Table 2.
Key characteristics of the community health worker models associated with notification impact
| LI (n = 15) | MI (n = 16) | HI (n = 19) | Total | p-value╪ | |
|---|---|---|---|---|---|
| Implementation activities | |||||
| Tuberculosis (TB) and other | 6 (40%) | 5 (31%) | 8 (42%) | 19 (38%) | 0.790 |
| TB only | 9 (60%) | 11 (69%) | 11 (58%) | 31 (62%) | |
| % time spent on TB activities | 77.3 | 85 | 77.9 | 80 | 0.649* |
| Community outreach1 | 13/15 (87%) | 13/15 (87%) | 16/19 (84%) | 42/49 (86%) | 1.00 |
| Verbal screening1 | 13/15 (87%) | 12/15 (80%) | 17/19 (89%) | 42/49 (86%) | 0.730 |
| HIV testing1 | 1/15 (7%) | 1/15 (7%) | 2/19 (11%) | 4/49 (8%) | 1.00 |
| Sputum collection and transportation | 12 (80%) | 11 (69%) | 13 (68%) | 36 (72%) | 0.712 |
| Linkage to treatment | 11 (73%) | 9 (56%) | 15 (79%) | 35 (70%) | 0.326 |
| Treatment counseling | 5 (33%) | 7 (44%) | 13 (68%) | 25 (50%) | 0.106 |
| Recruitment and selection | |||||
| Had prior experience2 | 11/15 (73%) | 10/14 (71%) | 13/19 (68%) | 34/48 (71%) | 0.951 |
| Years of education3 | 12 (10–14) | 12 (10–12) | 10 (9–12) | 12 (10–12) | 0.378* |
| Provided written contracts | 13 (87%) | 15 (94%) | 15 (79%) | 43 (86%) | 0.462 |
| From TB REACH | 7 (54%) | 13 (87%) | 8 (53%) | 27 (65%) | 0.095 |
| From non-governmental organization | 4 (31%) | 2 (13%) | 4 (27%) | 10 (23%) | 0.513 |
| From government | 2 (15%) | 1 (7%) | 3 (20%) | 6 (14%) | 0.655 |
| Provided differentiated contracts5 | 10/12 (83%) | 14/15 (93%) | 9/15 (60%) | 33/42 (79%) | 0.075 |
| Pre-service training¶ | |||||
| Training method | |||||
| Expert | 14 (93%) | 15 (94%) | 18 (95%) | 47 (94%) | 1.000 |
| Peer-to-peer | 10 (67%) | 8 (50%) | 9 (47%) | 27 (54%) | 0.495 |
| Hands-on | 14 (93%) | 13 (81%) | 17 (90%) | 44 (88%) | 0.652 |
| E-learning | 0 (0%) | 0 (0%) | 2 (11%) | 2 (4%) | 0.323 |
| Training setting | |||||
| Classroom-based | 14 (93%) | 15 (94%) | 19 (100%) | 48 (96%) | 0.519 |
| Community-based | 8 (53%) | 9 (56%) | 15 (79%) | 32 (64%) | 0.223 |
| Average hours of pre-service trainings4 | 12 (5–24) | 12 (8–18) | 16 (8–30) | 16 (8–24) | 0.366* |
| Refresher training¶ | |||||
| Formal refresher trainings2 | 9/13 (69%) | 11/16 (73%) | 11/19 (58%) | 31/48 (65%) | 0.646 |
| Formal training method (n = 31) | |||||
| Expert | 8 (89%) | 11 (100%) | 10 (91%) | 29 (94%) | 0.740 |
| Peer-to-peer | 7 (78%) | 10 (91%) | 4 (36%) | 21 (36%) | 0.019 |
| Hands-on | 6 (67%) | 10 (91%) | 8 (73%) | 24 (77%) | 0.437 |
| E-learning | 0 (0%) | 0 (0%) | 2 (18%) | 2 (7%) | 0.314 |
| Formal training setting | |||||
| Classroom-based | 8 (89%) | 10 (91%) | 7 (64%) | 25 (81%) | 0.205 |
| Community-based | 4 (44%) | 6 (55%) | 7 (64%) | 17 (55%) | 0.692 |
| Frequency of refresher trainings4 | 3 (1–4) | 2 (2–4) | 2 (2–4) | 2 (2–4) | 0.978* |
| Average hours of refresher trainings3 | 4 (3–8) | 5 (3–8) | 7 (3–8) | 5.5 (3–8) | 0.725* |
| Supervision | |||||
| Issues addressed by direct supervisor4 | 13/15 (87%) | 13/13 (100%) | 16/19 (84%) | 42/47 (89%) | 0.420 |
| Issues addressed by upper management4 | 3/15 (20%) | 0/13 (0%) | 4/19 (22%) | 7/47 (15%) | 0.197 |
| Female supervisor (%)2 | 58.8 | 53.5 | 29.5 | 45.6 | 0.007 |
| Average # community health workers (CHWs) per supervisor3 | 13 (7–26) | 7 (6–10) | 15 (5–60) | 9 (6–25) | 0.247* |
| Average # of supervisor reviews per quarter2 | 9 (3–12) | 6 (3–12) | 6 (3–12) | 6 (3–12) | 0.775* |
| Average # of supervisor direct feedback per quarter2 | 9 (3–12) | 3 (1–12) | 6 (4–12) | 6 (3–12) | 0.293* |
| Sustainability and integration | |||||
| Promoted to a higher role | 5/14 (36%) | 8/13 (62%) | 8/19 (42%) | 21/46 (46%) | 0.426 |
| CHWs working on the project keep their jobs at the close of the project2 | |||||
| All kept their jobs after project | 7/14 (50%) | 1/15 (7%) | 5/19 (26%) | 13/48 (27%) | 0.064 |
| A subset kept their jobs after project | 4/14 (29%) | 11/15 (73%) | 11/19 (58%) | 26/48 (54%) | |
| None kept their jobs after project | 4/14 (29%) | 2/15 (13%) | 3/19 (16%) | 9/48 (19%) | |
| Continued with the same responsibilities (N = 39) | 8/11 (73%) | 3/12 (25%) | 13/16 (81%) | 24/39 (62%) | 0.007 |
Data are %, mean or median. % are calculated based on the total number of projects with available data. Percentages within each category are based on the total projects within each category. N sizes are listed for variables with missing values
1: 1 (2%) respondent missing information on questions asked
2: 2 (4%) respondents missing information on questions asked
3: 6 (12%) respondents missing information on questions asked
4: 3 (6%) respondents missing information on questions asked
5: 8 (16%) respondents missing information on questions asked
╪: Fisher’s exact test and Chi-square tests: comparing proportions that is conditional on frequencies; ANOVA test: comparing means
*: Median (IQR) and Kruskal–Wallis test
¶: As indicated by WHO CHW training guidelines, expert, peer-to-peer, and hands-on training indicates face-to-face interaction as opposed to distance learning (e-learning). Classroom-based training emphasizes theoretical knowledge; community-based training emphasizes practical application