Improving Outcomes
Clinical Highlights
Researchers have identified a prediction rule that may help clinicians who care for nursing home residents identify pneumonia without a chest x-ray. Another predictive model may help determine the most appropriate care by identifying nursing home residents at low risk of lower respiratory tract infection.
The Agency for Healthcare Research and Quality (AHRQ) sponsored the research.
Select to download the print version (PDF File, 597 KB). PDF Help.
Introduction / Methods / Findings and Discussion / For More Information
Lower respiratory tract infection (LRI),
which includes pneumonia, bronchitis,
and tracheobronchitis, is the leading
cause of mortality and hospitalization
in nursing home residents. Often, the
signs and symptoms of pneumonia, in
particular, are not apparent in elderly
patients, making diagnosis more
complicated. In addition, clinician
visits to these residents are few and
sporadic and radiological facilities are
not readily available. This results in
residents being transferred to the
hospital just for the purpose of x-ray
evaluation. Researchers in Missouri
have identified a prediction rule that
may help clinicians who care for
nursing home residents identify
pneumonia without a chest x-ray.
Once an LRI has been diagnosed,
clinicians can then determine the most
appropriate care. For nursing home
residents, many of whom are
chronically ill, determining the severity
of the illness is a critical step in
deciding whether they should be
treated in the nursing home or in the
hospital. The Missouri researchers have
identified patient characteristics
predictive of 30-day mortality and used
them to distinguish residents at low risk
of dying from LRIs. Residents at low
risk of dying may be able to receive
nursing home care, thereby avoiding
the potential complications associated
with hospitalization.
The studies "Clinical Findings Associated with
Radiographic Pneumonia in Nursing Home
Residents" and "Predicting Mortality in Nursing
Home Residents with Lower Respiratory Tract
Infection" were conducted in nursing homes in
central Missouri and the St. Louis, MO, area.
Both studies were funded by the Agency for Healthcare
Research and Quality (AHRQ).
Additional funding was provided by the
Robert Wood Johnson Foundation Generalist
Physician Faculty Scholars and an Institutional
National Research Service Award from the
Health Resources and Services Administration. |
Between August 15, 1995, and
September 30, 1998, researchers in
Missouri identified participants from
36 nursing homes in central Missouri
and the St. Louis, MO, area. As part of
the Missouri LRI Project, the
researchers analyzed 2,334 episodes of
illness in 1,474 nursing home residents.
Those residents with signs or symptoms
of LRI whose chest x-rays showed
evidence of pneumonia were evaluated
to determine the pattern of symptoms
that would indicate a high risk of
pneumonia. Researchers developed a
simple scoring system (range = -1 to 8)
based on eight factors that
independently predicted pneumonia
without obtaining a chest x-ray:
- Increased pulse.
- Increased respiratory rate (30 or higher).
- Temperature of 38 degrees C or higher.
- Somnolence or decreased alertness.
- Presence of acute confusion.
- Lung crackles on auscultation.
- Absence of wheezing.
- Elevated white blood cell count (Table 1).
Table 1. Scoring System for Estimating 30-Day Mortality from Lower Respiratory Infection in Nursing Home Residents
Variable |
Value |
Points assigned |
Patient value |
BUN (mg/dL) |
< 16
> 16 and up to 27
> 27 and up to 38
> 38 and up to 49
> 49 and up to 60
> 60 and up to 71
> 71 |
0
1
2
3
4
5
6 |
|
WBC (109 cells/L) |
< 14
> 14 and up to 24
> 24 |
0
1
2 |
|
Absolute lymphocyte count1 (109 cells/L) |
> .8
< .8 |
0
1 |
|
Pulse (beats/minute) |
< 72
> 72 and up to 102
> 102 and up to 132
> 132 |
0
1
2
3 |
|
Gender |
Female
Male |
0
1 |
|
Body mass index (kg/m2)2 |
> 31
> 25 and up to 31
> 19 and up to 25
> 13 and up to 19
< 13 |
0
1
2
3
4 |
|
ADL3 |
0
1 or 2
3 or 4 |
0
1
2 |
|
Mood deterioration over last 90 days |
No
Yes |
0
2 |
|
Sum of item scores for patient.
To derive risk score, sum the assigned points. Risk of 30-day mortality is as follows:
1-4 points, 2.4%; 5-6 points, 6.9%; 7-8 points, 15.6%; 9-10 points, 34.5%; 11-17 points, 61.6%. |
|
1To calculate absolute lymphocyte count, multiply WBC by percent lymphocytes. For an individual with a WBC of 8 x 109/L and percent lymphocytes of 15, (8 x 109/L x .15 = 1.2 x 109/L. This value would receive 0 points.
2BMl is calculated as (weight in kg) divided by (height in m)2. Divide weight in pounds by 2.2 to derive weight in kilograms (145 lb = 66 kg). Multiply height in inches by 2.54 to convert to cm, then divide by 100 to convert to meters (68 in = 1.727 m). BMI = 66/1.72722 = 22.1.
3ADL scoring is based on four ADL variables: grooming, toileting, locomotion, and eating. Each is assigned a 0 if the resident is independent, requires supervision, or requires limited assistance; a 1 is assigned if the resident requires extensive assistance or is totally dependent. The four scores are summed to derive an ADL score of 0 to 4 which is assigned points as shown above.
In a separate study of mortality risk due
to LRI conducted during the same time
period as the pneumonia study, these
same researchers evaluated 1,406
episodes of LRI in 1,044 residents of 36
nursing homes in central Missouri and
the St. Louis, MO, area. Study
participants were considered eligible if
they were age 60, in the facility at least
14 days, off an antibiotic in the last 7
days for a previous LRI, were expected
to live more than 30 days, did not have
AIDS, and if they and their physicians
and families agreed to participate.
Clinical evaluations were recorded for
several variables and compared for levels
of differentiation and mortality after 30
days.
Researchers developed a point
system for clinicians specific to nursing
home residents based on eight factors:
- Serum urea nitrogen.
- White blood cell count.
- Body mass index (BMI).
- Pulse rate.
- Activities of daily living (ADLs) status.
- Low total lymphocyte count of less than 800/µl.
- Male sex.
- Decline in mood over 90 days.
The scores derived from the point system are then
evaluated to determine the most
appropriate care for the resident. A
score of 0-4 indicates low risk for death
occurring 30 days after diagnosis of
LRI, and a score of 5-6, relatively low
risk (Table 2).
Table 2. Scoring System1 for Projecting Probability of Pneumonia
Factor |
Range |
Score |
White blood cells per mm3 |
< 10,000
10,000-14,999
> 15,000 |
0
1
2 |
Respiratory rate |
< 30 per minute
> 30 per minute |
0
1 |
Somnolence or decreased alertness |
Absent Present |
0
1 |
Wheezes |
Absent Present |
0
-1 |
Acute confusion |
Absent Present |
0
1 |
Temperature |
< 38°C
> 38°C |
0
1 |
Crackles |
Absent Present |
0
1 |
Pulse, beats per minute |
< 110
110-129
> 130 |
0
1
2 |
1Points for the appropriate range of each factor are summed to derive an overall score. Residents with a score of -1 or 0 had a 24.5% probability of radiographic pneumonia; those with a score of 1, a 37.7% probability; a score of 2, 44.4%; a score of 3, 55.6%; and a score of 4 or more, 69.4% probability of pneumonia.
In developing this new mortality
model, the researchers built on the
work of earlier research funded by
AHRQ and conducted by the Patient
Outcomes Research Team (PORT) on
Community-Acquired Pneumonia. The
PORT developed and validated the
Pneumonia Severity Index (PSI), which
uses broad categories to identify
pneumonia patients living in the
community who can be treated safely at
home. The LRI tool is more sensitive to
residents of nursing facilities, giving
more weight to variables such as ADLs,
mood decline, and markers of poor
nutritional status.
Inclusion criteria for both studies were
the same, with the exception that
residents in the mortality study had to
meet the study definition of an LRI,
whereas all those evaluated in the
pneumonia diagnosis study were
included to determine whether they
met the study definition of an LRI.
Researchers' findings confirmed that
pneumonia in nursing home residents
usually is associated with few
symptoms, but a simple clinical tool
can identify residents at high risk for
pneumonia without the use of a chest
x-ray. Among 2,334 episodes of illness
in 1,474 nursing home residents, 45
percent of chest x-ray reports suggested
possible or definite pneumonia. Yet in
80 percent of pneumonia episodes,
patients had three or fewer respiratory
or general symptoms. However, only 8
percent of patients had no respiratory
symptoms. The 33 percent of residents
scoring three or more points had more
than a 50-percent probability of
pneumonia, and the 24 percent of
residents who had a score of 2 points
had a 44-percent probability of
pneumonia.
Study limitations included
the caution that residents were from a
single State and the possibility of
incomplete or missed clinical findings.
The researchers note that if their
findings are confirmed, doctors should
consider treating residents at high risk
of pneumonia—those with a score of at
least two or three points on the scoring
system—without obtaining a chest x-ray.
For residents with a score of 1 or
less, doctors should obtain an x-ray as a
guide to treatment.
In the mortality risk study, researchers
developed a new predictive model for
30-day mortality risk in nursing home
residents with LRIs. Fifty-two percent
of residents had low (score of 0-4) or
relatively low (score of 5-6) predicted
30-day mortality, with 2.2 percent and
6.2 percent actual mortality,
respectively. Again, researchers
cautioned that the residents were from
a single State. In addition, they noted
limitations of some missing or
potentially misclassified data. Still, the
results are useful for identifying low-risk
residents. If confirmed in other States,
the prediction rule can be helpful to
clinicians and researchers in deciding
between nursing home care and
hospitalization for residents with LRI.
Study findings are presented in an
article in the November 2001 issue of
The Journal of Family Practice: Mehr
DR, Binder EF, Kruse RL, et al.
Clinical findings associated with
radiographic pneumonia in nursing
home residents. J Fam Pract 2001
Nov;50(11):931-7; and The Journal of
the American Medical Association: Mehr
DR, Binder EF, Kruse RL, et al.
Predicting mortality in nursing home
residents with lower respiratory
infection: The Missouri LRI Study.
JAMA 2001 Nov;286(19):2427-36.
AHRQ Publication No. 03-P014
Current as of June 2003
Internet Citation:
Nursing Home Residents with Lower Respiratory Tract Infection. June 2003. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/nursnglti.htm