Nursing Homes: Quality of Care More Related to Staffing than
Spending (13-JUN-02, GAO-02-431R).
Costs for nursing home care have almost doubled since 1990, from
$53 billion to $92 billion in 2000. Much of that spending has
been financed with public monies. Under the Medicare and Medicaid
programs, the federal government financed 39 percent of the
nation's nursing home spending in 2000, up from 28 percent in
1990. As federal outlays have grown, Congress has focused
attention on the quality of care delivered and the level of
staffing in nursing homes. Nursing home expenditures per resident
day varied considerably across Ohio, Mississippi, and
Washington--the three states covered in GAO's survey. Although
the total level of spending varied, the average share devoted to
resident-care activities such as nursing care and medical
supplies was relatively stable. The share of spending devoted to
buildings and equipment, by comparison, was more variable. Homes
in Ohio and Washington that provided more nursing hours per
resident day, especially nurses' aide hours, were less likely
than homes providing fewer nursing hours to have repeated serious
or potentially life-threatening quality problems. However, GAO
found no clear relationship between a nursing home's spending per
resident day and the number of serious quality problems.
-------------------------Indexing Terms-------------------------
REPORTNUM: GAO-02-431R
ACCNO: A03626
TITLE: Nursing Homes: Quality of Care More Related to Staffing
than Spending
DATE: 06/13/2002
SUBJECT: Health care costs
Health care services
Nursing homes
Comparative analysis
Medicare Program
Mississippi
Ohio
Washington
Medicaid Program
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO Product. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
******************************************************************
GAO-02-431R
GAO- 02- 431R Nursing Home Expenditures and Quality United States General
Accounting Office
Washington, DC 20548
June 13, 2002 The Honorable John B. Breaux Chairman The Honorable Larry E.
Craig Ranking Minority Member Special Committee on Aging United States
Senate
The Honorable Charles E. Grassley Ranking Minority Member Committee on
Finance United States Senate
Subject: Nursing Homes: Quality of Care More Related to Staffing than
Spending
Since 1990, national expenditures for nursing home care have almost doubled,
climbing from $53 billion to $92 billion in 2000. An increasing amount of
that spending has been financed with public monies. Under the Medicare and
Medicaid programs, the federal government financed 39 percent of the
nation?s nursing home spending in 2000, up from 28 percent in 1990. As
federal outlays have grown, the Congress has focused attention on the
quality of care delivered and the level of staffing in nursing homes.
Questions have arisen about how federal dollars are being spent and the
relationship between nursing homes? spending and quality of care. To better
understand what public monies are purchasing, whether nursing homes with
high total expenditures spend more on nursing care, and how individual
nursing homes? expenditures relate to the quality of care they furnish, you
asked us to examine (1) nursing home expenditures, particularly those
devoted to resident care, and (2) whether there is any relationship among
nurse staffing levels, quality of care, and expenditures.
We examined the spending and staffing for freestanding 1 nursing homes in
three states- Mississippi, Ohio, and Washington- that are geographically
diverse and that collect the necessary information to adjust homes? spending
for differences in residents? care needs. We analyzed 1999 cost data
included in Medicaid cost reports, which include nursing homes? spending for
all residents. We adjusted these spending data to account for differences in
the resource needs of residents across homes and
1 Freestanding nursing homes are not part of another facility such as an
acute care or rehabilitation hospital.
2 GAO- 02- 431R Nursing Home Expenditures and Quality for differences in
wages across geographic areas. We also analyzed federal data on the results
of state surveys of nursing home quality in those states. In addition, we
discussed cost reporting requirements and payment methods with Medicaid
officials from each state. Although our findings cannot be generalized to
the country as a whole, they provide insights into nursing home spending
patterns. We conducted our work from December 2000 through April 2002 in
accordance with generally accepted government auditing standards. (For a
detailed discussion of our scope and methodology, see encl. I.)
In summary, we found that nursing homes? expenditures per resident day
varied considerably across the three states. 2 After controlling for
differences in the care needs of residents and in area wages, average total
nursing home expenditures were $133 per resident day in Ohio and $132 per
resident day in Washington and were about 23 percent less in Mississippi.
Although the total level of spending varied, the average share devoted to
resident- care activities such as nursing care 3 and medical supplies was
relatively stable across the states, averaging slightly more than 50 percent
of total expenditures in all three states. The share of spending devoted to
buildings and equipment, by comparison, was more variable. For nursing homes
within each of the states, spending also varied widely. Nursing homes with
high total expenditures tended to have high nursing care expenditures, but
as spending per resident day increased, the proportion of spending devoted
to nursing care tended to decline. Some of the variation in spending within
the states may be due to Medicaid payment policies, which attempt to
influence the resources nursing homes use, generally by encouraging spending
on nursing services and limiting payments for other services. For two of the
states we examined, homes with a high proportion of Medicaid residents had
lower daily expenditures per resident than homes with a low share of
Medicaid residents.
Homes in Ohio and Washington that provided more nursing hours per resident
day, especially nurses? aide hours, were less likely than homes providing
fewer nursing hours to have had repeated serious or potentially life-
threatening quality problems, as measured by deficiencies detected during
state surveys. But we found no clear relationship between a nursing home?s
spending per resident day and the number of serious quality problems. Higher
spending on nursing was associated with fewer deficiencies only in
Washington; Mississippi homes with higher nursing expenditures had slightly
more deficiencies, while in Ohio we found no relationship between nursing
expenditures and deficiencies. We received comments from state officials
from Mississippi, Ohio, and Washington, and from two experts in nursing home
costs and quality, and we have incorporated into these into the report as
appropriate. 2 Due to differences in the state reporting requirements for
spending on ancillary services (such as
physical, occupational, and speech therapy, and drugs and laboratory
services), these services were excluded from this analysis. 3 Throughout
this letter, ?nursing? refers to services provided by registered nurses,
licensed practical nurses, and nurses? aides.
3 GAO- 02- 431R Nursing Home Expenditures and Quality BACKGROUND
Nursing homes in the United States play an essential role in our health care
system, caring for 1.6 million elderly and disabled persons who are
temporarily or permanently unable to care for themselves but who do not
require the level of care furnished in an acute care hospital. They provide
a variety of services to residents, including nursing and personal care;
physical, occupational, respiratory, and speech therapy; and medical social
services. On average, 67 percent of nursing home residents have their care
paid for through the Medicaid program, while 9 percent are covered by
Medicare, and 24 percent are covered by other payers or pay for the care
themselves.
Nursing homes treat people with a wide range of clinical conditions. Most
facilities historically have served residents whose primary need is
custodial care. Nursing homes also treat residents with more complex needs,
furnishing higher intensity rehabilitative therapies and nursing services-
such as ventilator care- that previously were provided only in hospital
settings. The mix and amount of resources nursing homes use determine the
cost of the care they provide. These resources include nurses and nurses?
aides (referred to in this letter as ?nursing?); medical supplies; other
resident care resources such as dieticians, social workers, directors of
nursing, and staff and supplies needed for medical recordskeeping; home
operations such as staff and supplies needed for housekeeping, food
services, laundry, and maintenance; capital such as depreciation on
buildings, equipment, and furnishings; and administration such as
administrator and clerical salaries and office supplies. The states and the
federal government share responsibility for oversight of the quality
of care provided in nursing homes. The federal government, through the
Centers for Medicare and Medicaid Services (CMS), 4 establishes the
requirements that nursing homes must meet to participate in the Medicare and
Medicaid programs. CMS contracts with state agencies to check compliance
with these standards through onsite surveys conducted at every home at least
once every 15 months. During these surveys, state surveyors spend several
days on site, conducting a broad review of care and services to ensure that
homes are meeting the needs of residents. 5 Deficiencies identified during
the survey process are placed in 1 of 12 categories
depending on the extent of resident harm (severity) and the number of
residents adversely affected (scope). The most serious category is for a
widespread deficiency that causes actual or potential for death or serious
injury to residents; the least
4 On July 1, 2001, the Health Care Financing Administration (HCFA) was
renamed the Centers for Medicare and Medicaid Services (CMS). This letter
refers to the agency as HCFA when referring to actions taken before the name
change and as CMS when referring to actions taken since the name change.
5 Surveyors assess the provision of services in residents? plans of care,
the use of physical restraints, the incidence of pressure sores, the
treatment of incontinence, the use of antipsychotic drugs, the rate of
medical errors, the adequacy of the nursing staff, the maintenance of
residents? quality of life and
personal dignity, the facility?s cleanliness, and the thoroughness of
employee background checks, among many other areas.
4 GAO- 02- 431R Nursing Home Expenditures and Quality serious category is
for an isolated deficiency that poses no actual harm and has potential only
for minimal harm.
Under their shared responsibility for nursing home oversight, state agencies
identify and categorize deficiencies and make referrals for proposed
sanctions to CMS. 6 Most homes are given a grace period, usually 30 to 60
days, to correct deficiencies. Usually, states do not refer homes to CMS for
sanctions unless the homes fail to correct their deficiencies within that
grace period. CMS policies call for states to
refer immediately for sanction those facilities found to have repeated
severe deficiencies.
The survey process has revealed many homes to be deficient in guaranteeing
the safety and welfare of their residents. Each year more than 25 percent of
nursing homes are found to have deficiencies that cause actual harm to
residents or place
them at risk of death or serious injury. 7 Even so, in previous work, we
concluded that state surveys of nursing home quality likely understate the
extent of serious care problems, for several reasons. 8 First, homes may be
able to mask some problems because they can predict the timing of annual
reviews and therefore can prepare for them. Surveyors can also miss problems
that affect the health and safety of nursing home residents because of the
sampling methods used to select the residents whose care will be reviewed
and because the reviews rely heavily on medical records, which are not
always accurate.
In addition, the subjective nature of the survey process means that
surveyors may apply standards unevenly. Indeed, we previously have reported
that during attempts to validate the findings of state surveyors, federal
surveyors have found more than three times the number of serious care
problems recorded by state surveyors. 9 Further, we have found considerable
variation nationwide in the reporting of
deficiencies: there was more than a five- fold difference across states in
the percentage of homes found by state surveyors to have actual harm and
immediate jeopardy deficiencies. 10 Such differences in reporting make
comparisons across states difficult since it cannot be determined whether
observed differences are due to real variations in quality or to
inconsistent application of standards. In spite of these shortcomings, the
deficiency data are the best available national source of information about
the quality of care provided in the nation?s nursing homes.
6 States are responsible for enforcing standards in homes with only Medicaid
certification. 7 U. S. General Accounting Office, Nursing Homes: Additional
Steps Needed to Strengthen Enforcement of Federal Quality Standards, GAO/
HEHS- 99- 46 (Washington, DC: March 1999). 8 U. S. General Accounting
Office, California Nursing Homes: Care Problems Persist Despite Federal and
State Oversight, GAO/ HEHS- 98- 202 (Washington, DC: July 1998). 9 U. S.
General Accounting Office, Nursing Homes: Sustained Efforts Are Essential to
Realize the Potential of the Quality Initiatives, GAO/ HEHS- 00- 197
(Washington, DC: Sept. 2000). 10 Immediate jeopardy deficiencies are those
that have caused or have the potential to cause serious injury or death.
5 GAO- 02- 431R Nursing Home Expenditures and Quality NURSING HOMES?
EXPENDITURES VARY CONSIDERABLY, BUT SHARE DEVOTED TO RESIDENT CARE
COMPARATIVELY UNIFORM
We found significant variation in nursing home spending across Mississippi,
Ohio, and Washington. Spending on capital was particularly variable. By
comparison, the proportion of spending devoted to resident- care activities,
such as nursing care and medical supplies, was relatively stable across the
states, averaging more than 50 percent of total expenditures for all three
states. Within each state, nursing home spending varied considerably. We
found that nursing homes with high total expenditures tended to have high
levels of spending in all expenditure categories, including nursing, but
those homes devoted a smaller share of their total expenditures to nursing
compared to homes with low levels of spending. The variation in spending
within states may be explained in part by differences across homes in their
reliance on Medicaid reimbursement. Homes with large shares of Medicaid
residents had lower daily expenditures than homes with low Medicaid shares.
Further, Medicaid policies appeared to influence the resources used by
nursing homes. Within each state, spending varied less for expenditure
categories for which Medicaid payments were more restricted.
We found that average total nursing home expenditures- excluding spending on
therapies, drugs, and laboratory services- were $133 per resident day in
Ohio and $132 in Washington, compared to $102 in Mississippi, even after
controlling for differences in the mix of residents and in area wages (see
fig. 1). Although total spending per resident day in Ohio and Washington was
similar, spending across expenditure categories differed somewhat. Capital
spending per resident day averaged $22 in Washington, while in Ohio it was
$13. On average, spending on nursing was $55 per resident day in Ohio,
compared with $48 in Washington. 11 But the amount spent on medical
supplies, other resident care, and home operations combined was almost the
same in Ohio and Washington, averaging $42 and $43 per resident day,
respectively.
11 Due to data limitations, we were unable to separate nursing- related
administrative expenditures from nursing expenditures in Ohio, which may
explain in part that state?s higher nursing costs per resident day.
6 GAO- 02- 431R Nursing Home Expenditures and Quality Figure 1: Average
Total Nursing Home Expenditures per Resident Day, by Expenditure Category,
1999
Note: Expenditures were adjusted to account for differences in the resource
needs of residents across homes and for differences in wages across
geographic areas. Due to data limitations, ancillary services, including
therapies, were excluded from this analysis. Source: GAO analysis of fiscal
year 1999 Medicaid nursing home cost report data from Mississippi, Ohio, and
Washington.
Although spending in Mississippi was lower than in the other two states, the
shares of spending devoted to the different expenditure categories were
similar to those in Ohio. Nursing expenditures in both states averaged about
two- fifths of total spending. Other resident care expenditures (such as
salaries and benefits for social workers and medical recordkeeping) were 13
percent in Mississippi and 14 percent in Ohio. The share of spending devoted
to capital was also similar, averaging 9 percent of total expenditures in
Mississippi and 10 percent in Ohio.
Within each of the three states, we found wide variation in spending across
nursing homes that was not explained by differences in area wages or the
care needs of residents. Washington nursing homes with the highest spending
levels had total expenditures per resident day that were, on average, 63
percent higher than the
0 20
40 60
80 100
120 140
Mississippi Ohio Washington
Capital Administrative Home operations Other resident care Medical supplies
Nursing
Dollars
7 GAO- 02- 431R Nursing Home Expenditures and Quality state?s lowest-
spending nursing homes (see fig. 2). 12 There was a similar difference
between the highest- spending and lowest- spending nursing homes in Ohio. In
contrast, the difference between the highest- and lowest- spending nursing
homes was much smaller in Mississippi (26 percent).
Figure 2: Average Total Nursing Home Expenditures per Resident Day for the
Lowest- and Highest- Spending Homes, 1999
0 20
40 60
80 100
120 140
160 180
Mississippi Ohio Washington
Lowest- spending homes Highest- spending homes
In dollars
Notes: Nursing homes with total expenditures per resident day that were in
the top 25 percent for a state were considered to be the highest- spending
homes. Homes with total expenditures per resident day that were in the
bottom 25 percent for a state were considered to be the lowest- spending
homes. Expenditures were adjusted to account for differences in the resource
needs of residents across homes and for differences in wages across
geographic areas. Due to data limitations, ancillary services, including
therapies, were excluded from this analysis. Source: GAO analysis of fiscal
year 1999 Medicaid nursing home cost report data from Mississippi, Ohio, and
Washington.
We also found that as total nursing home expenditures per resident day
increased, the amount spent in almost every expenditure category also
increased, but not proportionally. Although the level of their spending on
nursing was higher, the highest- spending homes in Ohio and Washington
devoted a smaller share of their total spending to nursing compared with the
lowest- spending homes in each state. In both those states, the share of
total expenditures devoted to capital increased as total
12 Nursing homes with total expenditures per resident day that were in the
top 25 percent for a state were considered to be the highest- spending
homes. Homes with total expenditures per resident day that were in the
bottom 25 percent for a state were considered to be the lowest- spending
homes.
8 GAO- 02- 431R Nursing Home Expenditures and Quality expenditures
increased. In Mississippi, the share of total expenditures devoted to
nursing was the same in the highest- spending homes and lowest- spending
homes,
although the highest- spending homes provided 25 percent more nursing hours
per resident day than did the lowest- spending homes.
Some of the variation in spending within states may be attributable to
Medicaid payment policies and their effects on homes with different
proportions of Medicaid residents. We found an inverse relationship between
spending and Medicaid share in Ohio and Washington, where average total
expenditures per resident day were lower in homes that had a high proportion
of Medicaid residents and higher in homes that had a low proportion of
Medicaid residents. In Ohio, homes with a high Medicaid share had total
expenditures per resident day that were 10 percent lower than lowMedicaid-
share homes, and in Washington, high- Medicaid- share homes had total
expenditures per resident day that were 13 percent lower (see fig. 3). In
Ohio, the total expenditure difference between high- Medicaid- share and
low- Medicaid- share homes was driven mostly by a difference in expenditures
for administration, while in Washington, the difference between the two
types of homes was due to differences in expenditures for capital,
administration, and nursing. No relationship between spending and Medicaid
share was found in Mississippi. Since the proportion of residents covered by
Medicaid was greater than 65 percent for almost all Mississippi nursing
homes, those homes may be uniformly influenced by Medicaid payment policies.
9 GAO- 02- 431R Nursing Home Expenditures and Quality Figure 3: Average
Total Nursing Home Expenditures per Resident Day for Homes with the Highest
and Lowest Medicaid Shares, 1999
0 20
40 60
80 100
120 140
160
Capital Administrative Home operations Other resident care Medical supplies
Nursing
In dollars High Medicaid Mississippi
Low Medicaid Mississippi High Medicaid Ohio
Low Medicaid Ohio High
Medicaid Washington Low Medicaid Washington
Notes: Homes in which the proportion of resident days paid by Medicaid was
in the top 25 percent for a state were considered to be high- Medicaid-
share homes. Homes in which the proportion of resident days paid by Medicaid
was in the bottom 25 percent for a state were considered to be low- Medicaid
share homes. Expenditures were adjusted to account for differences in the
resource needs of residents across
homes and for differences in wages across geographic areas. Due to data
limitations, ancillary services, including therapies, were excluded from
this analysis. Source: GAO analysis of fiscal year 1999 Medicaid nursing
home cost report data from Mississippi, Ohio, and Washington.
Medicaid payment policies influence spending by creating incentives for
homes to contain overall expenditures while encouraging spending on
resources that most directly affect resident care and well- being, like
nursing services (see table 1). Generally, the states we examined
established Medicaid payment rates prospectively based on a previous year?s
spending, so that homes were at risk for any spending that exceeded the
rate. For most spending categories, nursing homes were paid their costs up
to a certain limit. States encouraged nursing home spending on nursing care
and other resident care by applying higher limits or ceilings compared with
those applied to other spending categories. In addition, Mississippi and
Washington
10 GAO- 02- 431R Nursing Home Expenditures and Quality encouraged a minimum
level of spending on nursing care. Mississippi made additional payments to
homes with nursing expenditures above a certain level. In
Washington, payments to homes with low nursing expenses were established
based on a minimum level of nursing spending. At the end of the year, homes
that had not spent the minimum amount had to give back any unspent funds.
The Mississippi and Ohio Medicaid programs encouraged nursing homes to limit
their spending in certain areas by offering efficiency incentives, by which
homes received additional payments if they kept their home operations and
administrative expenditures below the daily maximum allowable Medicaid
payment amounts. Mississippi and Ohio also employed mechanisms to limit
capital payments, either by not tying payment directly to a home?s spending
or by setting maximum payment rates and offering incentives to homes with
lower expenditures. Washington?s reimbursement method for capital was
comparatively generous, which may help to explain higher average capital
expenditures in that state.
Table 1: Medicaid Payment Rules for Nursing Homes in Mississippi, Ohio, and
Washington, 1999
State Homes paid their incurred costs up to Mississippi Nursing 120% of
median costs for all homes a Other resident care 120% of median costs for
all homes a Medical supplies 120% of median costs for all homes a Home
operations 109% of median costs for similar homes b Administrative 109% of
median costs for similar homes b Ohio Nursing 124% of median costs for
similar homes a, c Other resident care 124% of median costs for similar
homes a, c Medical supplies No cap Home operations 112.5% of median costs
for similar homes d Administrative 112.5% of median costs for similar homes
d, e
Certain home office costs not capped
11 GAO- 02- 431R Nursing Home Expenditures and Quality Washington Nursing
115% of median costs for similar homes a, f Other resident care 115% of
median costs for similar homes a, f Medical supplies 115% of median costs
for similar homes a, f Home operations Costs associated with food services,
housekeeping, and laundry capped at 110% of median for similar homes; other
costs capped at median costs for similar homes f Administrative Median costs
for similar homes f Note: Payment rules for capital- related expenditures
are omitted from this table.
a Costs are adjusted to reflect differences in the resource needs of
residents across homes. b Median costs are calculated separately for large
and small homes. c Median costs are calculated separately for four
geographic areas. d Median costs are calculated separately for eight groups
of homes that are similar in number of beds and geographic location. e
Nursing- related administrative costs paid as nursing costs. f Median costs
calculated separately for urban and rural homes.
Source: GAO analysis of Medicaid payment rules for Mississippi, Ohio, and
Washington. Across states, spending per resident day varied more and was
higher for expenditure categories that were less constrained by Medicaid
policies. Though Ohio and Washington have similar total expenditures per
resident day, spending on nursing was higher and varied more widely in Ohio.
This higher variation is consistent with that state?s policy of establishing
separate limits for more subgroups of homes, thereby accounting for more
differences in spending across homes compared with Washington?s system.
NURSING HOURS- MORE
THAN EXPENDITURES- RELATED TO QUALITY OF CARE DEFICIENCIES
In the states we examined, nursing hours per resident day- especially
nurses? aide hours- were related to quality of care deficiencies, with homes
providing more nursing hours being less likely to have identified quality
problems than homes providing fewer nursing hours. We found no clear
relationship between a nursing home?s total spending and the frequency of
quality of care deficiencies identified on
12 GAO- 02- 431R Nursing Home Expenditures and Quality state surveys, or
between spending on nursing and quality of care deficiencies. In Washington,
homes with higher nursing expenditures per resident day had fewer
quality of care deficiencies. In Mississippi, the opposite was true, while
we found no relationship between them in Ohio.
We found in Washington and Ohio that nursing homes providing more nursing
hours per resident day were less likely than homes with fewer nursing hours
to be cited by state surveyors for having repeated deficiencies involving
actual harm or immediate jeopardy to residents. This was especially true for
nurses? aide hours. In Washington, 16 percent of homes with the lowest
number of nurses? aide hours per resident day were found to have serious
deficiencies in successive surveys, compared with 3 percent of homes with
the highest number of nurses? aide hours per resident day. 13 In
Ohio, among homes with the lowest number of nurses? aide hours per resident
day, almost 6 percent were found to have repeated serious deficiencies,
compared to about half that many among homes with the highest number of
nurses? aide hours per resident day. In Mississippi, however, homes with
higher nursing hours per resident day were more likely to have deficiencies.
The findings in Washington and Ohio echo those of some other studies, which
have shown that staffing is positively correlated with quality of care,
although stronger associations were found between registered nurses? hours
and quality than between nurses? aide hours and quality. 14 A HCFA study on
nursing home staffing found that, for each type of nursing staff, there is a
minimum threshold of staff hours to residents below which homes are at
substantial risk of increased quality problems. 15 Studies of nursing home
spending have not found a clear association between
spending and quality. 16 One study, using defined outcomes as measures of
quality, found that higher quality can be associated with lower costs. 17 We
found no consistent relationship between spending and quality deficiencies
in the three states we examined. This may be indicative of the complex
factors influencing quality in
13 Nursing homes with nurses? aide hours per resident day that were in the
top 25 percent for a state were considered to be the highest- aide- hour
homes. Homes with nurses? aide hours per resident day that were in the
bottom 25 percent for a state were considered to be the lowest- aide- hour
homes.
14 Joel W. Cohen and William D. Spector, ?The Effect of Medicaid
Reimbursement on Quality of Care in Nursing Homes,? Journal of Health
Economics 1996; 15: 23- 48; John A. Nyman, ?Improving the Quality of Nursing
Home Outcomes: Are Adequacy- or Incentive- Oriented Policies More
Effective?? Medical
Care 1988: 26( 12): 1158- 1171. 15 Health Care Financing Administration,
Report to Congress: Appropriateness of Minimum Nurse Staffing Ratios in
Nursing Homes, Summer 2000. 16 See for example Mark A. Davis, ?On Nursing
Home Quality: A Review and Analysis,? Medical Care Review 1991; 48( 2): 129-
166; John A. Nyman, ?The Effect of Competition on Nursing Home Expenditures
Under Prospective Reimbursement,? Health Services Research 1988; 23( 4):
555- 574; Donald F. Vitaliano and Mark Toren, ?Cost and Efficiency in
Nursing Homes: A Stochastic Frontier Approach,? Journal of Health Economics
1994; 13: 281- 300. 17 Dana B. Mukamel and William D. Spector, ?Nursing Home
Costs and Risk- Adjusted Outcome Measures of Quality,? Medical Care 2000;
38( 1): 78- 89.
13 GAO- 02- 431R Nursing Home Expenditures and Quality nursing homes or
because increased spending does not necessarily purchase more hours of care.
18 In Ohio, the presence of repeated serious deficiencies appeared
unrelated to homes? spending on nursing. 19 However, in Mississippi, homes
with lower spending on nursing were less likely to have been found deficient
in state surveys. This is consistent with our finding that homes in
Mississippi with lower staffing levels had fewer deficiencies. By contrast,
in Washington, we found that homes with the lowest expenditures per resident
day on nursing were more likely to have repeated serious deficiencies than
were homes with highest nursing expenditures per resident day. Of the
lowest- spending homes, 17 percent had repeated serious deficiencies,
compared with 7 percent of the highest- spending homes. Further, a larger
share of the highest- spending homes had no serious
deficiencies compared with lowest- spending homes. Homes with serious
deficiencies in successive surveys had lower average nursing expenditures (9
percent less) than did homes with no serious deficiencies in successive
surveys. But these homes with repeated serious deficiencies also had higher
capital and administrative expenditures, and as a result, had total spending
that was 9 percent higher than homes with no such deficiencies.
COMMENTS FROM EXTERNAL REVIEWERS We received comments on a draft of this
report from Medicaid officials in Mississippi, Ohio, and Washington. They
provided technical comments, which we incorporated as appropriate. The
representative from Ohio reported that our findings of no relationship
between nursing home spending and quality of care were consistent with
analyses conducted by the state.
We also received comments from two researchers who have done extensive
analyses in the area of nursing home quality. We incorporated their
technical comments as appropriate. One researcher commented on the
complexity of the relationship between quality and staffing, noting that
factors such as management, tenure and training, staff mix, retention, and
turnover of staff may affect both the quality and the cost of care. We do
not disagree with this observation. The other researcher was concerned about
the possible interpretations of our results and noted that increased
spending does not necessarily increase hours of nursing care, but that
increased hours would increase homes? costs, which homes might afford by
decreasing their non- nursing costs or by lowering their profits. We point
out that nursing homes could increase nursing hours, and not necessarily
costs, and that the homes with higher spending had disproportionately higher
spending on capital, home operations, and administrative expenses- not
nursing care. 18 For example, a home may pay higher wages or it may hire a
different (and more skilled) mix of personnel, which would increase a home?s
costs without raising care hours. 19 Reviews of 1999 Ohio nursing home cost
report data by that state?s Bureau of Long Term Facilities
also found no clear relationship between nursing home spending and the
frequency of quality of care deficiencies.
14 GAO- 02- 431R Nursing Home Expenditures and Quality - - - - - As agreed
with your office, unless you publicly announce the contents of this report
earlier, we plan no further distribution until 30 days from the report date.
At that time, we will send copies of this report to Medicaid officials in
Mississippi, Ohio, and Washington and to interested congressional
committees. In addition, the letter will be available at no charge on the
GAO Web site at http:// www. gao. gov. If you or your staff have any
questions, please call me at (202) 512- 7114 or Carol Carter, Assistant
Director, at (312) 220- 7711. Staff who made major contributions to this
letter include Christine DeMars, Dana Kelley, and Daniel Lee.
Laura A. Dummit Director, Health Care- Medicare Payment Issues Enclosure
Enclosure Enclosure 15 GAO- 02- 431R Nursing Home Expenditures and Quality
Scope and Methodology
To determine levels of spending in nursing homes and the factors influencing
that spending, we analyzed the 1999 Medicaid cost reports of nursing homes
in three states, Mississippi, Ohio, and Washington. The cost reports capture
nursing home expenditures associated with the care of all residents living
in Medicaid- certified nursing homes in the states. Because we wanted to
uniformly adjust nursing home spending for differences in the care needs of
residents, we selected 3 of the 18 states that use the Resource Utilization
Group (RUG) case- mix classification system in their Medicaid payment
systems, were geographically diverse, and could provide us with electronic
data in a timely fashion. 20 We examined the expenditures of freestanding
nursing homes only; homes that were part of another facility, such as an
acute care or rehabilitation hospital, were excluded from the analysis. 21
Medicaid nursing home reimbursement practices vary considerably across
states and, therefore, the results of our analyses cannot be generalized to
the rest of the country.
The three states? cost reports included slightly different categorizations
of costs. We aggregated the more detailed spending available in some of the
cost reports into six uniformly defined categories: nursing (salaries and
benefits for registered nurses, licensed practical nurses, and nurses?
aides), medical supplies, other resident care (such as salaries and benefits
for dieticians, social workers, and directors of nursing; and staff and
supplies for medical recordkeeping), home operations (such as staff and
supplies needed for housekeeping, food service, laundry, and maintenance),
administrative (such as administrative and clerical salaries and office
supplies), and capital (such as depreciation on buildings, equipment, and
furnishings; interest; leases; and rentals). Ancillary services, including
therapies, were not included in our analysis because spending for these
services was not uniformly reported on the state cost reports that we
examined.
We excluded nursing homes that had cost- reporting periods of less than 10
months or greater than 14 months. We also excluded homes that had aberrant
values for total expenditures per resident day and nursing hours. Finally,
we excluded homes with missing data. In total, we excluded 5 percent of
free- standing nursing homes in Washington and 5 percent of free- standing
homes in Ohio. In Mississippi, we excluded 31 percent of free- standing
nursing homes, largely because of missing data. 22 Our final sample sizes
were 105 homes for Mississippi, 826 homes for Ohio, and 232
homes for Washington. In Washington and Ohio, our final sample of homes did
not 20 This classification system sorts nursing home residents into groups
based on their clinical condition, functional status, and expected use of
certain services. Payments for each group are adjusted up or down to reflect
the level of resources needed to care for the average resident in the group,
relative to
the overall average cost. 21 Because of their affiliations with hospitals
and other health systems, the cost structures of facilitybased nursing homes
can differ substantially from those of freestanding homes. For this reason,
we excluded those homes from our analysis. 22 The most common missing
information was the staffing data. The voluntary nature of the special
staffing survey may have contributed to the number of homes with missing
data.
Enclosure Enclosure 16 GAO- 02- 431R Nursing Home Expenditures and Quality
differ from the population of free- standing homes in terms of number of
beds, ownership, and rural and urban location. In Mississippi, our final
sample of homes
included a slightly higher share of proprietary homes and more large homes
and fewer small homes, compared to the population of free- standing homes.
To compare spending across facilities, we adjusted nursing expenditures for
differences in resident complexity, as calculated by the states for each
nursing home using the RUG classification system. Nursing, other resident
care, and administrative expenditures were also adjusted for differences in
wages across geographic areas using the Medicare hospital wage index.
Our staffing analysis assessed the total number of registered nurse,
licensed practical nurse, and nurses? aide hours per resident day, as
reported on Ohio and Washington cost reports. 23 Staffing hours were not
reported on Mississippi cost reports. For that state, we used data from a
voluntary 1- month study (December 1999) of total staffing hours (all
nursing hours and nurses? aide hours combined). We extrapolated these data
to a 12- month period. 24 We measured quality using deficiency data reported
in CMS?s On- line Survey
Certification and Reporting system. 25 We examined deficiencies recorded in
the following areas: physical restraints, abuse, quality of life, dignity,
pressure sores, indwelling catheters, treatment of incontinence, nutrition,
dehydration, unnecessary drugs, antipsychotic drugs, and nursing staff. Good
quality homes were determined to be those that CMS considers to be ?in
substantial compliance? in the areas we examined. In Ohio, 56 percent of
homes fell into this category, compared with 43 percent and 24 percent of
homes in Mississippi and Washington, respectively.
Because of differences in the frequency of deficiencies across the states,
we used a relative measure to identify poor quality homes. We defined as
poor quality those homes that had G- level deficiencies or worse in
successive surveys. These are homes that were found in two consecutive
surveys to have deficiencies in the same area that caused actual harm,
potential for death or serious injury, or actual death or serious injury.
Five percent of Ohio homes met this definition of poor quality and 10
percent
23 Due to data limitations, we were unable to separate out spending on
rehabilitation aides from spending on nurses? aides in Washington nursing
homes. 24 We also examined Mississippi nursing home staffing data from
another point in time and determined that the December 1999 data were a
reasonable representation of the year.
25 These data record the findings of routine and follow- up state surveys to
assess compliance with federal standards. We used results from the most
recent survey for each home; generally, homes? most recent surveys were in
1999 or 1998. Deficiencies identified in the survey process are placed in 1
of 12 categories, labeled ?A? through ?L? depending on the extent of
resident harm (severity) and the number of residents adversely affected
(scope). The most dangerous category (L) is for a widespread deficiency that
causes actual or potential for death or serious injury to residents; the
least dangerous category (A) is for an isolated deficiency that poses no
actual harm and has potential only for minimal harm. Homes with deficiencies
that do not exceed the C level are considered in ?substantial compliance,?
and as such, providing an acceptable level of care.
Enclosure Enclosure 17 GAO- 02- 431R Nursing Home Expenditures and Quality
of Washington homes did. No homes in Mississippi met this definition, so we
chose a less stringent definition for that state: Poor quality was defined
as homes that had
four or more D- level deficiencies (deficiencies that have the potential for
more than minimal harm) or two or more G- level deficiencies in a single
survey. Fourteen percent of Mississippi nursing homes fell into this
category. (290009)
*** End of document. ***