This paper presents empirical findings from a multi-year applied naturalistic study that focused on changes in clinical and work outcomes after using an employee assistance program in the United States. Self-report outcomes assessed with standardized measures were collected at the start of counseling from 33,683 employees during the normal course of business at CuraLinc Healthcare during 2017 to June 2022. A total of 5,159 employees had valid Pre and 30-day Post use data on one or more of the outcomes. Tests within each outcome sample found significant improvement after counseling (most had a 5 or 6 session model). Among the 487 cases using the EAP for a depression issue, the average severity level of depression symptoms (PHQ-9) was reduced by 59%. Among the subset of these cases at-risk for clinical depression disorder (n = 292), 89% had a reliable clinical improvement in severity. Among the 440 cases using the EAP for an alcohol issue, the average severity level of alcohol misuse (AUDIT-10) was reduced by 67%. Among subset of these cases at-risk for alcohol misuse disorder (n = 288), 73% had a reliable clinical improvement in severity. Across all cases with absence data (n = 3,732), the average hours of work absence in the past month per employee (Workplace Outcome Suite) was reduced by 80%, changing from 8.9 hours at Pre to 1.8 hours at Post. At the start of counseling, 35% of these cases were classified as having an absence problem (i.e., missing more work than a typical healthy employee), but at Post only 7% had a work absence problem. Across all cases with productivity data (n = 3,845), the level of work productivity (Stanford Presenteeism Scale) improved by 35%. At the start of counseling, 34% of cases were initially classified as having a problem with work productivity (i.e., low performance and lack of focus), but at Post this rate was reduced to just 5% of cases. All primary results for each outcome were large size statistical effects. Moderator tests indicated the improvement in each outcome was generally consistent across different sub-groups of employees based on client age, sex, clinical use characteristics and other study context factors. Some moderator findings were identified for clinical issue, referral type (self or manager) and client sex having slightly different profiles on certain outcomes. Comparisons with past research, study design limitations, and best practices in conducting applied research on workplace mental health are discussed.