How Does Rurality Influence the Staffing of Social Service Departments in Nursing Homes?

How Does Rurality Influence the Staffing of Social Service Departments in Nursing Homes? Abstract Purpose of the Study Social service departments in nursing homes (NHs) are staffed by qualified social workers (QSWs) and paraprofessionals. Due to greater workforce challenges in rural areas, this article aims to describe the staffing levels and composition of these departments by rurality. Design and Methods Certification and Survey Provider Enhanced Reports data from 2009 to 2015 are used to examine the effect of rurality on social service staffing using random-effects linear panel regressions. Results The average NH employed 1.8 full-time equivalents (FTEs), with approximately two thirds of social services staffed by QSWs. Large NHs had more staff, but employed fewer staff hours per resident day. Staffing levels were lower and QSWs made up a smaller percentage of staff in rural areas. Implications National trends indicate variability in staffing by NH size and degree of rurality. Very low staffing within rural NHs is a concern, as staff may have less time to respond to residents’ needs and these NHs may utilize fewer QSWs. Social services, Workforce issues, Social work, Nursing homes, Staffing levels With the aging of the baby boomer generation, a growing number of older Americans will need nursing home (NH) care. In order to maintain and provide high-quality care, NHs need highly trained workers, including those providing social services. Social services play a vital role in developing care plans, assist in discharge planning, and help residents deal with a variety of psychosocial needs, along with other functions (Bern-Klug & Kramer, 2013). Within NHs, psychosocial care provided by social services should be “comprehensive in scope and address the psychological, social, emotional, and behavioral needs of residents” (Simons, Bern-Klug, & An, 2012, p. 800). Psychosocial care is becoming increasingly important due to the prevalence of depression and dementia within the NH population (Harris-Kojetin, Sengupta, Park-Lee, & Valverde, 2013), and the role psychosocial health can play in quality of life and care (Bowen & Zimmerman, 2008). Regulations are a key determinant of staffing in social service departments (Bern-Klug, 2008; Roberts & Bowblis, 2016). At the federal level, employment of a full-time professional is required only among NHs with more than 120 beds, and this individual may hold a degree outside of social work (State Operations Manual, 2016). States can exempt facilities or strengthen federal requirements (Bern-Klug, 2008), yet nationally, residents spend approximately 5 min a day interacting with a licensed social worker (Harris-Kojetin et al., 2013). Further, the National Association of Social Workers (NASW, 2016) has recently called for policies that require the employment of social workers and increase staffing levels in NHs. Even with the deficits in psychosocial care (Levinson, 2013; Rehnquist, 2003), there has not been enough focus by policy makers and researchers on social service departments. Social services are delivered by two types of staff: professionals and paraprofessionals. Though the roles of each are not well defined (Simons et al., 2012), professional social workers have higher educational, training, and licensure requirements (Bern-Klug et al., 2009). Both support psychosocial care, and commonly work together, yet nearly 11% of NHs in the United States rely exclusively on paraprofessionals (Roberts & Bowblis, 2016). This raises concerns about the qualifications, staffing levels, and the quality of social services. To understand national trends, we examine the effect of rurality on social service staffing in U.S. NHs. The rural context is crucial because other studies have found urban and rural differences in staffing and quality outcomes (Bowblis, Meng, & Hyer, 2013; Lutfiyya, Gessert, & Lipsky, 2013). Methods Empirical Strategy and Data Our empirical strategy is to examine how rurality influences the staffing of social service departments. To accomplish this, we estimate random-effects linear panel regressions with staffing as the dependent variable and rurality as the key explanatory variable. The Certification and Survey Provider Enhanced Reports (CASPER) provide data on NH characteristics and staffing. CASPER is a national database collected as part of an annual, required re-certification process for all NHs that receive Medicare or Medicaid funding. Our sample includes all CASPER surveys between 2009 and the first quarter for 2015 in free-standing NHs with at least 20 beds. We merged the CASPER data with rural–urban commuting area (RUCA) codes from the WWAMI Rural Health Research Center to measure rurality. The final analytic sample included 85,826 survey observations from 14,691 unique NHs. Variables CASPER data identify two levels of staff: CMS-Qualified Social Workers (QSWs) who earned at least a bachelor’s degree in social work or a human services field with at least 1 year of supervised social work experience in a health care setting working with the elderly; and paraprofessionals who have not completed a bachelor’s degree (Items F61 and F62 on CMS-Form 671). We measured staffing levels for overall social services, and separately for QSWs and paraprofessionals by full-time equivalents (FTEs) and hours per resident day (HPRD). HPRD adjusts for the number of residents in the facility and reflects the average number of hours a staff member could spend with a resident each day. We also examine the composition of departments as measured by the proportion staffed by QSWs. Utilizing WWAMI’s Categorization A of RUCA codes, we group NHs as urban or one of the three levels of rurality, including micropolitan city, small rural town, and isolated small rural town. From urban to the most rural, these categories make up 66.7%, 13.7%, 11.0%, and 8.6% of the sample, respectively. The regressions also control for other contextual and organizational factors that may affect staffing, such as facility characteristics, payer-mix, resident case-mix, availability of other staffing resources, and state and year fixed effects. Descriptive information about the covariates are presented in Table 1. Bivariate comparisons between rural and urban characteristics found statistically significant differences for all study variables (results not reported). Table 1. Descriptive Statistics All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 Notes: The descriptive statistics table above reports the mean and standard deviation (SD; in parentheses) for the sample. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The proportion of social service department staffed by CMS-QSWs was based on 83,828 observations due to some NHs not reporting any social services staff. Overall, 2.33% or 1,999 observations do not report any social service staff. The corresponding rates of NHs not reporting any social services staff for urban, micropolitan, small rural town, and isolated small rural town are 1.86%, 3.02%, 3.30%, and 3.61%, respectively. The reference group for ownership is for-profit. The Acuindex is a facility-level measure of resident acuity summed from the ADL index score and special treatment index (Cowles, 2003). Physician services are measured in four mutually exclusive categories: medical director only (reference group); a medical team consisting of physicians and nurse practitioners (NPs) or physician assistants (PAs); NPs and PAs only; and no services. The Herfindahl–Hirschman index is a measure of market competition for NH services within the county. A few variables included in the model were not reported in Table 1, including state fixed effects, year fixed effects, and errors rates in nursing or activities staffing. ADL = activities of daily living; CCRC = continuing care retirement community; CMS = Centers for Medicare and Medicaid Services; CNA = certified nurse aide; FTE = full-time equivalent; HPRD = hours per resident day; LPN = licensed practical nurse; NH = nursing home; QSW = qualified social worker; RN = registered nurse. View Large Table 1. Descriptive Statistics All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 Notes: The descriptive statistics table above reports the mean and standard deviation (SD; in parentheses) for the sample. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The proportion of social service department staffed by CMS-QSWs was based on 83,828 observations due to some NHs not reporting any social services staff. Overall, 2.33% or 1,999 observations do not report any social service staff. The corresponding rates of NHs not reporting any social services staff for urban, micropolitan, small rural town, and isolated small rural town are 1.86%, 3.02%, 3.30%, and 3.61%, respectively. The reference group for ownership is for-profit. The Acuindex is a facility-level measure of resident acuity summed from the ADL index score and special treatment index (Cowles, 2003). Physician services are measured in four mutually exclusive categories: medical director only (reference group); a medical team consisting of physicians and nurse practitioners (NPs) or physician assistants (PAs); NPs and PAs only; and no services. The Herfindahl–Hirschman index is a measure of market competition for NH services within the county. A few variables included in the model were not reported in Table 1, including state fixed effects, year fixed effects, and errors rates in nursing or activities staffing. ADL = activities of daily living; CCRC = continuing care retirement community; CMS = Centers for Medicare and Medicaid Services; CNA = certified nurse aide; FTE = full-time equivalent; HPRD = hours per resident day; LPN = licensed practical nurse; NH = nursing home; QSW = qualified social worker; RN = registered nurse. View Large Results The average NH employed 1.8 FTEs, of which 1.2 FTEs were QSWs and 0.6 were paraprofessionals, making approximately two thirds of social service departments staffed by QSWs. A total of 2.3% survey observations do not report any social service staffing. How these staffing levels change with NH size and rurality is illustrated in Figure 1. There are two notable trends. First, the overall staffing levels were positively correlated with NH size. Second, rural NHs were more likely to hire fewer FTEs regardless of NH size, although some variation exists around the smallest NHs and those close to 150 beds. Figure 1. View largeDownload slide Overall social service staffing levels (FTEs) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. FTE = full-time equivalent; NH = nursing home. Figure 1. View largeDownload slide Overall social service staffing levels (FTEs) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. FTE = full-time equivalent; NH = nursing home. Since rural NHs also tend to be smaller (Table 2), we also measured staffing levels in HPRD. Figure 2 shows how staffing levels decreased as NHs increased in size and that rural NHs had lower staffing levels. The smallest NHs had between 0.15 and 0.25 HPRD. However, as the size of the facility increased, staffing levels decreased until the NH had at least 100 beds. For NHs with greater than 100 beds, staffing levels ranged from 0.08 to 0.12 HPRD. Table 2. Number of Beds by Level of Rurality Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Note: The table reports the distribution of the number of beds for the overall sample and each level of rurality (n = 85,826). NH = nursing home; SD = standard deviation. View Large Table 2. Number of Beds by Level of Rurality Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Note: The table reports the distribution of the number of beds for the overall sample and each level of rurality (n = 85,826). NH = nursing home; SD = standard deviation. View Large Figure 2. View largeDownload slide Overall social service staffing levels (HPRD) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. HPRD = hours per resident day; NH = nursing home. Figure 2. View largeDownload slide Overall social service staffing levels (HPRD) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. HPRD = hours per resident day; NH = nursing home. To account for other confounding variables, Table 3 reports the regression results for all NHs with at least 20 beds. The number of FTEs decreased with rurality, as rural NHs had 0.08 fewer FTEs compared with urban NHs. Similarly, QSW staffing levels in rural areas were lower, ranging from 0.11 to 0.13 fewer FTEs. Not only was the staffing of social services lower in rural areas, but paraprofessionals may have been hired in place of QSWs, as rural NHs used more paraprofessionals. Examining the proportion of QSWs within NHs confirms this result. For the average NH in the sample, QSWs comprised nearly two thirds of social service departments. Compared with urban NHs, social services in micropolitan NHs consisted of fewer QSWs (4.5% points fewer), though the effect was largest among NHs in small isolated rural towns (7.5% points fewer). Table 3. Influence of Rurality on Social Service Staffing Levels and Composition Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Note: The table above reports the coefficient estimates and standard errors (SEs; in parentheses) of random-effects linear panel regression models for staffing levels and composition of social services for all NHs. The model also controls for NH characteristics, resident case-mix, other types of staffing, state fixed effects, and year fixed effects (see Table 1). The reference group is urban NHs. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The final sample is 85,826 observations from 14,658 NHs, except in the case of the proportion of social service department staffed by CMS-QSWs which has 83,828 observations. The difference in the number of observations is because some NHs reported no staffing in social services. SEs are adjusted for clustering at the facility level. CMS = Centers for Medicare and Medicaid Services; FTE = full-time equivalent; HPRD = hours per resident day; NH = nursing home; QSW = qualified social worker; SD = standard deviation. *p < .1. ** p < .05. *** p < .01. View Large Table 3. Influence of Rurality on Social Service Staffing Levels and Composition Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Note: The table above reports the coefficient estimates and standard errors (SEs; in parentheses) of random-effects linear panel regression models for staffing levels and composition of social services for all NHs. The model also controls for NH characteristics, resident case-mix, other types of staffing, state fixed effects, and year fixed effects (see Table 1). The reference group is urban NHs. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The final sample is 85,826 observations from 14,658 NHs, except in the case of the proportion of social service department staffed by CMS-QSWs which has 83,828 observations. The difference in the number of observations is because some NHs reported no staffing in social services. SEs are adjusted for clustering at the facility level. CMS = Centers for Medicare and Medicaid Services; FTE = full-time equivalent; HPRD = hours per resident day; NH = nursing home; QSW = qualified social worker; SD = standard deviation. *p < .1. ** p < .05. *** p < .01. View Large When adjusting for the number of residents, both overall and social worker staffing levels were lower the more rural a facility. Compared with urban NHs, micropolitan NHs had staffing levels that were 0.009 and 0.012 HPRD lower for overall and social worker staffing, respectively. The effects for the most isolated rural NHs were lower by nearly double, at 0.020 and 0.024 HPRD. Interestingly, the staffing levels of paraprofessionals in terms of HPRD were not statistically different by rurality. Discussion This study highlights the influence of rurality on the staffing levels and composition of NH social service departments. In our national sample, more FTEs were hired as facility size increased; however, larger facilities staff fewer HPRD. Furthermore, rural facilities were disadvantaged by having lower staffing levels than urban NHs, which produces higher workload in rural areas. To illustrate, overall staffing levels adjusted for the size of the facility were 8% to 22% lower in rural areas compared with urban NHs. This implies that staff in rural areas were expected to do more work if they were to maintain the same level of service quality as urban areas. Our findings also indicate that QSWs make up a smaller proportion of social service departments in rural areas, which may encourage the substitution of paraprofessionals for professionally trained and experienced staff. Very low staffing levels in rural areas are a concern, as workers may have less time to respond to the psychosocial needs of residents or alter the kinds of care provided. Though more research is needed on social service staffing, quality and resident outcomes, the existing literature suggests that better psychosocial care is associated with higher qualifications of social service staff (Simons, 2006; Vongxaiburana, Thomas, Frahm, & Hyer, 2011; Zhang, Gammonley, Paek, & Frahm, 2008). Staffing is strongly influenced by federal regulations that require one full-time CMS-QSW for large facilities, and do not currently address staffing levels relative to caseload. Guidelines based on a ratio of residents per social service staff, similar to minimum nursing staff ratios (Bowblis, 2011; Mueller et al., 2006; Park & Stearns, 2009), will support consistency in the structure of social services and take steps toward requiring a manageable caseload. Bern-Klug, Kramer, Sharr, and Cruz (2010) found that NH directors of social services recommend a caseload of 60 long-term stay residents or less than 20 residents in skilled care per FTE worker. Relative to this standard, social services are seriously understaffed, especially in rural areas. A significant challenge with minimum staffing levels is the limited availability of skilled professionals in the rural workforce and cost of hiring additional staff. Facing these challenges, rural NHs should increase professional development opportunities for existing staff, with attention to opportunities for mentoring, training, and higher education. Limitations Several limitations are noted. First, we examined two types of staff: QSWs and paraprofessionals. While CASPER defines a QSW as anyone with a college degree in related fields, at minimum, the NASW (2003, 2016) requires a bachelor’s degree in social work. Therefore, our study should be interpreted with CMS definitions and not the standards of the NASW. Second, CASPER reports staffing levels for the 14-day period before the inspection, and this may not reflect average staffing throughout the year. Third, there are a limited number of isolated rural NHs with over 150 beds. While our analyses used a sample of NHs with at least 20 beds, sensitivity analyses that limited the regressions to under 150 beds were found to be robust. Finally, we used HPRD to adjust for the number of residents. Analyses utilizing FTEs per 100 resulted in similar conclusions, but found greater use of paraprofessionals in rural areas. Conclusion Insufficient regulation and staffing continue to challenge how social services are provided in NHs (Simons et al., 2012). National trends indicate considerable staffing variation in social service departments by size of facility and degree of rurality. Modifications to policy at the state and federal level are needed to formalize the structure of NH social services, ensure realistic caseloads, and define specific responsibilities according to the educational preparation and experience of staff to provide high-quality social services. Lastly, awareness of the unique challenges of rural NHs should be part of ongoing advocacy and research efforts to support the role of social workers and paraprofessionals in delivering care in NHs. References Bern-Klug M . ( 2008 ). State variations in nursing home social worker qualifications . Journal of Gerontological Social Work , 51 , 379 – 409 . doi: 10.1080/01634370802039734 Google Scholar CrossRef Search ADS PubMed Bern-Klug M. Kramer K. W . ( 2013 ). Core functions of nursing home social service departments in the United States . Journal of the American Medical Directors Association , 14 , e1 – e7 . Google Scholar CrossRef Search ADS Bern-Klug M. Kramer K. W. Chan G. Kane R. Dorfman L. T. Saunders J. B . ( 2009 ). Characteristics of nursing home social services directors: How common is a degree in social work ? Journal of the American Medical Directors Association , 10 , 36 – 44 . doi: 10.1016/j.jamda.2008.06.011 Google Scholar CrossRef Search ADS PubMed Bern-Klug M. Kramer K. W. Sharr P. Cruz I . ( 2010 ). Nursing home social services directors’ opinions about the number of residents they can serve . Journal of Aging & Social Policy , 22 , 33 – 52 . doi: 10.1080/08959420903396426 Google Scholar CrossRef Search ADS PubMed Bowblis J. R . ( 2011 ). Staffing ratios and quality: An analysis of minimum direct care staffing requirements for nursing homes . Health Services Research , 46 , 1495 – 1516 . doi: 10.1111/j.1475-6773.2011.01274.x Google Scholar CrossRef Search ADS PubMed Bowblis J. R. Meng H. Hyer K . ( 2013 ). The urban-rural disparity in nursing home quality indicators: The case of facility-acquired contractures . Health Services Research , 48 , 47 – 69 . doi: 10.1111/j.1475-6773.2012.01431 Google Scholar CrossRef Search ADS PubMed Bowen S. E. Zimmerman S . ( 2008 ). Understanding and improving psychosocial services in long-term care . Health Care Financing Review , 30 , 1 – 4 . Google Scholar PubMed Cowles C. M . ( 2003 ). 2002 Home Statistical Yearbook . Montgomery Village, MD : Cowles Research Group . Harris-Kojetin L. Sengupta M. Park-Lee E. Valverde R . ( 2013 ). Long-term care services in the United States: 2013 overview. National Center for Health Statistics . Vital Health Statistics , 3 , 1 – 105 . Levinson D . ( 2013 ). Skilled nursing facilities often fail to meet care planning and discharge planning requirements . Washington, DC : Department of Health and Human Services, Office of the Inspector General . Lutfiyya M. N. Gessert C. E. Lipsky M. S . ( 2013 ). Nursing home quality: A comparative analysis using CMS Nursing Home Compare data to examine differences between rural and nonrural facilities . Journal of the American Medical Directors Association , 14 , 593 – 598 . doi: 10.1016/j.jamda.2013.02.017 Google Scholar CrossRef Search ADS PubMed Mueller C. Arling G. Kane R. Bershadsky J. Holland D. Joy A . ( 2006 ). Nursing home staffing standards: Their relationship to nurse staffing levels . The Gerontologist , 46 , 74 – 80 . doi: 10.1093/geront/46.1.74 Google Scholar CrossRef Search ADS PubMed National Association of Social Workers . ( 2003 ). Standards for social work services in long-term care facilities . Washington, DC : NASW Press . National Association of Social Workers . ( 2016 ). NASW comments to the Senate Committee on Finance Bipartisan Chronic Care Working Group . Retrieved from http://www.socialworkers.org/advocacy/letters/2016/Senate_Finance_chronic_care.pdf Park J. Stearns S. C . ( 2009 ). Effects of state minimum staffing standards on nursing home staffing and quality of care . Health Services Research , 44 , 56 – 78 . doi: 10.1111/j.1475-6773.2008. 00906.x Google Scholar CrossRef Search ADS PubMed Rehnquist J . ( 2003 ). Psychosocial services in skilled nursing facilities . Washington, DC : Department of Health and Human Services, Office of the Inspector General . Roberts A. R. Bowblis J. R . ( 2016 ). Who hires social workers? Structural and contextual determinants of social service staffing in nursing homes . Health & Social Work . Advance online publication . doi: 10.1093/hsw/hlw058 Simons K. V . ( 2006 ). Organizational characteristics influencing nursing home social service directors’ qualifications: A national study . Health & Social Work , 31 , 266 – 274 . doi: 10.1093/hsw/31.4.266 Google Scholar CrossRef Search ADS PubMed Simons K. V. Bern-Klug M. An S . ( 2012 ). Envisioning quality psychosocial care in nursing homes: The role of social work . Journal of the American Medical Directors Association , 13 , 800 – 805 . doi: 10.1016/j.jamda.2012.07.016 Google Scholar CrossRef Search ADS PubMed State Operations Manual . ( 2016 ). Appendix PP – Guidance to Surveyors for Long Term Care Facilities (Rev. 157, 06-10-16) . Retrieved from www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/som107ap_pp_guidelines_ltcf.pdf Vongxaiburana E. Thomas K. S. Frahm K. A. Hyer K . ( 2011 ). The social worker in interdisciplinary care planning . Clinical Gerontologist , 34 , 367 – 378 . doi: 10.1080/07317115.2011.588540 Google Scholar CrossRef Search ADS Zhang N. J. Gammonley D. Paek S. C. Frahm K . ( 2008 ). Facility service environments, staffing, and psychosocial care in nursing homes . Health Care Financing Review , 30 , 5 – 17 . Google Scholar PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Gerontologist Oxford University Press

How Does Rurality Influence the Staffing of Social Service Departments in Nursing Homes?

Loading next page...
 
/lp/ou_press/how-does-rurality-influence-the-staffing-of-social-service-departments-XwpPmJh0Qy
Publisher
Oxford University Press
Copyright
© The Author(s) 2017. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0016-9013
eISSN
1758-5341
D.O.I.
10.1093/geront/gnw208
Publisher site
See Article on Publisher Site

Abstract

Abstract Purpose of the Study Social service departments in nursing homes (NHs) are staffed by qualified social workers (QSWs) and paraprofessionals. Due to greater workforce challenges in rural areas, this article aims to describe the staffing levels and composition of these departments by rurality. Design and Methods Certification and Survey Provider Enhanced Reports data from 2009 to 2015 are used to examine the effect of rurality on social service staffing using random-effects linear panel regressions. Results The average NH employed 1.8 full-time equivalents (FTEs), with approximately two thirds of social services staffed by QSWs. Large NHs had more staff, but employed fewer staff hours per resident day. Staffing levels were lower and QSWs made up a smaller percentage of staff in rural areas. Implications National trends indicate variability in staffing by NH size and degree of rurality. Very low staffing within rural NHs is a concern, as staff may have less time to respond to residents’ needs and these NHs may utilize fewer QSWs. Social services, Workforce issues, Social work, Nursing homes, Staffing levels With the aging of the baby boomer generation, a growing number of older Americans will need nursing home (NH) care. In order to maintain and provide high-quality care, NHs need highly trained workers, including those providing social services. Social services play a vital role in developing care plans, assist in discharge planning, and help residents deal with a variety of psychosocial needs, along with other functions (Bern-Klug & Kramer, 2013). Within NHs, psychosocial care provided by social services should be “comprehensive in scope and address the psychological, social, emotional, and behavioral needs of residents” (Simons, Bern-Klug, & An, 2012, p. 800). Psychosocial care is becoming increasingly important due to the prevalence of depression and dementia within the NH population (Harris-Kojetin, Sengupta, Park-Lee, & Valverde, 2013), and the role psychosocial health can play in quality of life and care (Bowen & Zimmerman, 2008). Regulations are a key determinant of staffing in social service departments (Bern-Klug, 2008; Roberts & Bowblis, 2016). At the federal level, employment of a full-time professional is required only among NHs with more than 120 beds, and this individual may hold a degree outside of social work (State Operations Manual, 2016). States can exempt facilities or strengthen federal requirements (Bern-Klug, 2008), yet nationally, residents spend approximately 5 min a day interacting with a licensed social worker (Harris-Kojetin et al., 2013). Further, the National Association of Social Workers (NASW, 2016) has recently called for policies that require the employment of social workers and increase staffing levels in NHs. Even with the deficits in psychosocial care (Levinson, 2013; Rehnquist, 2003), there has not been enough focus by policy makers and researchers on social service departments. Social services are delivered by two types of staff: professionals and paraprofessionals. Though the roles of each are not well defined (Simons et al., 2012), professional social workers have higher educational, training, and licensure requirements (Bern-Klug et al., 2009). Both support psychosocial care, and commonly work together, yet nearly 11% of NHs in the United States rely exclusively on paraprofessionals (Roberts & Bowblis, 2016). This raises concerns about the qualifications, staffing levels, and the quality of social services. To understand national trends, we examine the effect of rurality on social service staffing in U.S. NHs. The rural context is crucial because other studies have found urban and rural differences in staffing and quality outcomes (Bowblis, Meng, & Hyer, 2013; Lutfiyya, Gessert, & Lipsky, 2013). Methods Empirical Strategy and Data Our empirical strategy is to examine how rurality influences the staffing of social service departments. To accomplish this, we estimate random-effects linear panel regressions with staffing as the dependent variable and rurality as the key explanatory variable. The Certification and Survey Provider Enhanced Reports (CASPER) provide data on NH characteristics and staffing. CASPER is a national database collected as part of an annual, required re-certification process for all NHs that receive Medicare or Medicaid funding. Our sample includes all CASPER surveys between 2009 and the first quarter for 2015 in free-standing NHs with at least 20 beds. We merged the CASPER data with rural–urban commuting area (RUCA) codes from the WWAMI Rural Health Research Center to measure rurality. The final analytic sample included 85,826 survey observations from 14,691 unique NHs. Variables CASPER data identify two levels of staff: CMS-Qualified Social Workers (QSWs) who earned at least a bachelor’s degree in social work or a human services field with at least 1 year of supervised social work experience in a health care setting working with the elderly; and paraprofessionals who have not completed a bachelor’s degree (Items F61 and F62 on CMS-Form 671). We measured staffing levels for overall social services, and separately for QSWs and paraprofessionals by full-time equivalents (FTEs) and hours per resident day (HPRD). HPRD adjusts for the number of residents in the facility and reflects the average number of hours a staff member could spend with a resident each day. We also examine the composition of departments as measured by the proportion staffed by QSWs. Utilizing WWAMI’s Categorization A of RUCA codes, we group NHs as urban or one of the three levels of rurality, including micropolitan city, small rural town, and isolated small rural town. From urban to the most rural, these categories make up 66.7%, 13.7%, 11.0%, and 8.6% of the sample, respectively. The regressions also control for other contextual and organizational factors that may affect staffing, such as facility characteristics, payer-mix, resident case-mix, availability of other staffing resources, and state and year fixed effects. Descriptive information about the covariates are presented in Table 1. Bivariate comparisons between rural and urban characteristics found statistically significant differences for all study variables (results not reported). Table 1. Descriptive Statistics All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 Notes: The descriptive statistics table above reports the mean and standard deviation (SD; in parentheses) for the sample. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The proportion of social service department staffed by CMS-QSWs was based on 83,828 observations due to some NHs not reporting any social services staff. Overall, 2.33% or 1,999 observations do not report any social service staff. The corresponding rates of NHs not reporting any social services staff for urban, micropolitan, small rural town, and isolated small rural town are 1.86%, 3.02%, 3.30%, and 3.61%, respectively. The reference group for ownership is for-profit. The Acuindex is a facility-level measure of resident acuity summed from the ADL index score and special treatment index (Cowles, 2003). Physician services are measured in four mutually exclusive categories: medical director only (reference group); a medical team consisting of physicians and nurse practitioners (NPs) or physician assistants (PAs); NPs and PAs only; and no services. The Herfindahl–Hirschman index is a measure of market competition for NH services within the county. A few variables included in the model were not reported in Table 1, including state fixed effects, year fixed effects, and errors rates in nursing or activities staffing. ADL = activities of daily living; CCRC = continuing care retirement community; CMS = Centers for Medicare and Medicaid Services; CNA = certified nurse aide; FTE = full-time equivalent; HPRD = hours per resident day; LPN = licensed practical nurse; NH = nursing home; QSW = qualified social worker; RN = registered nurse. View Large Table 1. Descriptive Statistics All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 All NHs Urban Level of rurality Micropolitan city Small rural town Isolated small rural town M (SD) M (SD) M (SD) M (SD) M (SD) FTEs  Total social services 1.838 (1.394) 2.061 (1.508) 1.607 (1.059) 1.333 (1.011) 1.123 (0.747)  CMS-QSWs 1.207 (1.159) 1.401 (1.241) 0.976 (0.888) 0.773 (0.895) 0.624 (0.653)  Paraprofessionals 0.631 (0.956) 0.660 (1.040) 0.631 (0.849) 0.560 (0.694) 0.499 (0.671) HPRD  Total social services 0.117 (0.203) 0.122 (0.236) 0.109 (0.141) 0.100 (0.068) 0.109 (0.092)  CMS-QSWs 0.076 (0.170) 0.084 (0.195) 0.065 (0.123) 0.056 (0.068) 0.059 (0.094)  Paraprofessionals 0.041 (0.106) 0.038 (0.121) 0.044 (0.081) 0.044 (0.057) 0.050 (0.065) % of department staffed by CMS-QSWs 67.325 (39.608) 70.833 (36.830) 62.604 (41.836) 59.506 (44.121) 57.238 (46.444) Control variables  Ownership: government 0.045 (0.208) 0.028 (0.165) 0.062 (0.241) 0.074 (0.261) 0.117 (0.321)  Ownership: not-for-profit 0.222 (0.416) 0.224 (0.417) 0.205 (0.403) 0.186 (0.389) 0.282 (0.450)  121+ bed facility 0.320 (0.467) 0.394 (0.489) 0.263 (0.440) 0.146 (0.353) 0.066 (0.248)  No. of beds (/10) 11.226 (6.064) 12.256 (6.572) 10.432 (4.488) 9.172 (3.724) 7.137 (3.384)  No. of beds (/10) × 121+ bed facility 5.612 (8.941) 7.066 (9.722) 4.142 (7.300) 2.240 (5.622) 0.993 (3.950)  Multi-facility chain 0.562 (0.496) 0.561 (0.496) 0.593 (0.491) 0.586 (0.493) 0.492 (0.500)  Part of CCRC 0.099 (0.298) 0.116 (0.320) 0.093 (0.290) 0.048 (0.214) 0.038 (0.191)  Occupancy rate 81.392 (16.478) 82.875 (16.100) 79.438 (16.689) 76.887 (17.338) 78.771 (16.225)  % Medicaid residents 60.874 (22.089) 59.276 (24.141) 63.728 (17.277) 65.097 (16.222) 63.320 (16.739)  % Medicare residents 14.559 (12.916) 16.031 (14.328) 13.612 (9.513) 11.344 (8.389) 8.760 (6.967)  Acuindex 10.192 (1.481) 10.359 (1.526) 10.056 (1.304) 9.871 (1.287) 9.524 (1.337)  % Dementia 47.606 (17.796) 46.803 (18.635) 47.736 (15.747) 49.453 (15.753) 51.267 (15.937)  % Psychiatric illness 28.057 (19.075) 26.996 (19.191) 29.873 (18.528) 30.459 (18.324) 30.327 (19.273)  % Depression 47.587 (22.469) 45.414 (22.177) 50.134 (22.402) 52.168 (22.286) 54.517 (22.350)  % Developmental disability 2.419 (5.386) 2.067 (5.487) 2.976 (5.743) 3.222 (4.423) 3.241 (4.821)  Dementia special care unit 0.166 (0.372) 0.159 (0.365) 0.189 (0.391) 0.194 (0.395) 0.147 (0.354)  Other special care unit 0.053 (0.224) 0.063 (0.243) 0.037 (0.189) 0.029 (0.168) 0.027 (0.161)  RN staffing (HPRD) 0.427 (0.328) 0.463 (0.353) 0.372 (0.279) 0.338 (0.242) 0.353 (0.232)  LPN staffing (HPRD) 0.805 (0.371) 0.825 (0.386) 0.808 (0.350) 0.758 (0.311) 0.711 (0.325)  CNA staffing (HPRD) 2.252 (0.729) 2.287 (0.745) 2.204 (0.706) 2.175 (0.673) 2.162 (0.693)  Physician services: medical team 0.586 (0.493) 0.649 (0.477) 0.497 (0.500) 0.451 (0.498) 0.410 (0.492)  Physician services: NP and PAs only 0.018 (0.132) 0.015 (0.121) 0.025 (0.156) 0.018 (0.132) 0.028 (0.166)  Physician services: none 0.111 (0.314) 0.077 (0.266) 0.141 (0.348) 0.187 (0.390) 0.230 (0.421)  Any mental health staff 0.575 (0.494) 0.665 (0.472) 0.454 (0.498) 0.385 (0.487) 0.316 (0.465)  Activities staff (HPRD) 0.187 (0.136) 0.190 (0.144) 0.177 (0.122) 0.176 (0.110) 0.201 (0.122)  Herfindahl–Hirschman index 0.190 (0.228) 0.094 (0.128) 0.279 (0.183) 0.423 (0.261) 0.498 (0.301) Number of observations 85,826 57,236 11,786 9,413 7,391 Notes: The descriptive statistics table above reports the mean and standard deviation (SD; in parentheses) for the sample. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The proportion of social service department staffed by CMS-QSWs was based on 83,828 observations due to some NHs not reporting any social services staff. Overall, 2.33% or 1,999 observations do not report any social service staff. The corresponding rates of NHs not reporting any social services staff for urban, micropolitan, small rural town, and isolated small rural town are 1.86%, 3.02%, 3.30%, and 3.61%, respectively. The reference group for ownership is for-profit. The Acuindex is a facility-level measure of resident acuity summed from the ADL index score and special treatment index (Cowles, 2003). Physician services are measured in four mutually exclusive categories: medical director only (reference group); a medical team consisting of physicians and nurse practitioners (NPs) or physician assistants (PAs); NPs and PAs only; and no services. The Herfindahl–Hirschman index is a measure of market competition for NH services within the county. A few variables included in the model were not reported in Table 1, including state fixed effects, year fixed effects, and errors rates in nursing or activities staffing. ADL = activities of daily living; CCRC = continuing care retirement community; CMS = Centers for Medicare and Medicaid Services; CNA = certified nurse aide; FTE = full-time equivalent; HPRD = hours per resident day; LPN = licensed practical nurse; NH = nursing home; QSW = qualified social worker; RN = registered nurse. View Large Results The average NH employed 1.8 FTEs, of which 1.2 FTEs were QSWs and 0.6 were paraprofessionals, making approximately two thirds of social service departments staffed by QSWs. A total of 2.3% survey observations do not report any social service staffing. How these staffing levels change with NH size and rurality is illustrated in Figure 1. There are two notable trends. First, the overall staffing levels were positively correlated with NH size. Second, rural NHs were more likely to hire fewer FTEs regardless of NH size, although some variation exists around the smallest NHs and those close to 150 beds. Figure 1. View largeDownload slide Overall social service staffing levels (FTEs) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. FTE = full-time equivalent; NH = nursing home. Figure 1. View largeDownload slide Overall social service staffing levels (FTEs) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. FTE = full-time equivalent; NH = nursing home. Since rural NHs also tend to be smaller (Table 2), we also measured staffing levels in HPRD. Figure 2 shows how staffing levels decreased as NHs increased in size and that rural NHs had lower staffing levels. The smallest NHs had between 0.15 and 0.25 HPRD. However, as the size of the facility increased, staffing levels decreased until the NH had at least 100 beds. For NHs with greater than 100 beds, staffing levels ranged from 0.08 to 0.12 HPRD. Table 2. Number of Beds by Level of Rurality Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Note: The table reports the distribution of the number of beds for the overall sample and each level of rurality (n = 85,826). NH = nursing home; SD = standard deviation. View Large Table 2. Number of Beds by Level of Rurality Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Mean (SD) Median Range Interquartile range (25th–75th percentile) All NHs 112.274 (60.641) 103 20–908 70–132 Urban 122.562 (65.714) 118 20–908 81–148 Micropolitan city 104.321 (44.880) 100 20–815 75–122 Small rural town 91.718 (37.236) 90 20–499 63–114 Isolated small rural town 71.368 (33.838) 61 20–549 49–90 Note: The table reports the distribution of the number of beds for the overall sample and each level of rurality (n = 85,826). NH = nursing home; SD = standard deviation. View Large Figure 2. View largeDownload slide Overall social service staffing levels (HPRD) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. HPRD = hours per resident day; NH = nursing home. Figure 2. View largeDownload slide Overall social service staffing levels (HPRD) by rurality. The analysis for the figure above used the entire sample. The figure is limited to NHs with under 150 beds due to the small percentage of large facilities in rural areas. HPRD = hours per resident day; NH = nursing home. To account for other confounding variables, Table 3 reports the regression results for all NHs with at least 20 beds. The number of FTEs decreased with rurality, as rural NHs had 0.08 fewer FTEs compared with urban NHs. Similarly, QSW staffing levels in rural areas were lower, ranging from 0.11 to 0.13 fewer FTEs. Not only was the staffing of social services lower in rural areas, but paraprofessionals may have been hired in place of QSWs, as rural NHs used more paraprofessionals. Examining the proportion of QSWs within NHs confirms this result. For the average NH in the sample, QSWs comprised nearly two thirds of social service departments. Compared with urban NHs, social services in micropolitan NHs consisted of fewer QSWs (4.5% points fewer), though the effect was largest among NHs in small isolated rural towns (7.5% points fewer). Table 3. Influence of Rurality on Social Service Staffing Levels and Composition Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Note: The table above reports the coefficient estimates and standard errors (SEs; in parentheses) of random-effects linear panel regression models for staffing levels and composition of social services for all NHs. The model also controls for NH characteristics, resident case-mix, other types of staffing, state fixed effects, and year fixed effects (see Table 1). The reference group is urban NHs. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The final sample is 85,826 observations from 14,658 NHs, except in the case of the proportion of social service department staffed by CMS-QSWs which has 83,828 observations. The difference in the number of observations is because some NHs reported no staffing in social services. SEs are adjusted for clustering at the facility level. CMS = Centers for Medicare and Medicaid Services; FTE = full-time equivalent; HPRD = hours per resident day; NH = nursing home; QSW = qualified social worker; SD = standard deviation. *p < .1. ** p < .05. *** p < .01. View Large Table 3. Influence of Rurality on Social Service Staffing Levels and Composition Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Staffing levels Staffing levels Proportion of department staffed by CMS-QSW FTE HPRD Total social services CMS-QSW Paraprofessionals Total social services CMS-QSW Paraprofessionals Mean of dependent variable 1.838 1.207 0.631 0.117 0.076 0.041 67.325% Micropolitan city −0.081*** (0.019) −0.119*** (0.016) 0.041** (0.019) −0.009*** (0.002) −0.012*** (0.002) 0.002 (0.002) −4.501*** (0.850) Small rural town −0.085*** (0.025) −0.111*** (0.022) 0.029 (0.020) −0.017*** (0.003) −0.018*** (0.002) −0.001 (0.002) −5.457*** (1.070) Isolated small rural town −0.077*** (0.025) −0.138** (0.022) 0.064*** (0.024) −0.020*** (0.004) −0.024*** (0.004) 0.003 (0.002) −7.455** (1.329) Note: The table above reports the coefficient estimates and standard errors (SEs; in parentheses) of random-effects linear panel regression models for staffing levels and composition of social services for all NHs. The model also controls for NH characteristics, resident case-mix, other types of staffing, state fixed effects, and year fixed effects (see Table 1). The reference group is urban NHs. Any observation with a total social service FTE that is 4 SDs above the sample mean was identified as an outlier and excluded from the study. The final sample is 85,826 observations from 14,658 NHs, except in the case of the proportion of social service department staffed by CMS-QSWs which has 83,828 observations. The difference in the number of observations is because some NHs reported no staffing in social services. SEs are adjusted for clustering at the facility level. CMS = Centers for Medicare and Medicaid Services; FTE = full-time equivalent; HPRD = hours per resident day; NH = nursing home; QSW = qualified social worker; SD = standard deviation. *p < .1. ** p < .05. *** p < .01. View Large When adjusting for the number of residents, both overall and social worker staffing levels were lower the more rural a facility. Compared with urban NHs, micropolitan NHs had staffing levels that were 0.009 and 0.012 HPRD lower for overall and social worker staffing, respectively. The effects for the most isolated rural NHs were lower by nearly double, at 0.020 and 0.024 HPRD. Interestingly, the staffing levels of paraprofessionals in terms of HPRD were not statistically different by rurality. Discussion This study highlights the influence of rurality on the staffing levels and composition of NH social service departments. In our national sample, more FTEs were hired as facility size increased; however, larger facilities staff fewer HPRD. Furthermore, rural facilities were disadvantaged by having lower staffing levels than urban NHs, which produces higher workload in rural areas. To illustrate, overall staffing levels adjusted for the size of the facility were 8% to 22% lower in rural areas compared with urban NHs. This implies that staff in rural areas were expected to do more work if they were to maintain the same level of service quality as urban areas. Our findings also indicate that QSWs make up a smaller proportion of social service departments in rural areas, which may encourage the substitution of paraprofessionals for professionally trained and experienced staff. Very low staffing levels in rural areas are a concern, as workers may have less time to respond to the psychosocial needs of residents or alter the kinds of care provided. Though more research is needed on social service staffing, quality and resident outcomes, the existing literature suggests that better psychosocial care is associated with higher qualifications of social service staff (Simons, 2006; Vongxaiburana, Thomas, Frahm, & Hyer, 2011; Zhang, Gammonley, Paek, & Frahm, 2008). Staffing is strongly influenced by federal regulations that require one full-time CMS-QSW for large facilities, and do not currently address staffing levels relative to caseload. Guidelines based on a ratio of residents per social service staff, similar to minimum nursing staff ratios (Bowblis, 2011; Mueller et al., 2006; Park & Stearns, 2009), will support consistency in the structure of social services and take steps toward requiring a manageable caseload. Bern-Klug, Kramer, Sharr, and Cruz (2010) found that NH directors of social services recommend a caseload of 60 long-term stay residents or less than 20 residents in skilled care per FTE worker. Relative to this standard, social services are seriously understaffed, especially in rural areas. A significant challenge with minimum staffing levels is the limited availability of skilled professionals in the rural workforce and cost of hiring additional staff. Facing these challenges, rural NHs should increase professional development opportunities for existing staff, with attention to opportunities for mentoring, training, and higher education. Limitations Several limitations are noted. First, we examined two types of staff: QSWs and paraprofessionals. While CASPER defines a QSW as anyone with a college degree in related fields, at minimum, the NASW (2003, 2016) requires a bachelor’s degree in social work. Therefore, our study should be interpreted with CMS definitions and not the standards of the NASW. Second, CASPER reports staffing levels for the 14-day period before the inspection, and this may not reflect average staffing throughout the year. Third, there are a limited number of isolated rural NHs with over 150 beds. While our analyses used a sample of NHs with at least 20 beds, sensitivity analyses that limited the regressions to under 150 beds were found to be robust. Finally, we used HPRD to adjust for the number of residents. Analyses utilizing FTEs per 100 resulted in similar conclusions, but found greater use of paraprofessionals in rural areas. Conclusion Insufficient regulation and staffing continue to challenge how social services are provided in NHs (Simons et al., 2012). National trends indicate considerable staffing variation in social service departments by size of facility and degree of rurality. Modifications to policy at the state and federal level are needed to formalize the structure of NH social services, ensure realistic caseloads, and define specific responsibilities according to the educational preparation and experience of staff to provide high-quality social services. Lastly, awareness of the unique challenges of rural NHs should be part of ongoing advocacy and research efforts to support the role of social workers and paraprofessionals in delivering care in NHs. References Bern-Klug M . ( 2008 ). State variations in nursing home social worker qualifications . Journal of Gerontological Social Work , 51 , 379 – 409 . doi: 10.1080/01634370802039734 Google Scholar CrossRef Search ADS PubMed Bern-Klug M. Kramer K. W . ( 2013 ). Core functions of nursing home social service departments in the United States . Journal of the American Medical Directors Association , 14 , e1 – e7 . Google Scholar CrossRef Search ADS Bern-Klug M. Kramer K. W. Chan G. Kane R. Dorfman L. T. Saunders J. B . ( 2009 ). Characteristics of nursing home social services directors: How common is a degree in social work ? Journal of the American Medical Directors Association , 10 , 36 – 44 . doi: 10.1016/j.jamda.2008.06.011 Google Scholar CrossRef Search ADS PubMed Bern-Klug M. Kramer K. W. Sharr P. Cruz I . ( 2010 ). Nursing home social services directors’ opinions about the number of residents they can serve . Journal of Aging & Social Policy , 22 , 33 – 52 . doi: 10.1080/08959420903396426 Google Scholar CrossRef Search ADS PubMed Bowblis J. R . ( 2011 ). Staffing ratios and quality: An analysis of minimum direct care staffing requirements for nursing homes . Health Services Research , 46 , 1495 – 1516 . doi: 10.1111/j.1475-6773.2011.01274.x Google Scholar CrossRef Search ADS PubMed Bowblis J. R. Meng H. Hyer K . ( 2013 ). The urban-rural disparity in nursing home quality indicators: The case of facility-acquired contractures . Health Services Research , 48 , 47 – 69 . doi: 10.1111/j.1475-6773.2012.01431 Google Scholar CrossRef Search ADS PubMed Bowen S. E. Zimmerman S . ( 2008 ). Understanding and improving psychosocial services in long-term care . Health Care Financing Review , 30 , 1 – 4 . Google Scholar PubMed Cowles C. M . ( 2003 ). 2002 Home Statistical Yearbook . Montgomery Village, MD : Cowles Research Group . Harris-Kojetin L. Sengupta M. Park-Lee E. Valverde R . ( 2013 ). Long-term care services in the United States: 2013 overview. National Center for Health Statistics . Vital Health Statistics , 3 , 1 – 105 . Levinson D . ( 2013 ). Skilled nursing facilities often fail to meet care planning and discharge planning requirements . Washington, DC : Department of Health and Human Services, Office of the Inspector General . Lutfiyya M. N. Gessert C. E. Lipsky M. S . ( 2013 ). Nursing home quality: A comparative analysis using CMS Nursing Home Compare data to examine differences between rural and nonrural facilities . Journal of the American Medical Directors Association , 14 , 593 – 598 . doi: 10.1016/j.jamda.2013.02.017 Google Scholar CrossRef Search ADS PubMed Mueller C. Arling G. Kane R. Bershadsky J. Holland D. Joy A . ( 2006 ). Nursing home staffing standards: Their relationship to nurse staffing levels . The Gerontologist , 46 , 74 – 80 . doi: 10.1093/geront/46.1.74 Google Scholar CrossRef Search ADS PubMed National Association of Social Workers . ( 2003 ). Standards for social work services in long-term care facilities . Washington, DC : NASW Press . National Association of Social Workers . ( 2016 ). NASW comments to the Senate Committee on Finance Bipartisan Chronic Care Working Group . Retrieved from http://www.socialworkers.org/advocacy/letters/2016/Senate_Finance_chronic_care.pdf Park J. Stearns S. C . ( 2009 ). Effects of state minimum staffing standards on nursing home staffing and quality of care . Health Services Research , 44 , 56 – 78 . doi: 10.1111/j.1475-6773.2008. 00906.x Google Scholar CrossRef Search ADS PubMed Rehnquist J . ( 2003 ). Psychosocial services in skilled nursing facilities . Washington, DC : Department of Health and Human Services, Office of the Inspector General . Roberts A. R. Bowblis J. R . ( 2016 ). Who hires social workers? Structural and contextual determinants of social service staffing in nursing homes . Health & Social Work . Advance online publication . doi: 10.1093/hsw/hlw058 Simons K. V . ( 2006 ). Organizational characteristics influencing nursing home social service directors’ qualifications: A national study . Health & Social Work , 31 , 266 – 274 . doi: 10.1093/hsw/31.4.266 Google Scholar CrossRef Search ADS PubMed Simons K. V. Bern-Klug M. An S . ( 2012 ). Envisioning quality psychosocial care in nursing homes: The role of social work . Journal of the American Medical Directors Association , 13 , 800 – 805 . doi: 10.1016/j.jamda.2012.07.016 Google Scholar CrossRef Search ADS PubMed State Operations Manual . ( 2016 ). Appendix PP – Guidance to Surveyors for Long Term Care Facilities (Rev. 157, 06-10-16) . Retrieved from www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/som107ap_pp_guidelines_ltcf.pdf Vongxaiburana E. Thomas K. S. Frahm K. A. Hyer K . ( 2011 ). The social worker in interdisciplinary care planning . Clinical Gerontologist , 34 , 367 – 378 . doi: 10.1080/07317115.2011.588540 Google Scholar CrossRef Search ADS Zhang N. J. Gammonley D. Paek S. C. Frahm K . ( 2008 ). Facility service environments, staffing, and psychosocial care in nursing homes . Health Care Financing Review , 30 , 5 – 17 . Google Scholar PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The GerontologistOxford University Press

Published: Jan 10, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off