Journal of Nursing Administration:
Turnover, Staffing, Skill Mix, and Resident Outcomes in a National Sample of US Nursing Homes
Trinkoff, Alison M. ScD, RN; Han, Kihye PhD, RN; Storr, Carla L. ScD; Lerner, Nancy DNP, RN; Johantgen, Meg PhD, RN; Gartrell, Kyungsook MS, RN
Author Affiliations: Professor (Drs Trinkoff and Storr), Assistant Professor (Dr Lerner), Associate Professor (Dr Johantgen), and PhD candidate (Ms Gartrell), School of Nursing, University of Maryland, Baltimore Maryland; Assistant Professor (Dr Han), Red Cross College of Nursing, Chung-Ang University, Seoul, South Korea.
This study was funded by the National Council of State Boards of Nursing Grant R40009 (to Dr Trinkoff, principal investigator).
The authors declare no conflicts of interest.
Correspondence: Dr Trinkoff, School of Nursing, University of Maryland, 655 W Lombard St, Room 625, Baltimore, MD 21201 (firstname.lastname@example.org).
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.jonajournal.com).
OBJECTIVES: The authors examined the relationship of staff turnover to selected nursing home quality outcomes, in the context of staffing and skill mix.
BACKGROUND: Staff turnover is a serious concern in nursing homes as it has been found to adversely affect care. When employee turnover is minimized, better care quality is more likely in nursing homes.
METHODS: Data from the National Nursing Home Survey, a nationally representative sample of US nursing homes, were linked to Nursing Home Compare quality outcomes and analyzed using logistic regression.
RESULTS: Nursing homes with high certified nursing assistant turnover had significantly higher odds of pressure ulcers, pain, and urinary tract infections even after controlling for staffing, skill mix, bed size, and ownership. Nurse turnover was associated with twice the odds of pressure ulcers, although this was attenuated when staffing was controlled.
CONCLUSIONS: This study suggests turnover may be more important in explaining nursing home (NH) outcomes than staffing and skill mix and should therefore be given greater emphasis.
Aspects of the quality of care provided to the 1.44 million elderly and disabled Americans residing in NHs are monitored via resident outcomes.1 An array of nursing care quality indicators (QIs) (eg, adverse outcomes and NH deficiencies) have been found to be inversely related to staffing and skill mix.2-6 However, a recent systematic review demonstrated mixed evidence supporting such relationships.7 A focus on staff turnover is a gap in the literature exploring the relationships of staffing and skill mix to QIs in NHs.8 A consistent staffing ratio and skill mix can be challenging in the presence of high turnover.
In 2007, it was estimated that 41% of RNs, 50% of licensed practical nurses (LPNs) and 66% of certified nursing assistants (CNAs) turned over annually.9 Estimates from 2010 indicate turnover continues to be a concern, as rates have increased for RNs (62.8%) and remain high for LPNs (43%) and CNAs (55%).10 Variation in local and regional rates also occurs. For example, state-to-state annual turnover in 2011 varied widely across each position, with RN rates ranging from 15.6% to 70%, LPN rates from 0% to 50.4%, and CNA rates from 15% to 80.3%.11
Staff turnover remains a serious concern in NHs as it has been found to adversely affect care. RN turnover has been associated with higher hospitalization rates for infection, with the infection rate increased by 30% for every RN lost (per full-time equivalent [FTE]/100 beds).12 Licensed nurse turnover is associated with a decline in activities of daily living.13 High turnover rates for CNAs and nurses have been directly associated with all 14 Centers for Medicare & Medicaid Services (CMS) reported QIs in NHs.14 This relationship remained for at least 4 QIs (physical restraint use, catheter use, pressure ulcers [PUs] and pain management) when agency usage and total staffing were controlled.8 This suggests that additional attention needs to be paid to the impact of turnover on outcomes, when considering other organizational factors such as staffing and skill mix. Data support that better care and a greater appreciation for when and how resident care should be delivered are more likely to occur when there is a stable set of employees.8
Our study examined the relationship between turnover and several NH QIs, representing resident outcomes, in the presence of skill mix and staffing, among a national sample of NHs. The conceptual framework was based on balance theory, which provides a human factors engineering perspective for studying nursing care in relation to resident outcomes.15,16 In balance theory, excessive demands (such as inadequate staffing) increase the “load” of an employee and adversely affect job performance, unless these demands are balanced by enhancing more positive aspects of the job and/or by reducing demands. If demands are not reduced or offset, employees may not be able to provide adequate care or may choose to leave their jobs, resulting in turnover of the position. We hypothesized higher adverse outcome rates would be found in NHs where turnover is high, and staffing and skill mix were lower.
Design and Data Sources
This study was a secondary analysis of cross-sectional data from 2 public nationally representative databases. The NH quality measures were in the form of resident outcomes obtained from resident assessment data (RAI).17 These data are collected through the NH Minimum Data Set and converted by CMS into Quality measures for Nursing Home Compare.18 Facility data on CNA/licensed nurse turnover, staffing, and skill mix were obtained from the 2004 National Nursing Home Survey (NNHS) conducted by the National Center for Health Statistics (NCHS).18 Data were available from the 1174 NHs of 1500 selected, using a multistage sampling scheme stratified on several facility and geographic characteristics. These NHs are nationally representative of the 16628 NHs in the United States.18 The NCHS Research Data Center (NCHS-RDC) created a restricted-use data set specifically for this study by linking facility-specific NNHS data to MDS resident outcomes data using the NH identifier.19 Thirty-two of the NHs in the NNHS sample did not have MDS data and were therefore excluded. Approval for the study was obtained from NCHS-RDC and the University of Maryland, Baltimore, institutional review board.
The resident outcome measures generated by CMS were risk-adjusted facility-level rates19 produced as part of Nursing Home Compare.18 These QIs were developed based on data from the RAI/MDS17 along with newer quality measures released in 2002 that incorporated a regression-based risk adjustment process that differs for each outcome measure. For this study, we used QIs from MDS version 2.0 that reflect skin care, pain management, and bowel and bladder health (Table 1). These outcomes have previously been used to evaluate NH quality.19,20
The NNHS provided turnover data calculated for a 3-month period, using formulas derived by Temple et al21 for licensed nurses (RNs + LPNs) as well as for CNAs. Turnover included both voluntary (resignation) and involuntary (termination) departures calculated for full- and part-time staff on all shifts. CNA turnover was defined as the total CNA FTE for those who left during the past 3 months (full- and part-time, with an assumption of part-time as 50% FTE) divided by the total number of CNA FTEs, expressed as a percentage.21-23 In addition, licensed nurse turnover was calculated for RNs plus LPNs using this same formula. We excluded 145 and 128 NHs with missing CNA or licensed nurse turnover data, respectively; 2 additional NHs were excluded for illogical values on turnover rate. The NH characteristics of our sample showed no differences from the overall US estimates for skill mix (ie, licensed proportion), staffing, bed size, or profit status (data not shown).
Staffing and Skill Mix
Variables were obtained from NNHS and include data on licensed nurse (RN + LPN), CNA, and total staffing. For this study, staffing was defined as total nurse staffing; this was equal to the sum of the RN, LPN, CNA, and aides total hours per resident day (HPRDs) and was dichotomized at 5.0+ hours versus less than 5.0 HPRDs.18,21 A cut point of 5.0 or more hours reflected higher nursing care hours for residents than the 4.0 hours recommended by CMS.24 Using this more stringent cut point allowed us to compare highly staffed NHs to those with average or lower staffing levels. Skill mix was calculated by dividing licensed nurse FTEs (RN + LPN) by all nursing staff FTEs (RN, LPN, CNA, and aides).
Control and Contextual
Variables related to NHs were included in the analyses to remove effects of any potentially confounding variables related to resident outcomes, staffing, skill mix, and turnover. These included NH bed size (3-49, 50-99, 100-199 vs ≥200 beds as the reference group) and profit status (for-profit vs all others).21,25,26
Facilities were the unit of analysis to evaluate the relationship of CNA and licensed nurse turnover to resident outcomes. Analyses were conducted through the remote access system (ANDRE) of the NCHS-RDC, using SAS-callable SUDAAN version 10.0.1 (Research Triangle Institute International, Research Triangle Park, North Carolina). Sampling weights were applied using SUDAAN to correct for design effects of the complex sampling strategy used in the NNHS. Although these data undergo extensive cleaning prior to release, additional data quality and completeness checks were conducted, and frequencies were compared with benchmarks from the nonconfidential (public use) NNHS data. This began with descriptive analyses to explore associations between CNA and licensed nurse turnover and NH resident outcomes and to test the assumptions of the statistical models. Using Proc Rlogist in SUDAAN for logistic regression, facility-level resident outcomes were regressed onto turnover variables with adjustment for nurse staffing and skill mix.
Logistic regression was chosen based on interpretability of the odds ratios (ORs), although results were similar across the models using both linear and logistic regression. Findings were significant if the 95% confidence interval (CI) did not include 1.0. Each resident outcome indicator was dichotomized at the 75th percentile to define NHs with adverse outcome rates (>;75th percentile) versus all others. Turnover rates were categorized as high (>;75th percentile) or medium (>;25th to ≤75th percentile), with low as the reference group (≤25th percentile).
According to our data, average CNA and licensed nurse 3-month turnover rates in US NHs were 16% and 12%, respectively (see Table, Supplemental Digital Content 1, http://links.lww.com/JONA/A273). This would be comparable to an annual turnover rate of 65% for CNAs and 47% for licensed nurses, assuming turnover rates are relatively stable across 1 year. CNA and licensed nurse turnover were weakly correlated (r = 0.29). The skill mix (licensed nurse proportion out of total nursing complement) was 34% for the sample. As far as total nurse staffing, 88% of NHs had less than 5 nursing care HPRDs. Most NHs had 50 to 199 beds (80%) and were for-profit facilities (62%). CNA turnover was highest in NHs with 50 to 99 and 100 to 199 beds (both 17% on average), compared with homes with less than 50 beds (11%) and homes with 200 or more beds (13%) (Table 2). CNA turnover was also significantly higher in for-profit NHs averaging 19% compared with 12% in not-for-profit NHs. Licensed nurse turnover was also higher in for-profit NHs (mean, 13.2%) than not-for-profit homes (mean, 8.9%) but did not differ by bed size.
Model 1 (Table 3) contains the crude estimates relating turnover to resident outcomes. To examine the relationship to staffing and skill mix, as well as potentially confounding influences of NH contextual variables, model 2 (Table 3) added adjustment for staffing and skill mix, NH bed size, and profit status. Resident outcomes data from NH Compare were suppressed by MDS using their rules requiring suppression of indicators with less than 11 observations. As this was more likely to occur in smaller NHs, we also ran models that excluded NHs with less than 50 beds, and no changes occurred in the findings. We therefore present the models for all NHs in the results.
According to model 1, NHs with high CNA turnover had significantly higher rates for 3 of the 6 resident outcomes as compared with NHs with low turnover (Table 3). Control for skill mix, staffing, ownership, and bed size did not alter these findings (model 2). As seen in Table 3, NHs with high CNA turnover were significantly more likely to have higher rates of low-risk PUs (OR, 2.49; 95% CI, 1.29-4.82), pain (OR, 2.69; 95% CI, 1.60-4.53), and urinary tract infection (UTI) (OR, 1.66; 95% CI, 1.02-2.72) than those with low CNA turnover, after adjusting for nurse staffing and skill mix along with bed size and ownership. Adverse outcome rates for homes with medium CNA turnover were not significantly different from those for homes with low turnover. For licensed nurse turnover, high turnover was associated with twice the odds of high-risk PUs compared with low turnover (OR, 2.05; 95% CI, 1.11-3.80). However, this relationship was no longer significant after adjustment for staffing, skill mix, bed size, and ownership, suggesting that these covariates explained some of these findings (Table 3). Higher nurse staffing was independently related to lower catheter use in the presence of CNA turnover (OR, 0.43; 95% CI, 0.20-0.93) and licensed nurse turnover (OR, 0.41; 95% CI, 0.19-0.88).
We provided national estimates of 3-month turnover for CNAs and licensed nurses (RNs + LPNs). Our annual turnover estimates (multiplying the quarterly rates by 4) were quite consistent with other studies and are also reflective of national estimates. For example, we estimated 65% annual turnover for CNAs compared with 70% by Decker27 and 46% licensed nurse annual turnover compared with 49% to 50% by Decker.27 In addition, we modeled turnover separately for CNAs and licensed nurses, while controlling for staffing and skill mix, and found that CNA turnover was related to NH outcomes independent of staffing and skill mix, although licensed turnover was not. In summary, although many large-scale studies find higher skill mix improves care,28,29 our work, as has others, suggests that CNA turnover and its influence on staffing may be more important to address as a potential explanation for these findings.
Although NHs may have sufficient nursing staff, high turnover and the related decreased staffing stability could increase the complement of inexperienced personnel, hindering quality care environments in NHs. Analysis of data from the 2004 National Nursing Assistant Survey showed that almost half of CNAs made less than $10 per hour, or less than $20,000 annually, and those with experience reported only minimal differences in pay compared with those entering the CNA workforce.30 Licensed nurses have a mixed experience regarding pay: RNs tend to receive less pay in NHs compared with other workplaces, whereas LPNs receive more in NHs compared with other settings.31 Increasing pay for CNAs and licensed staff in NHs can be difficult because of low reimbursement rates for patients covered by Medicaid, which pays for 70% of NH care, with prospective Medicare payment covering another 20%.32 Nonetheless, preventing costly PUs could mitigate some of this cost to reduce turnover. Estimated costs of treating PUs escalate rapidly depending on the stage: $2,000 for stage I, $3,000 to $10,000 for stage II, $5,900 to $14,840 for stage III, and $18,730 to $21,410 for stage IV.33
While pay considerations are significant, to reduce CNA turnover, it is also critical to enhance the work environment for NH staff because this is also essential for improving resident care outcomes.21 Although an initiative designed to improve retention of workers and reduce vacancies, Better Jobs, Better Care, showed promise in addressing turnover through global workplace changes, more focused efforts are required.34 Castle35 found that providing consistent resident assignments for CNAs led to a pronounced decrease in turnover rates. Similarly, establishment and maintenance of good relationships with residents were important for quality of care by CNAs.36
Findings should also be considered in light of limitations. Cross-sectional secondary data were used, limiting us to previously gathered information without causal conclusions. In regard to turnover, we were unable to differentiate between voluntary and involuntary separations. Voluntary turnover relates to employee perceptions and opportunities, whereas involuntary turnover is associated with the economic environment or employee competency.37 The 3-month period could be influenced by seasonality.38 However, our annualized turnover rates are quite similar to other annual estimates,27 and shorter turnover time frames have led to improved recall.22 Data were collected in 2004 and may not reflect the latest developments in long-term care or shifts in the economy. In addition, some have questioned the validity of staffing information in these kinds of data sources. Nonetheless, NNHS has allowed many critical questions to be addressed across hundreds of NHs, supporting continued use of these data.26,39,40
Findings suggest that attention to reducing staff turnover in NHs might improve resident outcomes. As the care demands of NH residents become increasingly more complex, the need for a well-trained and stable NH workforce will become even more critical.41 Creation and support of a stable workforce benefits facilities and residents, making turnover reduction an important priority for NHs.
1. Houser A, Fox-Grage W, Gibson MJ. Across the States: Profiles of Long-Term Care and Independent Living. 8th ed. Washington, DC: American Association of Retired Persons; 2009.
2. Alexander GL. An analysis of nursing home quality measures and staffing. Qual Manag Healthc. 2008; 17 (3): 242–251.
3. Grabowski DC, Stevenson DG. Ownership conversions and nursing home performance. Health Serv Res. 2008; 43 (4): 1184–1203.
4. Hyer K, Thomas KS, Branch LG, Harman JS, Johnson CE, Weech-Maldonado R. The influence of nurse staffing levels on quality of care in nursing homes. Gerontologist. 2011; 51 (5): 610–616.
5. Kim H, Kovner C, Harrington C, Greene W, Mezey M. A panel data analysis of the relationships of nursing home staffing levels and standards to regulatory deficiencies. J Gerontol B Psychol Sci Soc Sci. 2009; 64 (2): 269–278.
6. Harrington C, Olney B, Carillo H, Kang T. Nursing staffing and deficiencies in the largest for-profit nursing home chains and chains owned by private equity companies. Health Serv Res. 2012; 47: 106–128.
7. Spilsbury K, Hewitt C, Stirk L, Bowman C. The relationship between nurse staffing and quality of care in nursing homes: a systematic review. Int J Nurs Stud. 2011; 48 (6): 732–750.
8. Castle NG, Anderson RA. Caregiver staffing in nursing homes and their influence on quality of care: using dynamic panel estimation methods. Med Care. 2011; 49 (6): 545–552.
9. American Health Care Association (AHCA). Report of Findings 2007 AHCA Survey Nursing Staff Vacancy and Turnover In Nursing Facilities. Washington, DC: AHCA Department of Research; 2008.
10. American Health Care Association (AHCA). Report of Findings 2010 AHCA Survey Nursing Staff Vacancy and Turnover In Nursing Facilities. Washington, DC: AHCA Department of Research; 2011.
11. AHCA. Report of Findings 2011 Nursing Facility Staff Vacancy, Retention and Turnover Survey. Washington, DC: AHCA Department of Research; 2012.
12. Zimmerman S, Gruber-Baldini AL, Hebel JR, Sloane PD, Magaziner J. Nursing home facility risk factors for infection and hospitalization: importance of registered nurse turnover, administration, and social factors. J Am Geriatr Soc. 2002; 50 (12): 1987–1995.
13. Horn SD, Sharkey SS, Hudak S, Smout RJ, Quinn CC, Yody B. Beyond CMS quality measure adjustments: identifying key resident and nursing home facility factors associated with quality measures. J Am Med Dir Assoc. 2010; 11 (7): 500–505.
14. Castle NG, Engberg J, Men A. Nursing home staff turnover: impact on nursing home compare quality measures. Gerontologist. 2007; 47 (5): 650–661.
15. Wilson JR. Fundamentals of ergonomics in theory and practice. Appl Ergon. 2000; 31 (6): 557–567.
16. Carayon P, Gurses AP. A human factors engineering conceptual framework of nursing workload and patient safety in intensive care units. Int Crit Care Nurs. 2005; 21 (5): 284–301.
17. Zimmerman DR, Karon SL, Arling G, Clark BR, Collins T, Ross R. Development and testing of nursing home quality indicators. Health Care Financ Rev. 1995; 16 (4): 107–127.
20. Nakrem S, Vinsnes AG, Harkless GE, Paulsen B, Seim A. Nursing sensitive quality indicators for nursing home care: international review of literature, policy and practice. Int J Nurs Stud. 2009; 46 (6): 848–857.
21. Temple A, Dobbs D, Andel R. Exploring correlates of turnover among nursing assistants in the National Nursing Home Survey. Health Care Manag Rev. 2009; 34 (2): 182–190.
22. Castle NG. Measuring staff turnover in nursing homes. Gerontologist. 2006; 46 (2): 210–219.
23. Donoghue C. Nursing home staff turnover and retention: an analysis of national level data. J Appl Gerontol. 2010; 29 (1): 89–106.
25. Castle NG, Engberg J. Staff turnover and quality of care in nursing homes. Med Care. 2005; 43 (6): 616–626.
26. Decker FH, Castle NG. Relationship of the job tenure of nursing home top management to the prevalence of pressure ulcers, pain, and physical restraint use. J Appl Gerontol. 2011; 30( 5): 539–561.
27. Decker FH. The relationship of nursing staff to the hospitalization of nursing home residents. Res Nurs Health. 2008; 31 (3): 238–251.
28. Shin JH, Hall H. Relationship between nursing staffing and quality of life in nursing homes [published online ahead of print February 25, 2013]. Contemp Nurse.
29. Weech-Maldonado R, Meret-Hanke L, Neff MC, Mor V. Nurse staffing patterns and quality of care in nursing homes. Health Care Manag Rev. 2004; 29 (2): 107–116.
30. Squillace MR, Remsburg RE, Harris-Kojetin LD, Bercovitz A, Rosenoff E, Han B. The National Nursing Assistant Survey: improving the evidence base for policy initiatives to strengthen the certified nursing assistant workforce. Gerontologist. 2009; 49 (2): 185–197.
33. Spetz J, Brown D, Aydin C, Donaldson N. The value of reducing hospital-acquired pressure ulcer prevalence. JONA. 2013: 43 (4): 235–241.
34. Bishop CE, Weinb, erg DB, Leutz W, Dossa A, Pfefferle SG. Nursing assistants’ job commitment: effect of nursing home organizational factors and impact on resident well-being. Gerontologist. 2008; 48 (suppl 1): 36–45.
35. Castle NG. Consistent assignment of nurse aides: association with turnover and absenteeism. J Aging Soc Pol. 2013; 25 (1): 48–64.
36. Bowers BJ, Esmond S, Jacobson N. Turnover reinterpreted CNAs talk about why they leave. J Gerontol Nurs. 2003; 29 (3): 36–43.
37. Abelson MA, Baysinger BD. Optimal and dysfunctional turnover: toward an organizational level model. Acad Manag Rev. 1984; 9: 331–341.
38. Barry TT, Kemper P, Brannon SD. Measuring worker turnover in long-term care: lessons from the better jobs better care demonstration. Gerontologist. 2008; 48 (3): 394–400.
39. Bercovitz A, Decker FH, Jones A, Remsburg RE. End-of-life care in nursing homes: 2004 National Nursing Home Survey. National Center for Health Statistics Report. Hyattsville, MD: U.S. Department of Health and Human Services, CDC, NCHS. 2008: 9: 1–23.
40. Dwyer LL, Han B, Woodwell DA, Rechtsteiner EA. Polypharmacy in nursing home residents in the United States: results of the 2004 National Nursing Home Survey. Am J Geriatr Pharmacother. 2010; 8 (1): 63–72.
41. McConnell ES, Lekan D, Corazzini KN. Assuring the adequacy of staffing of long-term care, strengthening the caregiving workforce, and making long-term care a career destination of choice. NC Med J. 2010; 71 (2): 153–157.
Supplemental Digital Content
© 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins