Table 2 Characteristics of studies
# | Authors (year), country | Study design/ sample/data collection time | Measures/mean values of nurse turnover | Measures of factors | Data analysis/ quality score/quality | Main findings (p < .05) |
|---|---|---|---|---|---|---|
1 | Brewer et al. (2012) [16], US | Longitudinal panel design/ 1653 RNs/ 2006, 2007 | Whether at time two the NLRNs had stayed or left their employers identified at time one/14.5% (external turnover rate), 13.3% (internal turnover rate) | Personal characteristics, work attributes, opportunity, shocks, work attitudes [33, 34] | Binomial probit regression/ 11/high | - Work attributes (more than one job, working full time, voluntary overtime), shocks (sprains or strains), and work attitudes (intent to stay) influenced turnover. - Non-significant relationships between personal characteristics (age, gender, ethnicity, first language, marital status, children less than 6 years old, overall health, positive/negative affectivity, spousal income, income from other sources, first nursing degree, times took NCLEX), work attributes (staff/general duty, first RN job: formal orientation, first RN job: reduced workload, imputed wage, RN benefits-formal education, total number of RN benefits, number of months worked in current job, number of shifts floating, previous work experience, externship, internship, Magnet hospital, mandatory overtime, patient load, type of shift, typical work schedule, unit type, number of times change in supervisor, part of a union, importance of benefits from RN job), opportunity (local/non-local job opportunity, percent of all persons in HMOs, beds per 1000 population, unemployment rate, MSA), shocks (needle sticks, verbal abuse), work attitudes (job search behavior, job satisfaction, organizational commitment, variety, autonomy, mentor support, supervisory support, workgroup cohesion, distributive justice, procedural justice, promotional opportunities, work motivation, quantitative workload, organizational constraints, collegial RN-MD relations, work family conflict, family work conflict) and nurse turnover were found. |
2 | Cho et al. (2012) [21], South Korea | Longitudinal design/ 351 RNs/ 2006, 2007, 2008 | RN-reported turnover/45% (before the 3rd year) | Age, gender, marital status, father’s education, nursing education, hospital size, location, presence of labor unions, job satisfaction | Cox proportional hazards regression analysis/ 9/medium | - Married RNs working in small, nonmetropolitan areas, nonunionized hospitals reported higher turnover. Several aspects of job dissatisfaction (work content, physical work environment, interpersonal relationships) were significantly associated with turnover. - Non-significant relationships between age, gender, father with 4+ years college, nursing education (nursing degree, reason for choosing nursing major), job dissatisfaction (working hours, social insurances and fringe benefits, advancement system, social reputation for the job) and nurse turnover were found in univariate analysis. |
3 | Halfer (2011) [27], US | Descriptive longitudinal design/ 116 RNs/2008, 2009 | Whether the new graduate nurses stayed or left the organization between 2008 and 2009/12% (within a year) | Personal characteristics, job embeddedness factors [35] | Logistic regression modeling/ 6/medium | - Nurses who reported higher turnover were younger and had significantly lower agreement on two items of organizational embeddedness, which are: “I feel part of my work team” and “I feel as if I am a good match for this hospital”. - Non-significant relationships between personal characteristics (gender, marital status, number of children living at home, organizational tenure, education, ethnicity, part-time or full-time status, length of commute, whether nursing was a second career, and current position), job embeddedness (community embeddedness, organizational and community embeddedness item summed scores) and nurse turnover were found. |
4 | Han et al. (2019) [17], South Korea | Descriptive and prospective longitudinal study design/ 464 RNs/the first day of orientation before ward placement (between September 2014 and December 2015), 6 weeks after starting work, and at the 6th, 12th, 18th, and 24th month of work | Whether the nurse participants had stayed on or resigned by December 31, 2017/not reported | Pre-employment health lifestyle variables: sleep [36], diet behaviors, physical activity (Korean International Physical Activity Questionnaire Short Form [37, 38]), alcohol consumption [39], depressive symptoms (the Korean Center for Epidemiologic Studies-Depression scale [40, 41];), and self-rated health [42] , job stress (demands, control, and support) | Cox proportional hazards regression/ 7/medium | - Compared with the new nurses in the discordant group (low levels of sleep disturbance, depression, and poor self-rated health but relatively high levels of poor diet and physical inactivity behaviors), those in the unhealthy lifestyle group (high levels of poor sleep quality, not eating 3 meals a day, irregular diet, depression, and poor self-rated health) had significantly higher probabilities of resigning. - Non-significant relationships between job stress (psychological demand, physical demand, control, boss support, peer support) and nurse turnover were found. |
5 | Han et al. (2020) [28], South Korea | Prospective longitudinal design/ 465 RNs/the first day of orientation before ward placement (baseline, T0, between September 2014 and December 2015), 6 weeks after starting work (T1), and at the 6th, 12th, 18th, and 24th month of work (T 2–5) | Whether the new nurses had left the hospital at each data collection point/ 2.6% (until T1), 9.2% (until T2), 12.3% (until T3), 18.7% (until T4) 23% (until T5) | Personal characteristics (age, gender, marital status, highest degree achieved, alcohol consumption, smoking), trajectories of sleep disturbance (General Sleep Disturbance Scale [43, 44]) classified into the high/low symptomatic group | Latent growth curve analysis/ 7/medium | - Turnover rates of the high symptomatic group of sleep disturbance trajectories were higher than those of the low symptomatic group, and sleep disturbance was more severe among leavers than among stayers in the high symptomatic group. - Nurse turnover was not different according to the personal characteristics (age, gender, marital status, highest degree achieved, alcohol consumption, smoking) in univariate analysis. |
6 | Kim & Lee (2016) [29], South Korea | Secondary data analysis with retrospective study/ 323 RNs/ 2010 GOMS (2011, 2013) | Whether the new nurses had left their first job/24.5% (within the 1st year), 39% (within the 2nd year), 47.4% (within the 3rd year), 51.7% (within the 4th year) | General characteristics, job-related characteristics, satisfaction, characteristics of region | Multilevel survival analysis/ 8/medium | - Job-related characteristics (job status, monthly income), satisfaction (physical work environments, work hours, interpersonal relationship, social insurance and fringe benefits), and characteristics of region (number of hospitals in region, number of nurses per 100 beds) were significant predictors of turnover among new nurses. - Non-significant relationships between general characteristics (gender, age), job-related characteristics (number of employees), satisfaction (pay or income, stability of employment, potential for personal growth, advancement system, social reputation for the job, autonomy and authority for the job), characteristics of region (number of beds per 1000 population) and nurse turnover were found. |
7 | Lee (2019) [22], South Korea | Longitudinal panel design/ 652 RNs/ 2008 GOMS (2009, 2011), 2009 GOMS (2010, 2012) 2010 GOMS (2011, 2013) | Whether newly graduated nurses transferred profession or organization/25% (within the 1st year), 50% (within the 4th year) | Individual factors, organizational factors, job satisfaction | Cox regression analysis/ 8/medium | - Individual factors (graduation year, gender), hospital factors (interaction of hospital size and monthly salary, presence of union), dissatisfaction with organization and profession were major factors that affected the turnover rate. - Non-significant relationships between individual factors (gender, marital status, educational level of father, family income, admission type, selection reason), hospital factors (hospital location, moving direction for working, shift), job dissatisfaction (salary, welfare benefit, performance appraisal) and nurse turnover were found in univariate analysis. |
8 | Suzuki et al. (2006) [30], Japan | Longitudinal design/ 923 RNs/ 2003 (June), 2003 (December) | Turnover between novice nurses who quit their jobs between June 2003 to December 2003/4% | Nurse attributes, clinical department, burnout (the Japanese version of the MBI [45, 46]), assertiveness (the Japanese version of RAS [47, 48]), stressful life events (Social Readjustment Rating Scale [49]), reality shock, ward assignment preference, transfer preference, job satisfaction, social support and coping mechanisms [50] | Multiple logistic regression with the stepwise variable selection method/ 8/medium | - Graduation from vocational nursing school, dissatisfaction about being assigned to a different ward, and lack of social support from peers were the factors affecting rapid turnover. - Non-significant relationships between nurse attributes (gender, age, living arrangements, location of the hospital, nursing arrangement, ward, department), burnout (emotional exhaustion and depersonalization, personal accomplishment), assertiveness, stressful life events, reality shock, job satisfaction (salary), workload, overtime, social support, coping and nurse turnover were found in univariate analysis. |
9 | Suzuki et al. (2008) [31], Japan | Longitudinal design/ 923 RNs/2003 (June), 2003 (December), 2004 (June), 2005 (March) | Turnover between novice nurses who quit their jobs between June 2003 and March 2005/12.7% | Nurse attributes, burnout, assertiveness, stressful life events, reality shock, ward assignment preference, transfer preference, job satisfaction, social support and coping mechanisms | Cox’s proportional hazard model with the stepwise variable selection method/ 8/medium | - Support from peers and dissatisfaction with the workplace affected turnover of novice nurses. - Non-significant relationships between nurse attributes (gender, age, highest education level, living arrangements, department), burnout (emotional exhaustion and depersonalization, personal accomplishment), assertiveness, stressful life events, job satisfaction (salary), workload, overtime, social support, coping and nurse turnover were found in univariate analysis. |
10 | Suzuki et al. (2010) [32], Japan | Longitudinal design/ 762 RNs/2003 (December), 2004 (June) | Turnover between novice nurses who quit their jobs between December 2003 and June 2004/4.6% | Demographic attributes, burnout, assertiveness, stressful life events, reality shock, ward assignment preference, transfer preference, job satisfaction, social support and coping profiles | Multiple logistic regression analysis with the stepwise variable selection method/ 8/medium | - Hospital location being in Tokyo, burnout (physical exhaustion, emotional exhaustion and depersonalization), and dissatisfaction with the workplace affected turnover of novice nurses. - Non-significant relationships between demographic attributes (gender, age, education, ward, department), burnout (personal accomplishment), assertiveness, stressful life events, reality shock, ward assignment preference, job satisfaction (salary), workload, overtime, social support, coping and nurse turnover were found in univariate analysis. |