Table 6. Cox proportional hazards regression analysis of factors associated with United States mental health trial discontinuation.
| Univariable | Multivariable | |||
|---|---|---|---|---|
| n = 5,818 | HR (95% CI) | p-value | aHR (95% CI) | p-value |
| Primary Objective | ||||
| Treatment | Reference | Reference | ||
| Basic Science | 0.74 (0.47–1.18) | 0.21 | 0.63 (0.39–1.01) | 0.058 |
| Other | 0.86 (0.65–1.14) | 0.29 | 1.02 (0.76–1.36) | 0.92 |
| Prevention | 0.66 (0.46–0.94) | 0.021 | 0.97 (0.65–1.43) | 0.87 |
| Intervention | ||||
| Behavioral (Beh) | Reference | Reference | ||
| Pharmaceutical (Pharm) | 3.47 (2.75–4.36) | <0.0001 | 2.44 (1.73–3.45) | <0.0001 |
| Other (Oth) | 1.84 (1.30–2.60) | 0.00058 | 1.42 (0.99–2.04) | 0.059 |
| Beh + Pharm | 1.42 (0.84–2.41) | 0.19 | 1.34 (0.76–2.34) | 0.31 |
| Beh + Oth | 1.30 (0.71–2.36) | 0.40 | 1.59 (0.87–2.90) | 0.13 |
| Pharm + Oth | 4.49 (3.10–6.50) | <0.0001 | 3.65 (2.37–5.60) | <0.0001 |
| Beh + Pharm + Oth | 0.00 (0.00 - >100) | 0.99 | 0.00 (0.00 - >100) | 0.99 |
| Trial Phase | ||||
| Not Applicable | Reference | Reference | ||
| Phase 1 | 1.84 (1.32–2.57) | 0.00034 | 1.31 (0.92–1.88) | 0.14 |
| Phase 1/2–2 | 1.78 (1.40–2.27) | <0.0001 | 1.27 (0.94–1.71) | 0.10 |
| Phase 2/3–3 | 1.92 (1.42–2.60) | <0.0001 | 1.23 (0.86–1.75) | 0.26 |
| Phase 4 | 2.47 (1.91–3.17) | <0.0001 | 1.18 (0.86–1.61) | 0.30 |
| Year Registered | ||||
| 2007–2012 | Reference | Reference | ||
| 2013–2018 | 1.17 (0.97–1.40) | 0.10 | 1.17 (0.97–1.42) | 0.10 |
| Funder | ||||
| US Govt | Reference | Reference | ||
| Industry | 5.38 (4.05–7.15) | <0.0001 | 2.86 (2.07–3.96) | <0.0001 |
| AMC/Hosp/Oth | 3.40 (2.67–4.34) | <0.0001 | 2.59 (2.01–3.35) | <0.0001 |
| Blinding | ||||
| None | Reference | Reference | ||
| Single | 0.43 (0.33–0.57) | <0.0001 | 0.58 (0.42–0.80) | 0.0011 |
| Double | 1.44 (1.18–1.75) | 0.00026 | 0.95 (0.69–1.29) | 0.73 |
| Randomization | ||||
| Non-Randomized | Reference | Reference | ||
| Randomized | 0.63 (0.51–0.78) | <0.0001 | 0.96 (0.71–1.31) | 0.81 |
| Reports DMC | ||||
| No | Reference | Reference | ||
| Yes | 0.58 (0.48–0.70) | <0.0001 | 0.64 (0.53–0.78) | <0.0001 |
| Studies Children | ||||
| No | Reference | Reference | ||
| Yes | 0.90 (0.71–1.16) | 0.42 | 1.18 (0.89–1.56) | 0.26 |
| Number of Sites | ||||
| One | Reference | Reference | ||
| ≥Two | 0.78 (0.62–0.99) | 0.038 | 0.74 (0.58–0.95) | 0.019 |
| Disorder Category | ||||
| Substance | 0.55 (0.44–0.69) | <0.0001 | 0.67 (0.49–0.92) | 0.013 |
| Depression | 1.23 (0.99–1.54) | 0.062 | 0.99 (0.74–1.32) | 0.96 |
| Schizophrenia | 1.51 (1.14–2.01) | 0.0042 | 1.04 (0.73–1.48) | 0.84 |
| Neurodevelopment | 0.99 (0.72–1.36) | 0.96 | 0.61 (0.41–0.92) | 0.017 |
| Anxiety | 1.48 (1.07–2.04) | 0.019 | 1.27 (0.89–1.82) | 0.19 |
| Trauma | 0.88 (0.65–1.20) | 0.43 | 1.29 (0.89–1.86) | 0.18 |
| Sleep | 1.29 (0.86–1.93) | 0.21 | 1.04 (0.67–1.62) | 0.85 |
| Bipolar | 1.39 (0.96–2.00) | 0.078 | 0.90 (0.60–1.35) | 0.62 |
| OCD | 1.07 (0.59–1.94) | 0.83 | 0.73 (0.38–1.40) | 0.34 |
| Feeding | 0.88 (0.45–1.70) | 0.70 | 0.81 (0.40–1.63) | 0.56 |
| Neurocognitive | 2.55 (1.53–4.27) | 0.00036 | 1.77 (1.00–3.15) | 0.051 |
| Non-DSM | 0.92 (0.71–1.18) | 0.50 | 0.89 (0.67–1.20) | 0.45 |
All listed variables were included in the multivariable regression. ‘HR’ denotes hazard ratio, and ‘aHR’ denotes adjusted hazard ratio. ‘95% CI’ denotes 95% confidence interval. ‘DMC’ denotes Data Monitoring Committee. ‘AMC/Hosp/Oth’ denotes Academic Medical Centers/Hospitals/Other. ‘US Govt’ denotes United States Government, which includes NIH (National Institutes of Health) and US Federal agency funders. Refer to S2 Table for a description of the disorders included under each disorder category. For intervention types, ‘Beh’ denotes Behavioral, ‘Pharm’ denotes pharmaceuticals, and ‘Oth’ denotes Other. See methods for the interventions included in each of these categories. Trials in which multiple intervention types were tested are labeled with all relevant interventions (e.g. a trial testing psychotherapy and pharmacotherapy was labeled ‘Beh + Pharm’).