TABLE 2.

Studies Assessing Interventions of Wellness Programs

Rigor Score	Author	Design	Modality	Subject(s)	Number of Employees	Key Findings	Limitations
High	Sorensen et al21	RCT	Cancer prevention: healthful eating, physical activity, tobacco control, and occupational health	24 multiethnic manufacturing businesses	50–150	The intervention group scored higher on all outcomes (physical activity, produce consumption, red meat consumption and vitamin use) compared with control worksites. For fruit and vegetable consumption, the intervention was more effective among women, nonmanagers, and all other ethnic groups more so than whites. For red meat consumption, the intervention was more effective for women and less educated employees	Businesses not nationally representative, outcomes self-reported yet validated, statistic power to detect subgroup differences
Intermediate	Merrill et al24	Case study	Physical, emotional, and mental health	One business	<500	Employees improved on almost all categories of the well-being indices over 3 yrs: physical, emotional, and mental health, in addition to health behaviors, access to safe drinking water, affordable produce, and a safe exercise space. Greatest improvements were among those with a high-risk baseline, and older employees	No control group, selection bias
Intermediate	Merrill25	Case-cohort descriptive study	Physical health, mental health, healthy behavior, and access to health services	One intervention company and 1276 control employees	<500	Four well-being indices improved significantly in the intervention group. Nevertheless, there was no significant difference of self-perceived access to basic needs. There was a positive association between income and scores on the well-being indices. Perceived physical health was greater for men than for women, yet perceived healthy behaviors were opposite	Nonrandomized intervention group, selection bias
Intermediate	Merrill26	Intervention evaluation	Exercise, sleep, stress, and diet programs	Five businesses	<500	Exercise, diet, sleep, health perception, and life satisfaction had significant improvements. Smallest companies (<50 employees) had greater fruit consumption, nights of restful sleep, smoking cessation, and job satisfaction. Participation was lower among those with poor health and men. Job satisfaction decreased significantly. There was a positive association between health perception and health behaviors	Nonrandomized (selection bias), no control group, behavior and health outcome data self-reported
Low	Devine et al45	Process evaluation	Walking and eating	Five businesses	<500	There was a positive association between site-specific intervention reach, dose, and employee achievement of health goals (walking and eating habits). The two sites with fewer than 25 employees had greater reach. High reach was also associated with younger, more educated, and mixed job-type employees. In the three sites with the lowest implementation scores, competing priorities for leaders’ time was noted as a barrier	Nonrandomized, small sample size, potential positive response bias, self-reported and recall bias
Low	Allen et al46	Intervention evaluation	Heart disease prevention: screening, education, and physical activity	29 intervention employees and 31 control employees	172	After 1 yr, the intervention group scored better on health outcomes compared with the comparison group. For cost effectiveness, there was a $10.2 per percentage point reduction in LDL cholesterol and a $454.2 per percentage point reduction in heart disease risk	Nonrandomized, selection bias, not true RCT, controls received some exposure to intervention, only one business
Low	Erfurt and Holtyn29	Analytic	Screenings, referral, education, and counseling	Three businesses	5–296	After the program, the employees showed improvements in blood pressure, cholesterol, cigarette smoking, weight control, and oxygen uptake	Only three worksites compared with much variability, no control group, low response rate
Very low	Taylor et al	Analytic: pre-/postassessment	Physical activity	14 employees	<15	After the program, a reduction in HDL cholesterol and weight was shown among employees	No control group, strong wellness culture before intervention, no correction for multiple comparisons

RCT, randomized controlled trial