This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

The Impact of Rural Residence on Multiple Hospitalizations in Nursing Facility Residents

^a Institute for Health Policy, Edmund S. Muskie School of Public Service, University of Southern Maine, Portland

Correspondence: Andrew F. Coburn, PhD, Institute for Health Policy, Edmund S. Muskie School of Public Service, University of Southern Maine, P.O. Box 9300, Portland, ME 04101. E-mail: Andyc{at}usm.maine.edu.

Decision Editor: Laurence G. Branch, PhD

	Abstract

TOP Abstract Methods Results Discussion Appendix ENDIX References

 
Purpose: This study explored issues surrounding hospitalization rates among rural and urban nursing facility (NF) residents. Design and Methods: Data from the Minimum Data Set (MDS+) collected by states participating in the national Multi-State Case Mix Demonstration were used to assess whether rural NF residents experience higher rates of hospitalization compared with urban residents and to understand the extent that resident, facility, and market/area characteristics contribute to these differences. Results: Rural NF residents were more likely than urban residents to have multiple hospitalizations. Further analysis demonstrated that the effect of rural residence on the probability of multiple hospitalizations is greater among newly admitted rural residents than among rural residents not classified as new admissions. In addition to rural residence, other factors associated with an increased likelihood of multiple hospitalizations included state of residence, diagnosis of congestive heart failure, and no discharge planned at the time of NF admission. Implications: The findings of this study have important implications for both clinical care and health policy related to the financing and administration of NFs.

Key Words: Health services for aged persons • Rural health • Patient transfer/admission • Quality of health care • Health services research

The problem of hospitalizations, particularly repeated hospitalizations, of nursing facility (NF) residents has important implications for both the cost and the quality of care. The experience of hospitalization, as well as the transition from one setting to another, puts residents at increased risk of iatrogenesis and significant relocation stress (Gillick and Steel 1983; Mor et al. 1997). These effects often result in the experience of cascading dependency, a progressive downward spiral of decreasing function and increasing cognitive impairment (Creditor 1993). Repeated transitions between hospital and NF have been described as a "Ping-Pong" effect and have been shown to have particularly negative effects on the health status of residents (Brandeis, Ooi, Hossain, Morris, and Lipsitz 1994; Lewis, Cretin, and Kane 1985; Tresch, Simpson, and Burton 1985). Although it is known that on a national basis almost 40% of NF discharges are to hospitals, geographic variations in hospitalization rates have not been reported in the literature (Saliba et al. 2000; Sekscenski 1990). Thus, this study was developed to investigate this issue.

There is reason to believe that the problem of repeated transfers to and from hospitals may be more prevalent among rural NF residents. Higher rates of nursing home use, barriers to the provision of certain types of medical and nursing care in rural NFs, greater supplies of hospital beds, and lower physician supplies in rural areas are all factors that influence hospital utilization in rural areas. In a prior study of NF discharge rates for rural and urban residents hospitalized with a hip fracture in Maine, we found that rural residents experiencing an initial NF admission following hip surgery had higher rates of rehospitalization followed by NF readmission than urban NF residents (Coburn, Bolda, Keith, Dushuttle, and Schultz 1997).

A variety of factors may be related to NF-to-hospital transfers (Castle and Mor 1996), including residents' specific health problems (Murtaugh and Freiman 1995), the capacity of NFs to care for seriously ill residents during acute episodes (Stearns, Kovar, Hayes, and Koch 1996), and policy and financial incentives for NFs to shift the costs of caring for seriously ill residents to hospitals (Rubenstein, Ouslander, and Wieland 1988; Stearns et al. 1996). With regard to the latter, the NF-to-hospital transfer problem may be symptomatic of fragmented or misaligned incentives in state nursing home and federal hospital reimbursement policies (Freiman and Murtaugh 1995; Ouslander, Weinberg, and Phillips 2000; Stearns et al. 1996), particularly as they pertain to persons who are dually eligible for Medicare and Medicaid. Evidence of such policy conflict is suggested by previous findings that the rate of NF-to-hospital-to-NF transfers may be sensitive to both nursing home reimbursement rates and hospital bed supply (Stearns et al. 1996).

Barker and colleagues 1994 have recommended policy and other strategies to reduce unnecessary hospitalizations, including modifying NF staffing, expanding the use of geriatric nurse practitioners, and increasing payments to physicians providing services in NFs. The feasibility and implications of such interventions, and the effects of reducing hospital use in rural areas, however, have not been addressed in the literature. In rural communities, where the supply of physicians and specialty health care providers is lower than in urban areas, efforts to increase NF capacity are more difficult to achieve. Moreover, the financial implications of shifting care from hospital to NFs in rural areas, with a potentially significant negative effect on rural hospital revenues, are not well understood.

Given these problems and the limited information available to address them, this study was undertaken to understand more about the problem of hospital-to-NF admissions in rural NF residents. The study addressed two questions:

1. Do rural NF residents experience higher rates of hospitalization than urban residents?
   
2. If they do, to what extent do resident, facility, and market/area characteristics explain the difference in hospitalization rates between rural and urban residents?
   

Differences in hospitalization rates have important potential implications for the quality and outcome of care for older rural residents. Understanding whether and why these rates differ is also important for gauging the potential impact of payment and other policy changes on the quality of care in rural NFs.

	Methods

TOP Abstract Methods Results Discussion Appendix ENDIX References

 
Data Sources
This study used resident assessment data from the Minimum Data Set (MDS+) collected from January 1, 1994, though December 31, 1995. The MDS+ assessments included all residents and NFs in Maine, Mississippi, and South Dakota and all skilled NFs in New York. The MDS+ is part of the National Nursing Home Resident Assessment Instrument, the use of which is congressionally mandated (Morris et al. 1990).

Assessments are completed at admission and then quarterly or whenever there is a significant change in status. The exception is New York, where subsequent assessments are required only on an annual basis. The MDS+ includes information on demographics, functional and medical status, diagnoses, and special treatments and procedures. Although the MDS+ was designed for regulatory and care-planning purposes, its reliability and validity for research have been well established (Hawes et al. 1995, Hawes et al. 1997; Sgadari et al. 1997). Admission data in the MDS+ include date of admission, location from where the resident was admitted (home, another NF, acute care hospital, or other), and previous living arrangements (alone, with others, or in a facility).

Facility data were obtained directly from the states and linked to the assessment data using the state facility identifier. County-level data for demographic and health care provider supply variables were obtained from the Bureau of Health Professions Area Resource File for 1996 and were linked to the facility file for each state using county codes.

Study Sample and File Construction
The 1994–1995 multistate MDS+ data contained assessments for 195,425 nursing home residents in the four states: Mississippi (n = 30,714), South Dakota (n = 13,977), Maine (n = 21,781), and New York (n = 128,953). Given the population size and degree of urbanization of New York relative to the other three states, we excluded residents in facilities in New York City, resulting in a final sample of 43,107. Even so, the New York sample still had only 13% of its residents in rural (nonmetropolitan) facilities compared with more than 50% of residents in the other states.

From this database, we derived an admissions cohort, which included all residents with either a first or a second admission to a NF in 1994. The final sample for the study included 35,535 residents: 11,892 in Mississippi, 4,942 in South Dakota, 7,155 in Maine, and 11,546 in New York.

Separate resident-level analytic files were constructed for each of the four states, containing information on nursing home admissions and discharges as well as baseline data from the first 1994 assessment on resident characteristics and linked facility, and county characteristics. Analyses were conducted using both the individual state files and the pooled data file from the four states.

Study Variables
Dependent Variables
The outcome of interest in this study was hospital admission of nursing home residents. We could have measured this either by counting discharges from a nursing home to a hospital or by counting admissions from a hospital to a nursing home. We chose the latter approach because the MDS+ data on where residents came from were generally more complete and reliable than the data on where they went to after leaving the NF. Furthermore, we were not simply concerned with how often nursing home residents are sent to a hospital, but also with how often patients have hospital stays between multiple nursing home readmissions.

In our analyses, we used two dichotomous resident-level variables, one indicating whether the resident had any NF readmissions from a hospital and the other indicating whether the resident had two or more such readmissions. Any admission to a NF after the initial admission was considered a readmission. To create this variable, we constructed a resident-level file of nursing home stays for each state. For each stay, this file contained admission date, location from which the resident was admitted, discharge date, and type of discharge. For those residents in a nursing home at the beginning of the year, the first admission in the file was before the beginning of 1994, but all subsequent admissions in the file occurred in 1994 or 1995. Duplicate and overlapping stays were eliminated, and the number of nursing home stays and number of readmissions from a hospital were calculated for each resident.

Independent Variables
Independent variables included geographic (urban–rural) indicators, a wide range of individual resident characteristics, and a more limited number of facility and county characteristics (Table 1 ). The geographic indicators were defined at the county level to ensure comparability across states. For each resident, the primary geographic variable indicated whether the county of residence was within a metropolitan county as defined by the U.S. Office of Management and Budget. Other independent variables used in the analysis are described in Table 1 .

	Results

TOP Abstract Methods Results Discussion Appendix ENDIX References

 
Urban–Rural Differences in Resident Characteristics
Analyses of resident characteristics from the four-state pooled data (Table 2 ) indicated significant differences among rural and urban NF residents in this sample. On average, the case-mix index for rural residents was significantly lower than for urban residents (see Appendix, Note 1). This difference appears to be largely accounted for by the significantly lower percentage of rural compared with urban NF residents in the rehabilitation case-mix group (13.8% vs 33.2%). The percentage of rural residents was also significantly lower in the not assigned case-mix category. In contrast, a significantly higher percentage of rural residents were classified in the extensive care, special care, clinically complex, cognitive, and physical case-mix groups.

Hospitalization Rates
Bivariate analyses indicated that rural residents were more likely to experience multiple NF readmissions from a hospital than urban residents (Table 3 ). This finding was the same in all four states where residents in rural facilities were more likely to have multiple NF stays, to have at least one readmission from a hospital, and to have multiple readmissions from a hospital. Rural residents in all four states had higher mean numbers of NF stays.

Other Predictors of Multiple Hospitalizations
State of residence was by far the single most important variable in the models in predicting multiple hospitalizations. In comparison with New York, Maine, Mississippi, and South Dakota each had significantly higher rates of multiple hospitalizations (Table 4 ). Even after adjusting for other key variables in the model, the ORs for multiple hospitalization in the separate states were highly variable, ranging from 6.92 in Mississippi (C.I. = 4.78–10.02) to 2.59 in Maine (C.I. = 1.80–3.70). NF residents with congestive heart failure, and residents for whom discharge was not planned at the time of NF admission, had a significantly greater risk of multiple readmissions from the hospital. Older age (80 or older) was associated with a lower likelihood of multiple hospitalizations. With the exception of residents classified in the extensive, behavior, and not classified groups, there was a significant association between a resident's case-mix classification and the probability of multiple hospitalization. Specifically, residents classified in the rehabilitation, special care, and clinically complex case-mix groups all had a significantly higher likelihood of multiple hospitalizations, and residents in the cognitive group had a significantly lower likelihood of such hospitalizations.

	Discussion

TOP Abstract Methods Results Discussion Appendix ENDIX References

 
The findings of this study indicate that rural NF residents are more likely to have multiple readmissions to the NF from a hospital. This association was evident in all four states included in this study and persisted even after controlling for other factors (state of residence and resident, facility, and market-level characteristics). Further multivariate analysis indicated that the effect of rural residence on the probability of multiple NF readmissions is greater among newly admitted rural residents than among rural residents not classified as new admissions. In addition to rural residence, other factors associated with an increased likelihood of multiple NF readmissions from a hospital included state of residence, diagnosis of congestive heart failure, and no discharge planned at the time of NF admission. Residents classified in the rehabilitation, special care, and clinically complex case-mix groups all had a significantly greater likelihood of multiple NF readmissions from a hospital. Older age (80 and older) and being in the cognitive case-mix group were associated with a lower likelihood of multiple hospitalizations.

This study used data from only four states. We must be cautious, therefore, in generalizing these findings to the entire population of NF residents in the United States. Although the findings appear to be quite robust, the significant variations across the states in this study also give reason for pause in attempting to generalize the results. This variation is not unexpected given the known differences in the characteristics of the states, their Medicaid and long-term care eligibility and payment policies, and the nursing home markets across the states.

The literature leaves little doubt that multiple hospitalizations and relocation of nursing home residents pose significant health risks to them (Castle 2001). The findings of this study, therefore, raise potentially important quality-of-care concerns. As Castle and Mor 1996 have noted, resolving the problem of NF-to-hospital transfers is complex and poses significant conceptual, methodological, and policy challenges. In addition to the problem of disentangling the many factors that contribute to this phenomenon, implementation of facility-level or policy remedies may be complicated by staffing shortages, the need for staff education and training, and/or the need for greater physician involvement and support for medically complex NF residents. Changing, sometimes conflicting, Medicare and Medicaid payment and regulatory policies and incentives may also make solving this problem difficult.

The findings from this and other research suggest a number of key factors that may influence the decision to hospitalize a nursing home resident, including the clinical capacity of the facility, the overall acuity of the facility's residents, and the individual characteristics of residents. In their study of NF characteristics associated with the hospitalization of nursing home residents, Intrator, Castle, and Mor 1999 demonstrated that the presence of additional medical resources, either physicians or physician extenders, significantly reduced the risk of hospitalization. This suggests that a strategy of enhancing the availability and use of physician and other medical services to reduce hospitalizations might be appropriate (Garrard et al. 1990; Kayser-Jones, Wiener, and Barbaccia 1989; Mor et al. 1997). Implementing such an approach in rural nursing homes, however, may be difficult. There is a shortage of qualified health personnel, such as geriatric nurse practitioners, in many rural areas. Likewise, the limited availability of physicians in rural areas, their heavier workloads, and the limited reimbursement for care provided in NFs make it difficult in rural areas to obtain greater physician attention to NF residents.

The clinical and other characteristics of NF residents are likely to continue to change as Medicare and state Medicaid programs implement new case-mix–based payment systems for NF care that encourage facilities to admit patients with heavier care needs. State efforts to restrict eligibility for nursing home care to those with the greatest medical need and to create residential care alternatives, such as assisted living and home- and community-based service options for patients who might otherwise have been admitted to a NF, will further contribute to these trends. For all of these reasons, it is critical that we better understand the challenges of, and potential strategies for, caring for medically complex residents in rural NFs.

	Acknowledgments

 
This study was funded by Grant CSUR00003-04 from the Federal Office of Rural Health Policy, Health Resources and Services Administration, Department of Health and Human Services. The conclusions and opinions expressed in this paper are ours, and no endorsement by the University of Southern Maine or the funding source is intended or should be inferred.

We acknowledge the contribution of W. Douglas Thompson, PhD, who provided biostatistical support. Leslie H. Nicoll, PhD, MBA, RN, provided editorial assistance in the preparation of the manuscript.

Received for publication January 23, 2002. Accepted for publication April 16, 2002.

	Appendix ENDIX

TOP Abstract Methods Results Discussion Appendix ENDIX References

 
Note
The case-mix groups and case-mix index we used are taken from the Resource Utilization Groups system used by Health Care Financing Administration (now the Centers for Medicare and Medicaid Services) for determining reimbursement in nursing homes. This system uses MDS data to assign residents to case-mix groups on the basis of their expected level of resource utilization as determined by their need for services and staff time measurements. The case-mix index is a measure of the relative resource utilization expected for each group. For more information on the case-mix groups, see Health and Human Services Inspector General 2002.

	References

TOP Abstract Methods Results Discussion Appendix ENDIX References

 

• Barker W. H., Zimmer J. G., Hall W. J., Ruff B. C., Freundlich C. B., Eggert G. M., 1994. Rates, patterns, causes, and costs of hospitalization of nursing home residents: A population-based study. American Journal of Public Health 84:1615-1620. [Abstract/Free Full Text]
• Brandeis G. H., Ooi W. L., Hossain M., Morris J. N., Lipsitz L. A., 1994. A longitudinal study of risk factors associated with the formation of pressure ulcers in nursing homes. Journal of the American Geriatrics Society 42:388-393. [Medline]
• Castle N. G., 2001. Relocation of the elderly. Medical Care Research and Review 58:291-333. [Abstract/Free Full Text]
• Castle N. G., Mor V., 1996. Hospitalization of nursing home residents: A review of the literature, 1980–1995. Medical Care Research and Review 53:123-148. [Abstract/Free Full Text]
• Coburn, A., Bolda, E., Keith, R., Dushuttle, P., & Schultz, A. (1997). Differences in nursing home discharge rates for urban and rural nursing facility residents with hip fracture (Working Paper No. 8). Portland: Maine Rural Health Research Center, Edmund S. Muskie School of Public Service, Institute for Health Policy, University of Southern Maine.
• Creditor M. C., 1993. Hazards of hospitalization of the elderly. Annals of Internal Medicine 118:219-223. [Abstract/Free Full Text]
• Freiman M. P., Murtaugh C. M., 1995. Interactions between hospital and nursing home use. Public Health Reports 110:545-554.
• Garrard J., Kane R. L., Radosevich D. M., Skay C. L., Arnold S., Kepferle L., et al. 1990. Impact of geriatric nurse practitioners on nursing-home residents' functional status, satisfaction, and discharge outcomes. Medical Care 28:271-283. [Medline]
• Gillick M., Steel K., 1983. Referral of patients from long-term to acute-care facilities. Journal of the American Geriatrics Society 31:74-78. [Medline]
• Hawes C., Morris J. N., Phillips C. D., Fries B. E., Murphy K., Mor V., 1997. Development of the nursing home Resident Assessment Instrument in the USA. Age and Ageing 26: (Suppl. 2) 19-25.
• Hawes C., Morris J. N., Phillips C. D., Mor V., Fries B. E., Nonemaker S., 1995. Reliability estimates for the Minimum Data Set for nursing home resident assessment and care screening (MDS). The Gerontologist 35:172-178. [Abstract]
• Health and Human Services Inspector General 2002Retrieved February 4, 2002, from http://oig.hhs.gov/oei/reports/a529.pdf. Trends in the assignment of resource utilization groups by skilled nursing facilities
• Intrator O., Castle N. G., Mor V., 1999. Facility characteristics associated with hospitalization of nursing home residents: Results of a national study. Medical Care 37:228-237. [Medline]
• Kayser-Jones J. S., Wiener C. L., Barbaccia J. C., 1989. Factors contributing to the hospitalization of nursing home residents. The Gerontologist 29:502-510. [Abstract]
• Lewis M. A., Cretin S., Kane R. L., 1985. The natural history of nursing home patients. The Gerontologist 25:382-388. [Medline]
• Mor V., Intrator O., Fries B. E., Phillips C., Teno J., Hiris J., et al. 1997. Changes in hospitalization associated with introducing the Resident Assessment Instrument. Journal of the American Geriatrics Society 45:1002-1010. [Medline]
• Morris J. N., Hawes C., Fries B. E., Phillips C. D., Mor V., Katz S., et al. 1990. Designing the national Resident Assessment Instrument for nursing homes. The Gerontologist 30:293-307. [Abstract]
• Murtaugh C. M., Freiman M. P., 1995. Nursing home residents at risk of hospitalization and the characteristics of their hospital stays. The Gerontologist 35:35-43. [Abstract]
• Ouslander J. G., Weinberg A. D., Phillips V., 2000. Inappropriate hospitalization of nursing facility residents: A symptom of a sick system of care for frail older people. Journal of the American Geriatrics Society 48:230-231. [Medline]
• Rubenstein L. Z., Ouslander J. G., Wieland D., 1988. Dynamics and clinical implications of the nursing home–hospital interface. Clinics in Geriatric Medicine 4:471-491.
• Saliba D., Kington R., Buchanan J., Bell R., Wang M., Lee M., et al. 2000. Appropriateness of the decision to transfer nursing facility residents to the hospital. Journal of the American Geriatrics Society 48:154-163. [Medline]
• Sekscenski E. S., 1990. Discharges from nursing homes: 1985 National Nursing Home Survey. Vital Health Statistics 13:1-87.
• Sgadari A., Morris J. N., Fries B. E., Ljunggren G., Jonsson P. V., DuPaquier J. N., et al. 1997. Efforts to establish the reliability of the Resident Assessment Instrument. Age and Ageing 26: (Suppl. 2) 27-30.
• Stearns S. C., Kovar M. G., Hayes K., Koch G. G., 1996. Risk indicators for hospitalization during the last year of life. Health Services Research 31:49-69. [Medline]
• Tresch D. D., Simpson W. M., Jr. Burton J. R., 1985. Relationship of long-term and acute-care facilities: The problem of patient transfer and continuity of care. Journal of the American Geriatrics Society 33:819-826. [Medline]

This article has been cited by other articles:

				N. M. Bellows and H. A. Halpin MDS-Based State Medicaid Reimbursement and the ADL-Decline Quality Indicator Gerontologist, June 1, 2008; 48(3): 324 - 329. [Abstract] [Full Text] [PDF]

				A. Gruneir, S. C. Miller, O. Intrator, and V. Mor Hospitalization of Nursing Home Residents With Cognitive Impairments: The Influence of Organizational Features and State Policies Gerontologist, August 1, 2007; 47(4): 447 - 456. [Abstract] [Full Text] [PDF]

				C. D. Phillips, S. Holan, M. Sherman, M. L. Williams, and C. Hawes Rurality and Nursing Home Quality: Results from a National Sample of Nursing Home Admissions Am J Public Health, October 1, 2004; 94(10): 1717 - 1722. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation