Sample characteristics
At baseline, 98.0% (
n = 148 of 150) of the sample reported moderate to severe depressive symptoms (Table
1). Severe depression scores were reported by 9.3% of the sample (
n = 29). The sample was entirely Hispanic and their ccountry of origin unknown, though 88% of Hispanics in Texas are of Mexican descent [
26]. The vast majority were women (
n = 133, 88.7%), Spanish speaking (
n = 136, 90.7%), and currently married (
n = 106, 72.1%). A majority of the sample reported some high school or less (
n = 79, 52.7%). The mean age was 39.36 ± 9.08 years at time of baseline measurement.
Table 1
Sample Characteristics and Randomization Check
Age, M ± SD | 39.36 ± 9.08 | 40.52 ± 8.46 | 38.20 ± 9.57 | t = 1.57 | d = 0.26 |
Sex, female, n (%) | 133 (88.7%) | 63 (84.0%) | 70 (93.3%) | 𝜒2 = 3.25 | V = 0.15 |
Spanish Speaking, yes, n (%) | 136 (90.7%) | 69 (92.0%) | 67 (89.3%) | 𝜒2 = 0.32 | V = 0.05 |
Marital Status, n (%) | | | | 𝜒2 = 0.46 | V = 0.06 |
Married/cohabitating | 106 (72.1%) | 54 (74.0%) | 52 (70.3%) | | |
Never married | 12 (8.2%) | 6 (8.2%) | 6 (8.2%) | | |
Widowed | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | | |
Divorced | 21 (14.3%) | 9 (12.3%) | 12 (16.2%) | | |
Other | 8 (5.3%) | 4 (5.5%) | 4 (5.5%) | | |
Education Level, n (%) | | | | 𝜒2 = 11.97+ | V = 0.29 |
5th grade or less | 21 (14.3%) | 13 (17.6%) | 8 (11.0%) | | |
6th to 8th grade | 35 (23.8%) | 20 (27.0%) | 15 (20.5%) | | |
Some high school | 23 (15.6%) | 13 (17.6%) | 10 (10.7%) | | |
High school or GED | 52 (35.4%) | 25 (33.8%) | 27 (37.0%) | | |
Vocational or trade school | 1 (0.7%) | 1 (1.4%) | 0 (0.0%) | | |
Some college | 10 (6.8%) | 2 (2.7%) | 8 (11.0%) | | |
College degree | 5 (3.4%) | 0 (0.0%) | 5 (6.8%) | | |
Attrition, yes, n (%) | 25 (16.7%) | 11 (14.7%) | 14 (18.7%) | 𝜒2 = 0.43 | V = 0.05 |
Sessions attended, M ± SD | 11.90 ± 6.97 | 11.20 ± 7.28 | 12.60 ± 6.23 | t = 1.23 | d = 0.21 |
PHQ-9, baseline, M ± SD | 15.32 ± 4.15 | 15.36 ± 4.09 | 15.28 ± 4.24 | | |
PHQ-9 severity category, n (%) | | | | 𝜒2 = 3.79 | V = 0.16 |
Mild depression, n (%) | 3 (2.0%) | 2 (2.7%) | 1 (1.3%) | | |
Moderate depression, n (%) | 70 (46.7%) | 32 (42.7%) | 38 (50.7%) | | |
Moderately severe depression, n (%) | 48 (32.0%) | 29 (38.7%) | 19 (25.3%) | | |
Severe depression, n (%) | 29 (19.3%) | 12 (16.0%) | 17 (22.7%) | | |
GAD7 scores, M ± SD | 12.52 ± 4.56 | 12.17 ± 4.57 | 12.87 ± 4.56 | t = 0.93 | d = 0.15 |
DKM scores, M ± SD | 11.73 ± 1.96 | 11.33 ± 2.12 | 12.12 ± 1.71 | t = 2.50* | d = 0.41 |
SCHMC scores, M ± SD | 0.43 ± 0.83 | 0.38 ± 0.81 | 0.48 ± 0.86 | t = .70 | d = 0.06 |
LSAS scores, M ± SD | 6.07 ± 3.18 | 5.57 ± 3.06 | 6.56 ± 3.24 | t = 1.90 | d = 0.31 |
SDS scores, M ± SD | 13.02 ± 3.51 | 12.95 ± 3.60 | 13.09 ± 3.44 | t = 0.26 | d = 0.04 |
Randomization and attrition
Patients were evenly randomized to each of the SE group (
n = 75) and the DEF group (
n = 75). Eight participants (5.3% of 150) dropped from the study prior to receipt of either SE or the DEF intervention, four from each group. Tests between the two groups (
n = 150) on patient indicated randomization produced largely comparable groups on relevant patient characteristics and study measures (Table
1). A significant difference between the two intervention groups was found on the DKM measure (t = 2.50, df = 148,
p = .013, Cohen’s d = 0.41) where the DEF group (12.12 ± 1.71) demonstrated significantly more depression knowledge than the SE group at baseline (11.33 ± 2.12). The DEF group (6.56 ± 3.24) demonstrated more bias towards psychiatric medications compared to SE group at baseline (5.57 ± 3.06), but this difference was not significant (t = 1.90, df = 146,
p = .060) despite the moderate effect size (Cohen’s d = 0.42).
Analyses were completed to identify differences between those completing the intervention and 12-month follow-up (n = 123) and those who were lost to follow-up (n = 27). Those who did not complete the study significantly differed by proportion of sex in each group (𝜒2 = 4.21, df = 1, p = .040, V = .17). All of those attritioning from the study were women. Moderate effect sizes were present for the difference between those completing the protocol and dropping out on GAD-7 scores with attritioners reporting greater levels of anxiety (14.33 ± 4.53) compared to those completing the study (12.12 ± 4.49). There were no significant differences between completers and attritioners in any other baseline characteristics or by intervention group (p > .05).
Depression knowledge and stigma outcomes
Depression knowledge scores (Table
2) reported on the DKM assessment significantly differed over time (F [3.52, 416.054] = 10.66,
p < .001, partial η
2 = .080) and by intervention group (F [1, 147] = 13.09,
p < .001, partial η
2 = .082). Across all timepoints, including baseline, the group receiving the DEF was found to have significantly higher depression knowledge (Cohen’s
d = 0.375 to 0.618) with the greatest difference observed directly after the educational visit. Despite the differences, no time by group interaction was found (F [3.52, 517.99] = 0.66,
p = .598, partial η
2 = .004). Due to baseline imbalance of DKM scores between groups, a sensitivity analysis was conducted controlling for baseline DKM scores across subsequent time points. The results indicated no significant group differences at 1-month (F [1147] = 1.05,
p = .306), 6-month (F[1147] = 1.16,
p = .206), or 12-month (F[1147] = 1.60,
p = .207) follow-up scores.
Table 2
Outcomes over time using all carry forward intention-to-treat analysis
PHQ9 | 15.36 ± 4.09 | – | 8.33 ± 5.37 | 4.84 ± 5.21 | 4.79 ± 5.46 | 15.28 ± 4.24 | – | 8.64 ± 5.99 | 5.68 ± 5.40 | 5.85 ± 6.15 |
Model | Within-subjects F = 197.69***; Between-groups F = 0.703; η2 = 0.005 |
GAD7 | 12.17 ± 4.57 | – | 7.13 ± 5.25 | 4.49 ± 4.85 | 4.08 ± 4.61 | 12.87 ± 4.56 | – | 7.79 ± 5.60 | 5.73 ± 5.26 | 5.71 ± 5.50 |
Model | Within-subjects F = 131.66***; Between-groups F = 2.833; η2 = 0.019 |
DKM | 11.42 ± 2.00 | 12.08 ± 1.86 | 11.88 ± 1.91 | 11.97 ± 2.21 | 12.16 ± 2.15 | 12.12 ± 1.71 | 13.23 ± 1.94 | 12.95 ± 1.95 | 12.87 ± 1.94 | 13.03 ± 1.97 |
Model | Within-subjects F = 10.66***; Between-groups F = 12.09***; η2 = 0.082 |
SCMHC | 0.38 ± 0.81 | 0.44 ± 0.88 | 0.37 ± 0.86 | 0.26 ± 0.67 | 0.16 ± 0.55 | 0.48 ± 0.86 | 0.16 ± 0.47 | 0.15 ± 0.56 | 0.13 ± 0.55 | 0.11 ± 0.48 |
Model | Within-subjects F = 6.26***; Between-groups F = 2.210; η2 = 0.015 |
LSAS | 5.58 ± 3.06 | 5.71 ± 2.73 | 5.77 ± 2.74 | 5.88 ± 2.90 | 5.44 ± 2.70 | 6.56 ± 3.24 | 6.49 ± 3.16 | 6.57 ± 3.06 | 6.27 ± 3.06 | 6.13 ± 3.13 |
Model | Within-subjects F = 0.91; Between-groups F = 3.48+; η2 = 0.023 |
SD | 13.21 ± 3.88 | 10.61 ± 1.92 | 10.76 ± 1.91 | 10.69 ± 2.21 | 10.86 ± 1.88 | 13.00 ± 3.60 | 10.66 ± 2.37 | 10.73 ± 2.09 | 10.72 ± 2.26 | 10.96 ± 2.09 |
Model | Within-subjects F = 35.11***; Between-groups F = 0.05; η2 = 0.001 |
Differences by intervention group were not significant for the SCMHC (F [1, 146] = 2.21, p = .140, partial η2 = .015) or the SDS measures (F [1, 146] = 0.053, p = .819, partial η2 = .001). However, scores for the SDS (F [4, 584] = 35.11, p < .001, partial η2 = .194) and SCMHC (F [4, 584] = 6.26, p < .001, partial η2 = .041) significantly decreased for both groups over time. While decreases in SCMHC score represent a decrease in participant stigma towards mental health care, decreases in SDS scores indicate greater stigma towards others with depression or receiving depression treatment.
In contrast, attitudes towards psychiatric medications did not significantly change over time (F [3.48, 507.31] = 0.987, p = .449, partial η2 = .007), but the DEF group reported nearly significantly greater stigma towards medication than the SE group (F [1, 146] = 3.48, p = .064, partial η2 = .023). The effect size of the difference was small, however.
Treatment engagement
One hundred and thirty-three participants (93.7%) engaged in treatment: 62 participants (43.7%) received counseling only without antidepressant medication and 71 participants (50.0%) received counseling and antidepressant medications. The type of educational intervention received by patients was not significantly associated with treatment engagement (𝜒2 = 0.21, df = 1, p = .65, V = .004).
Depression and anxiety outcomes
Results indicated that while PHQ-9 scores (Table
2) significantly decreased over time for participants (F [2.811, 416.054] = 197.69,
p < .001, partial η
2 = .572), no differences between the SE and DEF groups were found (F [1, 148] = 0.70,
p = .403, partial η
2 = .005). For all participants at the 12-month follow-up (
n = 123), 99 patients (80.5%) reported a 50% or greater reduction in depression scores from baseline, yet this reduction was not significantly associated with the educational intervention (𝜒2 = 2.52, df = 1,
p = .112,
V = .143) nor was it associated with type of treatment engagement (𝜒2 = 3.26, df = 1,
p = .071,
V = .157). Similarly, reported anxiety scores decreased over time (F [2.847, 421.286] = 131.66, p < .001, partial η
2 = .471) and no differences by educational intervention group (F [1, 148] = 2.83,
p = .094, partial η
2 = .019).