Journal of Gerontological Nursing

The Relationship Between Nurse Staffing in Nursing Homes and Quality Indicators

Mary Ellen Dellefield, PhD, RN

Abstract

TABLE 1

STUDIES RELATING TOTAL NURSING STAFF HOURS AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY…

A common belief among both nursing home staff and lay people is that the ity of a nursing home is related to the amount and type of nursing staff available to residents. Given the interpersonal dimensions of nursing practice, functional impairments of many nursing home residents, and the labor-intennature of nursing care in this setting, it is understandable that this association is commonly assumed to exist. Because RNs function in leadership positions in nursing homes, it is important for them to be familiar with the research evidence currently available about the relationship between nursing staffing levels and quality in nursing homes. Informed RNs will be in a better position to participate in policy discussions at the state and national levels focused on the relationship between nursing home staffing and quality.

The purpose of this article is to review the published research literature that has addressed the relationship between traditional nursing staffing in nursing homes and various types of quality indicators. Because Donabedian's (1980) quality framework was frequently used as the theoretical framework for many studies, it will be explained, and the literature reviewed in this context. Implications for clinical practice, research, and health care policy will be addressed.

Table

TABLE 1STUDIES RELATING TOTAL NURSING STAFF HOURS AND NURSING HOME QUALITY

TABLE 1

STUDIES RELATING TOTAL NURSING STAFF HOURS AND NURSING HOME QUALITY

DONABEDIAN'S QUALITY FRAMEWORK

Donabedian's quality framework is a widely used conceptual framework for the study of nursing home quality. The assumptions of Donabedian's framework are that quality is multidimensional and has both technical and interpersonal aspects. Determining the presence or absence of quality is complex and requires inferences about quality be made through the examination of evidence and data (Donabedian, 1992). Each piece of evidence is called a quality indicator and is typically organized under one of four categories: structural, process, outcome, and composite types of indicators. A composite indicator is a combination of process and outcome indicators. Indicators may be either objective or perceptual measures. Because of quality's multidimensional nature, various types of quality indicators are needed to provide evidence about the quality of a nursing home. No single indicator is used as a comprehensive measure of the facility's quality (Donabedian, 1980).

The major proposition of Donabedian's quality framework is that the three primary quality indicators are causally linked. "Structure leads to process, and process leads to outcome," according to Donabedian (1992, p. 357). Structure may also directly influence outcome (Sainfort, Ramsay, Ferreia, & Mezghani, 1994).

Structure

Structure refers to relatively stable organizational characteristics, including system, provider, and consumer characteristics. In the nursing home setting, examples of system and provider dimensions of structure that have been studied include staffing levels, staffing mix, staff turnover, wages and benefits, management and leadership structure, facility size, location, ownership, availability of private rooms, volunteers, governance of the organization, age and condition of the physical plant, payer mix, case mix, accreditation, and teaching status (Davis, 1991; Institute of Medicine [IOM], 1996). Perceptual measures of structure include centralization, formalization, standardization, and span of control (Mullins, Nelson, Busciglio, & Weiner, 1988). Consumer characteristics are dimensions such as functional level, comorbidities, age, gender, and ethnicity of patients and residents of nursing homes.

Process

Process refers to what is done for and with residents or patients receiving health services. In the nursing home, process quality indicators that have been studied include assistance with activities of daily living (ADLs), staff or patient injury, infection control, various resident services, restraint use, use of urinary catheters, bladder training, cleaning and maintaining resident rooms, assessment, abuse prevention, quality assurance, use of medical care, and resident rights (IOM, 1996). A perceptual measure of process is coordination (Georgopoulos & Mann, 1962). For example, it is commonly assumed that increased interdisciplinary coordination is related to the achievement of desired clinical outcomes.

Outcome

Outcome refers to the end results for consumers of health care services. As with the other quality indicators, outcomes of nursing home care may be measured using objective and perceptual measures. Objective outcomes that have been studied include mortality, hospitalization, facilityacquired pressure sores, functional status changes, injuries, urinary incontinence, weight loss, thefts or abuse, staff injury or illness, and infectious disease. Perceptual measures of outcomes include pain control, depression, patient satisfaction, family satisfaction, service quality, and staff satisfaction (IOM, 1996). Outcomes have the advantage of being the most comprehensive of the three quality indicators and have increasingly been the focus in health services research.

Composite

When process and outcome indicators are combined, they are referred to as composite indicators of quality. These often are perceptual measures of quality. Examples of composite indicators used in studies of nursing homes include the Nursing Home Rating Scale (Linn, 1966), the Multiphasic Environmental Assessment Procedure (MEAP) (Moos & Lemke, 1992), the Quality Assessment Index (Gustafson, Sainfort, Van Konigsveld, & Zimmerman, 1990), and SERVQUAL (Parasuraman, Zeithaml, & Berry, 1988).

Deficiencies, complaints, certification ratings, and compliance rates also function as composite quality indicators (Davis, 1991). For example, the deficiency that nursing homes may be given related to pressure sore prevention and treatment is a composite quality indicator. Receipt of a deficiency is intended to reflect how well the nursing staff intervene to treat existing pressure sores (a process quality indicator) and how well pressure sores are prevented by examination of the prevalence rate of pressure sores (an outcome quality indicator).

CATEGORIZING STUDIES FOCUSED ON STAFFING AND QUALITY

Some of the quality indicators used in research studies are similar to those used by Medicare and Medicaid nursing home surveyors and various nursing home stakeholders' to define quality. Several of the indicators identified in Tables 1 and 2 are consistent with the 15 quality standards used by the Health Care Financing Adminis-tration (HCFA) to evaluate if nursing homes participating in Medicare or Medicaid are achieving the required level of quality for consumers. These standards include: resident rights, admission, transfer and discharge rights, resident behaviors and facility practices, quality of life, resident assessment, quality of care, nursing services, dietary services, physician services, specialized rehabilitative services, dental services, pharmacy services, infection control, physical environment, and administration. Interestingly, in a survey of administrators, nursing directors, ombudsmen, and survey training and certification coordinators from all 50 states and nursing home advocates from 34 states, Harrington, Mullan, and Woodruff (1999) found that quality of care, quality of life, and resident rights were the highest ranked quality indicators among this group of stakeholders.

Comparing the results of research studies on nursing home staffing and quality is made easier by tracking whether the quality indicator used functions as the dependent or independent variable in the study and how the quality indicator is categorized. The research studies reviewed in this article are grouped according to the dependent or independent variable used in the study and whether the variables functioned as structural, process, outcome, or composite quality indicators.

Table 1 is a summary of studies that have used total nursing staff hours as either the independent or dependent variable. Because total nursing staff hours are defined and measured differently in studies, a footnote in the Table describes how the variable was defined in each study. Studies that have used nursing staff skill mix as the independent or dependent variable are summarized in Table 2. Nursing staff skill mix is abo defined and measured differently among the various studies, as indicated in the Table's footnote.

Several of the studies identified in Tables 1 and 2 have used data from the Online, Survey, Certification, and Reporting system (OSCAR). The OSCAR system contains data obtained during each nursing home's annual survey. Data include facility characteristics, nurse staffing, resident characteristics, and facility deficiencies assessed by surveyors (Harrington, Carrillo, Thollaug, & Summers, 1996).

CURRENT FEDERAL NURSING STAFFING REGULATIONS AND STAFFING LEVELS

Although all nursing facilities are required to have 24-hour licensed nursing staff and an RN for at least 8 hours a day, 7 days a week, there is no federal requirement for a specific total staffing level in nursing homes. The regulatory standard is that there will be "sufficient staff to meet the needs of the residents on a 24-hour a day basis to provide nursing care to all residents in accordance with resident care plans" (AHCA, p. 135.8) Although a specific federal staffing level is not defined, some states do have mandated minimum staffing levels.

The most current analysis of national nursing homes staffing levels comes from a study conducted by Harrington, Carrillo, Mullan, and Swan (1998). Using OSCAR data, Harrington and her colleagues analyzed trends in the average nurse staffing levels for certified nursing facilities in the United States from 1991 through 1995. Although they found a small overall increase in staffing levels during the 5 years for RNs, licensed vocational or practical nurses (LVNs/LPNs), and certified nursing assistants (CNAs), much state and regional variation in staffing levels was identified. Higher RN and LVN/LPN levels were associated with states having higher resident case mix levels. Higher LVN/LPN staffing levels were associated with states having a higher percentage of Medicaid residents. Lower RN and LVN/LPN hours were associated with states having higher percentages of large facilities. Lower RN staff levels were associated with states having higher percentages of for-profit facilities.

In the period of July 1, 1995 through June 30, 1996, nursing staff comprised 60% of the total personnel hours in nursing homes in the United States. In this same period, in nonhospital-based nursing homes, there were an average of 3.3 hours of nursing care per resident a day. In hospital-based facilities, nursing care averaged 4.8 hours per resident a day (Harrington, Zimmerman, Karon, Robinson, & Beutel, 1997).

In 1995, the average ratio of RN hours per resident day was 0.5 hours or 30 minutes. Distributed over three 8-hour shifts, this reflects 10 minutes of RN time per shift.

The average ratio of LVN/LPN hours per resident day was 0.7 hours or 42 minutes. Over three 8-hour shifts, each resident received 14 minutes of LPN/LVN time. The average ratio of CNA hours per resident a day was 2.0 hours or 120 minutes, or 40 minutes of CNA time per resident for each 8-hour shift (Harrington, Carrillo, Mullan & Swan, 1998).

These descriptive statistics suggest that, in general, nursing personnel are a scarce resource in American nursing homes, with RNs being the most limited category of nursing personnel.

TOTAL NURSING STAFF HOURS

Structural Indicators of Quality

Resident payor source and ownership are important structural quality indicators to consider in nursing homes because more than half of nursing home care is financed by the federal Medicare and Medicaid programs and the majority of nursing home corporations are for-profit entities. In 1996, 65.9% of facilities were for-profit, 27.6% were not-forprofit, and 6.6% were governmentowned. For-profit enterprises are thought to be primarily motivated by profit maximization and efficiency while not-for-profit enterprises are thought to be primarily motivated by service delivery (Holmes, 1996).

Table

TABLE 2STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

One of the first studies to examine the relationship between the proportion of Medicaid patients, ownership, and patient care quality was conducted by Fottler, Smith, and James (1981). Using 46 for-profit nursing homes in a geographically contiguous area of southern California, the researchers found a negative relationship between for-profit status and nursing hours per patient day. In 1984, Elwell used 424 New York state old age institutions (OAIs) to examine the relationship between staffing patterns and ownership. Although no adjustment was made for case-mix differences among the homes, Elwell found that for-profit facilities had higher staff-patient ratios than not-for-profit facilities. Using a nationally representative sample of nursing homes and adjusting for case-mix, Cohen and Spector (1996) did not find that ownership type had a significant impact on total nursing staffing levels.

Harrington et al. (1997) used OSCAR data from all 50 states for the period of July, 1, 1995 through June 30, 1996 and found a negative relationship between total nursing staff hours and for-profit ownership and a negative relationship between total nursing staff hours and the proportion of residents having Medicaid as a payor source. Strengths of this study were that the population of nursing homes in the United States was included in the study and total nursing staff hours were defined in a comprehensive manner.

Outcome Indicators of Quality

Few studies have examined the relationship between total nursing staff levels and outcome indicators of quality of care. Porell, Caro, Silva, and Monane (1998) used Massachusetts Medicaid data from 1991 to 1994 to investigate associations between total nursing staff levels, or nursing staff intensity, and resident ADL status, mental status, and incontinence status. These characteristics were the observed health states of the residents over time.

The only significant relationship found was between mental status and nursing staff intensity. In addition, Porell, Caro, Silva and Monane (1998) examined outcome measures of survival rate and found no significant association between nursing staff intensity and survival rates.

Spector and Takada (1991) found higher total nursing staff levels related to functional improvements in residents in 2500 nursing home residents in 80 nursing facilities in Rhode Island. Low overall staffing levels with highly dependent residents were associated with a reduced likelihood of improvement.

Table

TABLE 2STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

Table

TABLE 2STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

Composite Indicatore of Quality

Johnson-Pawlson (1993) found that higher total nursing staff levels were negatively related to the quality of care deficiency index and the technical care deficiency index in a study of 198 Maryland nursing homes. The deficiency indexes were calculated by multiplying selected deficiencies by a weighted score obtained from an expert panel of nurses and ombudsmen.

Johnson, Cowles, and Simmens (1996) selected 1429 high-quality facilities and 210 low-quality facilities from 16,000 certified facilities described in the 1994 OSCAR system. High quality homes were those in the lowest quartile for the number of residents with pressure sores per number of bedbound residents, the percent of residents in restraints, and the drug error rate. Low-quality facilities where those in the top 25% for each of the three characteristics. High quality facilities had more total nursing staffing than low quality facilities.

Using OSCAR data from 13,770 facilities in all 50 states, Harrington et al. (1998) found facilities with fewer nurses were more likely to receive deficiencies of quality of care and quality of life, as well as total deficiencies.

These study findings provide some support for the hypothesis that there is a relationship between the total nursing staff levels and specific structural, outcome, and composite quality of care indicators in nursing homes.

NURSING STAFF SKILL MIX

Historically, nursing homes have used a predominance of paraprofessional nursing staff, including CNAs and LVNs/LPNs, rather than RNs. This variation in skill and educational background of nursing staff in nursing homes is commonly referred to as the skill mix. The nursing staff skill mix is believed to be as important, if not more important, of a predictor of quality of care in nursing homes as are total nursing staff hours (Cohen & Spector, 1996; IOM, 1996).

Structural Indicators of Quality

Several studies have examined the relationship between ownership and nursing skill mix. Holmberg and Anderson (1968) found no relationship between type of ownership and RN hours, LPN hours, and CNA hours. However, they found that notfor-profit facilities were associated with increased administrator hours and for-profit facilities were associated with fewer physician hours. Winn (1974) found an insignificant relationship between ownership and RN hours, LPN hours, and CNA hours. Elwell (1984) found a negative relationship between for-profit and nurse hours, RN hours, staff hours, and physician hours.

Using survey information and cost reports of 209 facilities in Kentucky from 1989, Davis (1993) found higher RN ratios in for-profit facilities. However, this was not accompanied by better quality measures. Davis speculated that forprofit facilities included in the sample may have offset the higher costs of increased RN staffing levels by reducing levels of other types of nursing staff personnel. Zinn (1993) found that nursing home markets with a higher proportion of forprofit facilities had higher LPN staffing and lower RN staffing. Cohen and Dubay (1990) found that higher nursing staffing levels were associated with facilities having a high percentage of Medicaid patients. Controlling for potential endogeneity among system variables, Aaronson, Zinn, and Rosko (1994) found that for-profit facilities had lower RN ratios than not-forprofit facilities.

Using a nationally representative sample of Medicaid certified nursing homes and their residents, Cohen and Spector (1996) examined the relationship between ownership and nursing staff skill mix. They found there was a positive relationship between RN to resident ratios and a negative relationship between LVN/LPN to resident ratios and not-for-profit facilities. This suggests that, in comparison with for-profit facilities, the not-for-profit facilities had a more skilled mix of staff.

Using states as the unit of analysis, Harrington, Carrillo, Mullan, and Swan (1998) examined trends in average nurse staffing levels from 1991 through 1995. They found that states with a higher percentage of for-profit facilities had lower RN staff levels. In states with a higher percentage of Medicaid residents, higher LVN/LPN staff levels were identified as a staffing trend.

Process Indicators of Quality

Most studies that have examined the relationship between nursing staff skill mix and quality have used outcome indicators. The few studies that used process indicators of quality related to nursing staff skill mix examined care plans, Medicare plans, diet plans (Nyman, 1988) catheter use (Cherry, 1991; Zinn, Aaronson, & Rosko, 1993), and restraint use (Aaronson, Zinn, & Rosko, 1994; Graber & Sloane, 1995).

Nyman (1988) did not find any significant relationship between either diet plans or care plans and licensed nurse staffing or CNA staffing. Nyman (1988) found a significant relationship between licensed nursing hours per patient and the quality variables of quality of life, maintenance of plant, and furnishing of rooms. In explaining his finding, Nyman suggested there was something unique about nursing labor in relation to quality as opposed to the other labor resources present in nursing homes.

Using 1987 survey data, Zinn et al. (1993) found no significant relationship between RN staffing and riskadjusted rates of urethral catheterization and use of physical restraints. Aaronson, Zinn, and Rosko (1994) found no association between the rate of restraint use and for-profit status. Graber and Sloane (1995) analyzed 1991 survey data from North Carolina nursing homes and found that the ratio of LVNs and CNAs to residents was associated with overall restraint use. Facilities tended to have greater proportions of restrained residents with decreased LVN and CNA staffing levels.

Outcome Indicators of Quality

Nursing staff skill mix related to outcome indicators of quality comprise the bulk of the nursing staff skill mix-quality literature. Using the outcome measures of mortality, function, and discharge, Linn, Gurel, and Linn (1977) found a significant negative relationship between RN staffing and mortality, and no significant relationships with LPN or CNA staffing. A positive significant relationship was found between RN staffing and the two outcomes of function and discharge. No significant relationships between LPN or CNA staffing and function or discharge were found.

Duffy (1988) found a significant positive relationship between LPN staffing and pressure sores, but none between RN staffing and pressure sores. Nyman (1988) used resident care and quality of life as outcome measures related to staffing skill mix. No significant findings were obtained in examining resident care and either CNA staffing or licensed nurse staffing levels. Quality of life was positively associated with licensed nurse staffing levels but was not significantly associated with CNA staffing. Braun (1991) found RN hours significantly and inversely related to mortality.

Zinn et al. (1993) found no significant relationship between RN staffing and risk-adjusted rates of mortality and pressure sores. Kolanowski, Hurwitz, and Taylor (1994) found that residents living in facilities where there was a higher ratio of licensed personnel (i.e., RNs, LVNs) to residents had fewer episodes of agitated psychomotor behavior. Bleismer (1994) studied Minnesota nursing homes and found that in the first year after patient admission, higher licensed nursing hours were significantly related to an increased probability of discharge, a decreased probability of death, and improved functional ability. Aaronson, Zinn, and Rosko (1994) found that pressure sore rate was negatively associated with forprofit status.

Cohen and Spector (1996) examined outcome indicators including mortality within a year, having a bedsore, and ADL status, controlling for health status, as they related to overall nurse staffing levels. Controlling for case mix, they found that nursing staff intensity, or the number of fulltime equivalent (FTE) staff per 100 residents, had a significant effect on mortality and ADL outcomes, but not bedsores. They found that a higher RN intensity was associated with a lower rate of mortality. A higher LPN intensity was associated with improved ADL outcomes.

Using 1996 OSCAR data from California nursing homes, Dellefield (1999) found that an increased ratio of total RN and LVN hours per resident day was associated with a higher-than-expected prevalence rate of pressure sores. One explanation for this finding is that licensed nursing staff may be increasingly involved in documentation of care and less available to be involved in direct clinical assessment of residents. Another possibility is that the clinical skill level related to pressure sore care of the individual licensed nurses is as important as the actual numbers of licensed nursing staff.

Composite Indicators of Quality

Duffy (1988) examined the relationship between rates of regulatory compliance and RN and LPN staffing hours. Hours for LPNs were negatively associated with regulatory compliance rates and positively associated with pressure sore percentages. Hours for RNs were not significantly related to regulatory compliance.

Munroe (1990) used 1986 data from 455 Medicare-certified skilled nursing facilities in California to determine the impact of RN staffing patterns on nursing home quality. The number of health-related deficiencies received at the annual survey was used as the quality indicator in the study. Health-related deficiencies were chosen among the 550 standards to comprise a group of 269 standards defined as health related. Seventy-six (28.3%) of the standards involved nursing services. The criteria used to determine which standards were health-related were not explained.

Munroe found a small but significant positive relationship between nursing home quality and the ratio of RN hours to LVN hours per patient day. Fewer numbers of deficiencies were related to a higher ratio of RNs to LPNs. Control variables included size, payor mix, average daily costs, ownership, nursing personnel salary, and turnover. Case mix was included to reflect facility differences in the proportion of residents who were incontinent, had pressure sores, and needed assistance with ADLs and mobility.

Cherry (1991) studied 134 Medicaid and Medicare certified nursing homes in Missouri and found increased RN hours negatively associated with a calculated composite index of poor nursing care in intermediate care facilities and skilled nursing homes with a high RN to resident ratio. The index was composed of outcomes (incidence of pressure sores and urinary tract infections per incontinent resident) and care processes (rates of antibiotic use per resident and catheterizations per incontinent resident).

Graber and Sloane (1995) studied 195 North Carolina nursing homes using 1991 data from OSCAR, the North Carolina Division of Medical Assistance, and the North Carolina Office of State Health Planning. A probit model was used to identify predictors of the proportion of restrained patients in each facility and the receipt of a restraint violation. Receipt of a restraint violation was associated with facility size, direct costs per patient day, the proportion of restrained patients, use of bladder training in less than 3% of residents, and the proportion of residents with organic brain syndrome. The finding of an association of the proportion of restrained patients and the receipt of restraint violations lends some credibility to the survey process.

Johnson-Pawlson and Infeld (1996) hypothesized that a higher level of RN staffing would be associated with a higher quality of nursing care. They studied 198 Maryland nursing facilities using 1991 OSCAR data and questionnaire data from ombudsmen. Nursing care quality was measured using 78 items from the 1991 Long Term Care Survey identified by ombudsmen as specifically related to nursing care. The independent variable was the RN t'l'E per resident. Control variables included case mix, facility ownership, and payor mix.

An overall measure of nursing care quality was constructed for each facility by multiplying the actual deficiencies received among the 78 selected types by the average score for the strength of its relationship to nursing, as derived from the ombudsmens' rating of this relationship. This total score, called the overall deficiency index, was used to represent nursing care quality in the study. Dimensions of nursing care quality were identified as resident rights, resident behavior, quality of life, resident assessment, and quality of care and a deficiency index for each dimension of nursing care quality was calculated using the same method.

Johnson et al. (1996) found that high quality facilities were 1.73 times more likely to have increased RN staff hours per resident day than were low quality homes. These homes were 1.04 times more likely to have increased LPN/LVN staff hours per resident day than were low quality homes.

However, study findings did not support the hypothesis related to RN staffing. Contrary to what was hypothesized, the ratio of RNs to residents was directly related to a deficiency index of resident rights. As the ratio of RNs to residents increased, the number of resident rights deficiencies increased. This finding suggests that RNs may not have been sufficiently sensitive to issues related to resident rights. The study was strengthened by the use of case mix as a control variable.

Dellefield (1999) examined the relationship between nursing staff skill mix and the receipt of a deficiency for pressure sore prevention and treatment. Although there was no relationship between licensed nursing staffing levels and the receipt of this deficiency, facilities having a higher ratio of CNA hours per resident day were less likely to receive a deficiency.

These study findings provide inconsistent support for the hypothesis that there is a relationship between different types of nursing staff skill mix and specific structural, process, outcome, and composite indicators of quality.

STUDY LIMITATIONS IN THE NURSE STAFFING/INDICATORS OF QUALITY RELATIONSHIP

Researchers have not been consistent in their categorization and use of variables as structure, process, or outcome quality indicators. For example, structure, process, or outcome measures are used as quality indicators related to explanatory or predictor variables that are categorized as the same dimension of quality. An illustration of this is the study of the relationship between expenditures and nursing staffing levels. Both variables are types of structural quality indicators. One would expect that higher staffing levels would be related to increased expenditures. Another type of inconsistency occurs when a variable is defined as a structural quality indicator in one study but is used as an outcome quality indicator in another study.

There has been significant variability in the use of risk-adjusted outcome measures in studies. Riskadjusted outcome measures control for differences in the clinical profile of residents included in studies. If these structural differences in resident case mix are not accounted for in the analysis, study results may be spurious when quality comparisons between facilities are made (Zimmerman, Karon, Arling, Clark, Collins, Ross, & Sainfort, 1995).

All of these studies are constrained by an inherent limitation found in much of health services research. Although the randomized clinical trial would be the best method for studying effectiveness of nursing staffing on various types of quality indicators, this approach is not a practical possibility (Spector & Takada, 1991). Also, studies are undertaken at different times, in different states, within the context of a dynamic regulatory environment.

CLINICAL IMPLICATIONS

Total Nursing Staff Hours

Structural quality indicators. Although a relationship between forprofit ownership and lower total nursing staff hours has consistently been found in various studies, the difficulty with the findings is that two structural measures are being compared with one another. The impact of ownership status and staffing on patient outcomes was not examined. The inference that for-profit ownership and lower total nursing staff hours suggest a lower level of quality seems unwarranted.

Outcome quality indicators. Although limited in number, the research evidence for outcome quality indicators and total staffing levels makes a more compelling case that higher overall staffing levels are associated with better resident outcomes. Mental status and ADL status may be other global outcome measures that are especially sensitive to overall staffing levels. Based on these findings, a director of nursing (DON) might use the quality indicators, currently part of the revised long-term care survey process (AHCA, 1999), to better understand how overall staffing levels are associated with selected quality indicators. For example, a DON might compare trends in overall staffing levels with trends in the quality indicator domain of physical functioning. Physical functioning includes Minimum Data Set (MDS) data on mobility, ADL functioning, and range of motion. Similarly, trends in the domains of behavioral/emotional patterns and skin care (two other quality indicators) might be compared to trends in overall staffing levels.

Composite quality indicators. Deficiencies have been consistently associated with reduced staffing levels. While the consistency of these findings suggests that reduced staffing is associated with a wide variety of survey deficiencies, this association needs to be viewed cautiously. There is significant debate about the reliability and validity of the current survey process, from which these deficiencies are derived (Cella, 1996). Studies using survey deficiencies as composite quality-ofcare indicators have not made adjustments for confounding variables that may influence the determination of a deficiency by state and federal regulators. These include regional variations in surveyor practice and inability to factor in risk in determining the unavoidability of a negative outcome (Cella, 1996). Given this caveat, nurses might use these study findings to support their case for increased total nursing staffing levels to achieve better survey results.

Nursing Staff Skill Mix

Although the most recent and comprehensive study of the relationship between ownership and RN hours per patient in the United States found that for-profit facilities had lower levels of RN hours (Harrington et al., 1998), other studies have produced mixed results when nursing staff skill mix has been examined. Based on these findings, no specific clinical implications may be drawn regarding nursing staff skill mix and nursing home ownership.

Little has been studied regarding process quality indicators and nursing staffing skill mix, in part because of the difficulty in measuring processes of care. When one thinks about the great extent to which processes of care are defined by various types of health care regulation and accreditation agencies, it is noteworthy that so little is known about the relationship between processes of care and nursing staff skill mix in nursing homes. Consequently, it is difficult to make clinical inferences based on this research literature.

RNs. The relationship between outcome and composite quality indicators and nursing staff skill mix has been most extensively studied in the nursing-home-quality literature. The most consistent evidence of a positive relationship between RN staffing ratios and outcomes has been found with global outcomes such as mortality rates, improvements in physical functioning, and discharge rates. Mixed results have been produced when the relationship between RN staffing ratios and the prevalence rate of pressure sores was examined. While some of these are encouraging findings to support the view that increased RN staffing enhances the achievement of desired patient outcomes, a relatively limited number of types of outcomes have been examined.

Clinically, these findings suggest that RNs may make their significant contribution in the nursing home as managers of care, overseeing the achievement of certain global outcomes, and are uniquely qualified to provide this oversight. Using RNs to provide direct patient care, such as administering medications and doing treatments, may not be the most prudent use of this scarce and valuable resource. Time spent performing tasks that could be performed by an LVN leaves less time available to participate in the care planning and case management that the RN is uniquely qualified to perform.

This is not to suggest that direct patient care and patient contact is not essential for the RN. Because the RN is such a valuable resource to the nursing home, time needs to be strategically allocated to meet broad goals of the residents and patients. Professional nurses are challenged to develop RN roles, methods of communicating among LVN/LPN and CNA team members, and care delivery systems that capitalize on the unique contributions that RNs make to the achievement of quality outcomes.

In spite of significant problems related to the reliability and validity of the use of survey deficiencies as quality indicators, researchers have fairly consistently found a relationship between survey deficiencies and lower levels of RN hours per patient. These findings might be used to help professional nurses advocate for higher ratios of RNs in facilities that may have a pattern of significant deficiencies in the presence of minimal or low levels of RN staffing.

LVNs. The relationship between LVN staffing ratios and outcome and quality indicators has either been examined directly in studies, or LVNs and JRJMs have been grouped together into the category of licensed staff. These studies also used a limited range of outcomes and resulted in inconsistent findings, making it difficult to draw clinically relevant inferences from this research literature.

CNAs. Among the quality indicators studied, research findings do not provide support for the assertion that there is a relationship between CNA staffing ratios and those outcomes (Table 2). Only one study found that receipt of a survey deficiency was associated with a lower level of CNA staffing ratios (Dellefield, 1999). Admittedly, a limited number of outcomes and types of deficiencies have been studied, which may contribute to these interesting and unexpected findings.

Professional nurses need to explore other aspects of human resource management, in addition to CNA staffing levels that may improve the contributions that CNAs make to nursing home quality. Such issues as CNA orientation, inservice education for CNAs, improved supervisory skills, better care delivery systems, and improved clinical information management systems may significantly contribute to improved outcomes.

The nature of the research findings on nursing skill mix and quality indicators makes it difficult to make many specific clinical recommendations. In general, a DON might evaluate the skill mix of the nursing staff and make adjustments based on trends in patient acuity. With more routinized and predictable chronic care patients, CNAs and LVNs may predominate. For more medically and psychosocially complex patients and residents, a greater RN presence may be warranted. In addition, using effective management practices with CNAs may be as significant as focusing on staffing levels in achieving patient outcomes.

RESEARCH IMPLICATIONS

Both the inconsistencies of the findings and the limited types of quality indicators that have been studied in relation to nursing staffing suggest that more research is warranted.

Research recommendations made by the Institute of Medicine (IOM) Committee on the Adequacy of Nurse Staffing in Hospitals and Nursing Homes (1996) included a request for increased funding of research related to staffing levels, skill mix, and studies focused on quality of care and outcomes in nursing homes. The Health Care Financing Administration (HCFA) has contracted with Abt Associates to conduct a study to help Congress determine the impact of increased staffing ratios on improved care outcomes and costs (Childs, 2000).

More research on the relationship between nursing skill mix and quality indicators is of particular importance given some of the study findings that have been found when various ratios of types of nursing staff are compared. Results from nursing skill mix studies can provide insights into strategic ways in which the total nursing staff levels might be changed. This is important for several reasons. Registered nurses, LVNs, and CNAs have significantly different wage levels, and the cost of staffing recommendations will vary widely according to the types of nursing personnel being recommended. The nursing home industry has used LVNs/LPNs as a substitute for RNs, perhaps as a cost saving measure (Cohen & Spector, 1996). Although the professional nursing community has long maintained that RNs and LVNs are not substitutes for one another, there needs to be more research evidence in the nursing-home-quality literature to support this assertion.

When considering RN skill mix, nurse researchers need to examine the impact of different educational preparations of RNs on the achievement of patient outcomes because most RNs working in nursing homes have not earned baccalaureate degrees. Some of the inconsistencies in findings related to RN staffing and quality outcomes may be a result of this variation in education preparation.

More nurse researchers should use existing data sets, such as OSCAR and the MDS, to address staffing and quality questions relevant to nursing home practice. The advantage in using these data sets is that large study sample sizes may be achieved. Nurse researchers need to investigate what combination of skill levels of nursing staff members contribute to quality health outcomes in the most cost-effective manner.

More nursing homes are developing the capacity to perform institutional research and examine the relationship among nursing staffing and skill mix and various quality indicators. The recent implementation of the investigative protocol on sufficient staff in nursing services as a component of the standard nursing home survey process will motivate nursing home providers to take another look at the relationship between their staffing levels and achievement of desired resident outcomes (AHCA, 1999).

POLICY IMPLICATIONS

Public policy reflects the standards a society has for itself and expresses in laws that governmental bodies establish for society's benefit (Koff & Park, 1993). Increasingly, the public policy implications of nursing staffing levels and nursing-home quality have been explored by provider and consumer groups. For example, in 1996, the IOM Committee on the Adequacy of Nurse Staffing in Hospitals and Nursing Homes made recommendations on staffing and quality in nursing homes (IOM, 1996). The committee recommended that "Congress require by the year 2000 a 24-hour presence of registered nurse coverage in nursing facilities as an enhancement of the current 8-hour requirement specified under OBRA 87" (p. 18). The use of geriatric nurse specialists and geriatric nurse practitioners was recommended, as was increased emphasis on the educational preparation of directors of nursing.

Consumer groups such as the National Citizens' Coalition for Nursing Home Reform (NCCNHR) and the American Association of Retired Persons (AARP) have been active on a national and state level with various legislative initiatives focused in nursing staffing in nursing homes.

Some members of the 1998 New York University and the Hartford Institute for Geriatric Nursing conference sponsored by AHCPR recommended that a bachelors' degree needed to be the imnimum education preparation for the DON, that nursing homes would use DONs with advanced practice degrees, and that specific CNA-, LVN-, and RN-topatient ratios be instituted (Harrington et al., 1998).

When thinking about nursing home staffing and policy implications, one needs to be mindful that health care policy is established in the context of politics and special interests, as well as research-based evidence. How one views the strength of the research evidence about the relationship between quality indicators and nursing staffing and its policy implications may be as much a statement about one's political perspective as it is a statement about one's skills as a researcher. Researchers may genuinely differ in the inferences drawn from study findings. As nursing matures as a profession and discipline, there will be an increased tolerance for a diversity of perspectives about policy issues that affect nursing.

Based on the literature reviewed, it seems premature for professional nurses or consumers to advocate for mandated nursing staff skill mix to patient ratios. There seems to be more evidence to advocate for a mandated minimum level of total nursing care hours, allowing each provider to determine how to configure the skill mix. Advocacy for a 24-hour presence of RN coverage, as recommended by the 1996 IOM Committee, is better supported by the research evidence.

Greater advocacy for policy strategies designed to increase professional accountability of the nurse, the educational level of nursing staff currently working in nursing homes, and nurses' management skills may be a more prudent use of societal resources. These strategies are suggested because relatively little of the variation in quality indicators has been explained by nursing staffing variables in the studies summarized in this literature review. Professional nursing advocacy for mandated staff skill mix ratios might even be perceived as self-serving by some (Spetz, 1999).

CONCLUSION

Professional nurses share a common desire to assist residents and patients in experiencing a high level of quality of care and quality of life in nursing homes. Achieving this goal will undoubtedly require a variety of clinical, management, and policy interventions. As an understanding of the complex relationship between staffing levels, skill mix, and quality is enhanced, more effective clinical, organizational, and governmental strategies will be developed and implemented to achieve this goal.

REFERENCES

  • Aaronson, W., Zinn, J., & Rosko, M. (1994). Do for-profit and not-for-profit homes behave differently? The Gerontologist, 34, 775-786.
  • American Health Care Association. (1999). The long term care survey. Washington, DC: Author.
  • Bleismer, M. (1994). Outcomes of Minnesota nursing home residents and their relationship to structural and process-related attributes. Dissertation Abstracts International, AAT 9509233.
  • Braun, B. (1991). The effect of nursing home quality on patient outcome. Journal of the American Geriatrics Society, 39, 329-338.
  • Cella, M., Gabay, M., Teitelbaum, M., Walker, A., & Kramer, A. (1996). Evaluation of the long-term care survey process final report. Bethesda, MD: Abt Associates, Inc.
  • Cherry, R. (1991). Agents of nursing home quality of care: ombudsmen and staff ratios revisited. The Gerontologist, 31, 302-308.
  • Childs, N. (2000). Staffing standard takes center stage at senate forum. Provider, 26(1), 10-11.
  • Cohen, J., & Dubay, L. (1990). The effects of Medicaid reimbursement method and ownership on nursing home costs, case mix, and staffing. Inquiry, 27(2), 183-200.
  • Cohen, J., & Spector, W. (1996). The effect of Medicaid reimbursement on quality of care in nursing homes. Journal of Health Economics, /5(4), 23-48.
  • Davis, M. (1991). On nursing home quality: A review and analysis. Medical Care, 48(2), 129-166.
  • Davis, M. (1993). Nursing home ownership revisited: Market, cost, and quality relationships. Medical Care, 31(11), 1062-1068.
  • Dellefield, M.E. (1999). Predictors of quality of care in California nursing homes. Dissertation Abstracts International, 9943785.
  • Donabedian, A. (1980). The definition of quality and approaches to its assessment (Vol. 1). Ann Arbor, MI: Health Administration Press.
  • Donabedian, A. (1992). The role of outcomes in quality assessment and assurance. Quality Review Bulletin, 11, 356-360.
  • Duffy, J. (1988). Quality care measurement: Not always what it seems. Proceedings of Southwestern Academy of Management, 30, 260-264.
  • Elwell, F. (1984). The effects of ownership on institutional services. The Gerontologist, 24(1), 77-83.
  • Georgopoulos, B. & Mann, F. (1962). The Community General Hospital. New York: Macmillan.
  • Graber, D. & Sloane, P. (1995). Nursing home survey deficiencies for physical restraint use. Medical Care, 33(10), 1051-1063.
  • Gustafson, D., Sai nf ort, F., Van Konigsveld, & Zimmerman, D. (1990). The quality assessment index (QAI) for measuring nursing home quality. Health Services Research, 25(1), 97-110).
  • Harrington, C., Carrillo, H., Mullan, J., & Swan, J. (1998). Nursing facility staffing in the states: The 1991-1995 period. Medical Care Research and Review, 55(3), 334-363.
  • Harrington, C., Carrillo, H., Thollaug, S. & Summers, P. (1996). Nursing facilities, staffing, residents, and facility deficiencies, 1991-95. Department of Social and Behavioral Sciences, University of California San Francisco.
  • Harrington, C, Kovner, C, Mezey, M., Kayser-Joncs, J., Burger, S., Mohler, M., Reilly, K., Burke, R., & Zimmerman, D. (1998). Experts recommend minimum nurse staffing standards for nursing facilities in the US. Unpublished manuscript. University of California at San Francisco.
  • Harrington, C, Mullan, J., Woodruff, L.C., Burger, S.G., Carrillo, H., & Bedney, B. (1999). Stakeholders opinions regarding important measures of nursing home quality for consumers. American Journal of Medical Quality, 14(3), 124-132.
  • Harrington, C, Zimmerman, D., Karon, S., Robinson, J. & Beutel, P. (1997). Nursing home staffing and its relationship to quality. Unpublished manuscript, University of California, San Francisco.
  • Harrington, C, Zimmerman, D., Karon, S., Robinson, J. and Beutel, P (1998). Nursing home staffing and its relationship to deficiencies. Unpublished manuscript, University of California, San Francisco.
  • Holmberg, R. & Anderson, N. (1968). Implications of ownership and nursing home care. Medical Care, 6(4), 300-307.
  • Holmes, J. (1996). The effects of ownership and ownership change on nursing home industry costs. Health Services Research, 31(3), 327-346.
  • Institute of Medicine. (1996). Nursing staff in hospitals and nursing homes: Is it adequate ? Washington, DC: National Academy Press.
  • Johnson, J., Cowles, C, & Simmens, S. Quality of care and nursing staff in nursing homes. (1996). In G. Wunderlich, F. Sloan, & C. Davis (Eds). Nursing staff in hospitals and nursing homes: Is it adequate? (pp. 426452). Washington, DC: National Academy Press.
  • Johnson-Pawlson, J. (1993). The relationship between nursing staff variables and quality of care in nursing facilities. Dissertation Abstracts International, 9316112.
  • Johnson-Pawlson, J., & Infeld, D. (1996). Nurse staffing and quality of care in nursing facilities. Journal of Gerontological Nursing, 22(8), 36-45.
  • Koff, T, & Park, R. (1993). Aging public policy: Bonding the generations. Amityville, New York: Baywood.
  • Kolanowski, A. Hurwitz, S., Taylor, L., Evans, L., & Strumpf, N. (1994). Contextual factors associated with disturbing behaviors in institutionalized elders. Nursing Research, 43(2), 73-79.
  • Linn, M. (1996). A nursing home rating scale. Geriatrics, 21(5), 532-538.
  • Linn, M., Gurel, L., & Linn, B. (1977). Patient outcome as a measure of quality of nursing home care. American Journal of Public Health, 67(4), 337-344.
  • Moos, R., & Lemke, S. (1992). Multiphasic environmental assessment procedure user's guide. Palo Alto: Center for Health Care Evaluation, Department of Veterans Affairs and Stanford University Medical Centers.
  • Mullins, L., Nelson, C, Busciglio, H., & Weiner, H. (1988). Job satisfaction among nursing home personnel: The impact of organizational structure and supervisory power. The Journal of Long-Term Care Administration, Spring, 12-17.
  • Munroe, D. (1990). The influence of registered nurse staffing on the quality of nursing home care. Research in Nursing and Health, 13, 263-270.
  • Nyman, J. (1988). Improving the quality of nursing home outcomes: Are adequacy- or incentive-oriented policies more effective? Medical Care, 26(12), 1158-1171.
  • Parasuraman, A., Zeithaml, V, & Berry, L. (1988). SERVQUAL: A multi item scale for measuring consumer perception of service quality. Journal of Retailing, 64( Spring) 12-40.
  • Porell, E, Caro, F., Silva, ?., & Monane, M. (1998). A longitudinal analysis of nursing home outcomes. Health Services Research, 33(A), 835-865.
  • Sainfort, F, Ramsay, J., Ferreira, P., & Mezghani, L. (1994). A first step in total quality management of nursing facility care: Development of an empirical causal model of structure, process, and outcome dimensions. Quality: Journal of the American College of Medical Quality, 9(2), 74-86.
  • Spector, W, & Takada, H. (1991). Characteristics of nursing homes that affect resident outcomes. Journal of Aging and Health, 3(4), 427-454.
  • Spetz, Joanne. (1999). Victor Fuchs on health care, ethics, and the role of nurses. Image: Journal of Nursing Scholarship, 31(3), 255258.
  • Winn, S. (1974). Analysis of selected characteristics of a matched sample of nonprofit and proprietary nursing homes in the state of Washington. Medical Care, 12(3), 221-228.
  • Zimmerman, D., Karon, S., Arling, G., Clark, B., Collins, T., Ross, R., & Sainfort, F. (1995). Development and testing of nursing home quality indicators. Health Care Financing Review, /6(4), 107-127.
  • Zinn, J. (1993). Inter-SMSA variation in nursing home staffing and resident management practices. The Journal of Applied Gerontology, 12(2), 206-224.
  • Zinn, J., Aaronson, W, & Rosko, M. (1993). The use of standardized indicators as quality improvement tools: An application in Pennsylvania nursing homes. Quality: Journal of the American College of Medical Quality, 8(2), 72-78.

TABLE 1

STUDIES RELATING TOTAL NURSING STAFF HOURS AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

TABLE 2

STUDIES RELATING NURSING STAFF SKILL MIX AND NURSING HOME QUALITY

10.3928/0098-9134-20000601-05

Sign up to receive

Journal E-contents