | 9 months | The percentage of sick children under five treated without protocol error among 23 potential errors during the month |