Sample Size Proejctions

All projections are based on 24 months of active data collection, 90% of admissions being eligible, and 50% of eligible people agreeing to participate (this is very conservative – we actually get ~80% to agree to participate).

Admissions per year are admissions per unit per year (e.g., 2 units, 300 admissions/year = 600 total admissions).

Some lines don’t extend all the way across the x-axis once they reach power of .99 (in those cases, we’d have sufficent power with <50% compliance rate).

2 units, 300 admissions/year

(least conservative, but still very realistic)

Power Curve

Expected Counts

## [1] "Total suicide attempts expected: 24.75"

## [1] "Total suicide plans expected: 99"

Phenotype	Expected count
Low mean, low variability	62
Mod. mean, low variability	139
Mod mean, mod. variability	93
High mean, mod. variability	139
High mean, high variability	62

2 units, 200 admissions/year

Power Curve

Expected Counts

## [1] "Total suicide attempts expected: 18"

## [1] "Total suicide plans expected: 72"

Phenotype	Expected count
Low mean, low variability	45
Mod. mean, low variability	101
Mod mean, mod. variability	68
High mean, mod. variability	101
High mean, high variability	45

1 unit, 300 admissions/year

Power Curve

Expected Counts

## [1] "Total suicide attempts expected: 13.5"

## [1] "Total suicide plans expected: 54"

Phenotype	Expected count
Low mean, low variability	34
Mod. mean, low variability	76
Mod mean, mod. variability	51
High mean, mod. variability	76
High mean, high variability	34

1 unit, 200 admissions/year

Power Curve

Expected Counts

## [1] "Total suicide attempts expected: 9"

## [1] "Total suicide plans expected: 36"

Phenotype	Expected count
Low mean, low variability	22
Mod. mean, low variability	51
Mod mean, mod. variability	34
High mean, mod. variability	51
High mean, high variability	22