 |
|
Patient and Hospital Characteristics Associated With Recommended Processes of Care for Elderly Patients Hospitalized With Pneumonia
Results From the Medicare Quality Indicator System Pneumonia Module
Jonathan M. Fine, MD;
Michael J. Fine, MD, MSc;
Deron Galusha, MS;
Marcia Petrillo, MA;
Thomas P. Meehan, MD, MPH
Arch Intern Med. 2002;162:827-833.
ABSTRACT
| |
Background Unexplained wide variability exists in the performance of key initial
processes of care associated with improved survival of elderly patients (those  65
years) hospitalized with pneumonia. The objective of this study was to assess
which patient and hospital characteristics are associated with performance
of these key initial processes of care for hospitalized elderly patients with
pneumonia.
Methods A retrospective cohort analysis was performed using data from the Medicare
Quality Indicator System Pneumonia Module for 14 069 patients 65 years
or older hospitalized with pneumonia throughout the United States. Associations
were calculated using multivariate logistic regression analysis between specific
patient and hospital characteristics and 2 processes of care associated with
improved 30-day survival: administration of antibiotics within 8 hours of
hospital arrival and blood culture collection within 24 hours of arrival.
Results Timely antibiotic administration was negatively associated with nonwhite
race (African American: odds ratio [OR], 0.71; 95% confidence interval [CI],
0.60-0.85; and other racial minorities: OR, 0.79; 95% CI, 0.68-0.92), major
hospital teaching status (OR, 0.79; 95% CI, 0.67-0.93), and larger hospital
size (250 beds vs <100 beds: OR, 0.68; 95% CI, 0.59-0.80). Timely blood
culture collection was positively associated with larger hospital size (OR,
1.61; 95% CI, 1.39-1.87). Performance of both processes of care were positively
associated with registered nurse–bed ratios of 1.25 or higher (for antibiotic
administration: OR, 1.23; 95% CI, 1.10-1.38; and for blood culture collection:
OR, 1.43; 95% CI, 1.26-1.61) and fever (for antibiotic administration: OR,
1.35; 95% CI, 1.23-1.49; and for blood culture collection: OR, 3.07; 95% CI,
2.81-3.34) and were negatively associated with hospital location in the South
(for antibiotic administration: OR, 0.77; 95% CI, 0.69-0.86; and for blood
culture collection: OR, 0.85; 95% CI, 0.77-0.93).
Conclusions Minority race, fever, nurse-bed ratio, hospital size and teaching status,
and southern location are among the major patient and hospital characteristics
associated, either negatively or positively, with the timeliness of performance
of initial antibiotic administration and blood culture collection for patients
hospitalized with pneumonia. Because performance of these processes of care
is associated with improved likelihood of survival, medical providers should
seek to eliminate the variations in care associated with these patient and
hospital characteristics. In addition, the impact of nurse staffing changes
on performance of key time-sensitive processes of care should be weighed carefully.
INTRODUCTION
PNEUMONIA IS the most frequent infectious cause of death and the sixth
leading overall cause of death in the United States.1
Because of the high mortality and corresponding high hospital admission rates
and costs of care for patients with pneumonia,2
pneumonia care is one of the targeted concerns of the Centers for Medicare
and Medicaid Services' (CMS) Health Care Quality Improvement Program (HCQIP).3 A major aim of HCQIP is to analyze patterns of medical
practice and outcomes to identify opportunities for improvement in the quality
of care.3
The strongest measure of quality of care is performance of processes
of care that have confirmed links to important medical outcomes. To date,
investigations conducted under HCQIP have identified 3 processes of care for
pneumonia independently associated with reduced 30-day mortality. Administration
of antibiotics within 8 hours of hospital arrival was associated with a 15%
lower 30-day mortality.4 Blood culture collection
within 24 hours of arrival was associated with a 10% lower 30-day mortality.4 Initial empirical treatment with a second- or third-generation
cephalosporin along with a macrolide or with a quinolone alone was associated
with improved survival compared with treatment with a third-generation cephalosporin
alone.5 Performance of all of these processes
of care varied considerably from state to state.6
Recognition of the factors associated with the performance of these
initial processes of care would aid the formulation of focused efforts to
improve the quality of care for patients with pneumonia. This study was undertaken
to identify patient and hospital factors associated with performance of antibiotic
administration within 8 hours of arrival and blood culture collection within
24 hours of arrival of elderly patients hospitalized with this illness. These
2 processes of care were specifically chosen for analysis because they conform
to reasonable clinical standards as designated by an expert national panel
when the data set was collected,4 are amenable
to organizational quality improvement efforts, and are among the quality indicators
HCQIP is championing in the Sixth Scope of Work.3
PARTICIPANTS AND METHODS
STUDY SAMPLE
As previously described,4 the study sample
was drawn from cases represented in the Medicare National Claims History File
with a principal discharge diagnosis of pneumonia (International
Classification of Diseases, Ninth Revision, Clinical Modification codes
480.0-480.9, 481.0, 482.0-482.9, 483.0-483.8, 485.0, 486.0, 487.0, and 507.0)
or a principal discharge diagnosis of respiratory failure (code 518.81) and
a secondary diagnosis of pneumonia. Using these diagnostic codes, approximately
500 cases were selected from discharges in each state, the District of Columbia,
and Puerto Rico from October 1, 1994, through September 30, 1995. From these
cases, copies of 25 561 medical records were collected. The diagnosis
of pneumonia was confirmed if an initial working diagnosis of pneumonia was
documented by a clinician and a chest radiograph taken within 48 hours of
hospital presentation was interpreted as consistent with pneumonia. A total
of 17 627 cases met these confirmation criteria. Confirmed cases were
excluded if they had any of the following exclusion criteria: aged younger
than 65 years, prior acute-care hospitalization within 10 days, infection
with the human immunodeficiency virus or diagnosed as having the acquired
immunodeficiency syndrome, history of organ transplantation, recent chemotherapy
or immunosuppressive treatment, transfer from another acute-care facility,
or death or discharge on the day of admission. If patients were hospitalized
with pneumonia more than once during the study period (n = 112), only the
first hospitalization was included. Cases in which mortality outcomes could
not be ascertained (n = 33) were also excluded. Following application of these
eligibility criteria, the final study population was 14 069 patients.
DATA COLLECTION
The medical records of the identified patients were forwarded by hospitals
to 1 of 2 Clinical Data Abstraction Centers. Data elements were abstracted
from the medical record by trained staff using an electronic data collection
instrument, and were merged with hospital claims data provided by CMS. As
previously described,4 the consistency of data
abstraction was confirmed by a second independent abstraction of a subset
of cases and calculation of  7 statistics
for confirmation and exclusion criteria (, 0.48-1.00), clinical characteristics
(, 0.61-1.00), and processes of care (, 0.61-0.83).
DATA ELEMENTS
Three categories of data elements were analyzed in this study: (1) patient
characteristics, (2) hospital characteristics, and (3) 2 initial processes
of care (antibiotic administration within 8 hours of arrival and blood culture
collection within 24 hours of arrival at the hospital). Patient characteristics
analyzed included demographic factors (age, race, sex, and residence in a
skilled nursing or intermediate-care facility), treatment with an antibiotic
within 48 hours before hospitalization, comorbid illnesses (cerebrovascular
disease, congestive heart failure, coronary artery disease, neoplastic disease
excluding skin cancer, liver disease, and renal disease), and physical examination
findings at presentation (abnormal mental status, temperature, heart rate,
respiratory rate, and systolic blood pressure). Initial laboratory values
and test results (arterial pH; arterial partial pressure of oxygen; levels
of blood urea nitrogen, sodium, and glucose; hematocrit; and pleural effusion)
were collected for inclusion in the Pneumonia Severity Index (PSI) calculation.8 The PSI was derived and validated as part of the Pneumonia
Patient Outcomes Research Team cohort study to identify patients with pneumonia
at low risk for mortality. Based on the clinical characteristics listed, patients
are assigned to 1 of 5 risk classes for 30-day mortality. In this study, no
patients were classified as risk class I because all patients were older than
50 years.
All of the patient characteristic variables were obtained from abstracted
medical records, except liver disease and neoplastic disease, which were derived
from a combination of abstracted medical record data and coded secondary diagnoses,
and renal disease, which was assessed from secondary diagnostic codes alone.
The times of performance of initial antibiotic administration and blood culture
collection and the times of hospital arrival were recorded to the nearest
minute. For analyses of the relationships between time of arrival and performance
of the processes of care, times of arrival were grouped into intervals corresponding
to the 3 traditional nursing shifts of 7 AM to 3 PM, 3 PM to 11 PM, and 11
PM to 7 AM. Information pertaining to whether patients were admitted through
the emergency department or directly to a ward as either an urgent or an elective
admission was collected from the Medicare Provider Analysis and Review file.
Hospital characteristics were collected primarily from the American
Hospital Association 1994 annual survey.9 The
characteristics examined were geographic location, staffed bed numbers, average
bed occupancy rates, number of emergency department visits per year, registered
nurse–bed ratio, number of nurses, and teaching status. Teaching status
was denoted as nonteaching if no medical school affiliation or residency program
existed, limited teaching if a medical school affiliation or residency program
was present, or major teaching if the institution was a member of the Council
of Teaching Hospitals. Hospital locations were grouped according to the Bureau
of Census regional areas (South, Northeast, West, and Midwest)10
and classified as metropolitan and nonmetropolitan according to their designation
on the American Hospital Association survey.
DATA ANALYSES
Bivariate associations were conducted between performance of antibiotic
administration within 8 hours of arrival and blood culture collection within
24 hours of arrival at the hospital and each patient and hospital characteristic.
Separate multivariate backward stepwise logistic regression models were then
constructed for each of the 2 processes of care as the dependent outcome and
using all of the patient and hospital characteristics that had bivariate associations
with P .10 as independent variables. To adjust
for the initial severity of illness, the demographic and clinical characteristics
that determine the PSI were added to the multivariable model. Odds ratios
(ORs) and 95% confidence intervals (CIs) relating to the performance of each
process of care were determined for each patient and hospital characteristic.
An OR greater than 1 indicates increased performance and a value less than
1 indicates decreased performance of the process of care relative to the reference
category.
Partial residual plots revealed that model assumptions were appropriate.11 The  2 goodness-of-fit statistic showed
adequate fit for the models for blood culture collection within 24 hours of
arrival (P = .28) and good fit for the model for
antibiotic administration within 8 hours of arrival (P
= .66). The areas under the receiver operating characteristic curve revealed
good model discrimination (0.62 for the antibiotic model and 0.72 for the
blood culture model).12 Associations between
PSI risk classes and performance of antibiotic delivery within 8 hours of
arrival and blood culture collection within 24 hours of arrival were analyzed
for trend using the Mantel-Haenszel 2 test. Calculations were
performed using computer software (Stata, version 4.0; Stata Corp, College
Station, Tex).
RESULTS
COHORT DESCRIPTION
The mean age of the study population was 79.4 years, and 23% were admitted
from a skilled nursing or intermediate-care facility. Of the study population,
58% had more than 1 comorbid illness and 61% had at least 1 laboratory abnormality
or pleural effusion. Fifty-seven percent of the patients were admitted through
an emergency department. Finally, 80% of the study population received antibiotics
within 8 hours of admission and 66% had blood cultures collected within 24
hours of arrival.
FACTORS ASSOCIATED WITH TIMELY ANTIBIOTIC ADMINISTRATION
Multivariable logistic regression revealed several patient characteristics
that were independently associated with antibiotic administration within 8
hours of hospital arrival (Table 1).
African Americans and other minorities were less likely to receive initial
antibiotic therapy within the first 8 hours. Patients with cerebrovascular
disease were also less likely to receive antibiotics within 8 hours of hospital
arrival. Increasing heart rate, temperature, and respiratory rate were independently
associated with timely antibiotic administration (Table 1). Patients who had received antibiotics within 48 hours
before hospitalization were more likely to receive antibiotics in the hospital
within 8 hours of arrival (Table 1). The percentages of patients receiving antibiotics within 8 hours of hospital
arrival were similar across PSI risk classes: 81% for those in risk class
II, 79% for those in risk classes III and IV, and 78% for those in risk class
V.
|
|
|
|
Table 1. Multivariate Analysis of Patient Characteristics and Performance
of Initial Processes of Care*
|
|
|
Admission to a major teaching hospital was also independently associated
with less timely antibiotic administration. The administration of antibiotics
within 8 hours of hospital arrival was positively associated with higher registered
nurse–bed ratios. Patients were more likely to have antibiotics initiated
within 8 hours of presentation in hospitals with 1 to 1.24 nurses per bed
and with 1.25 nurses per bed or greater compared with patients in hospitals
with less than 0.75 nurse per bed. In contrast, patients admitted to hospitals
with more beds and higher occupancy rates had a decreased odds of receiving
antibiotic therapy within 8 hours of hospital arrival (Table 2).
|
|
|
|
Table 2. Multivariate Analysis of Hospital Characteristics and Performance
of Initial Processes of Care*
|
|
|
Hospital locations in the South, the Northeast, and metropolitan areas
were independently associated with less timely administration of initial antibiotics.
The percentage of patients receiving antibiotics within 8 hours of arrival
varied with the time of presentation to the hospital (Figure 1, A). Analysis of 2-hour intervals within a 24-hour cycle
revealed that performance of antibiotic administration within 8 hours of hospital
arrival varied from 72% to 85%. Compared with the 3 PM to 11 PM shift, arrival
at the hospital during the 7 AM to 3 PM and the 11 PM to 7 AM shifts was associated
with significantly reduced odds of receiving antibiotics within 8 hours of
arrival (Table 2). Most patients
with pneumonia arrived at the hospital during the daytime and especially around
midday (Figure 1, B).
|
|
|
|
A, Relationships between the time of arrival at the hospital in 2-hour
intervals and the percentage of patients arriving within each 2-hour interval
in whom antibiotics were administered within 8 hours of arrival and in whom
blood cultures were collected within 24 hours of arrival. B, Percentage of
all patients with pneumonia arriving at the hospital within each 2-hour interval
during the 24-hour period. All times are given in military form.
|
|
|
FACTORS ASSOCIATED WITH TIMELY BLOOD CULTURE COLLECTION
Admission from a skilled nursing or intermediate-care facility was positively
associated with blood culture collection within 24 hours of hospital arrival.
The only comorbid illness with an independent association with increased blood
culture collection was cerebrovascular disease (Table 1). The presence of abnormal mental status on presentation
increased the odds of blood culture collection. Abnormal vital signs were
also associated with an increased odds of blood culture collection, especially
fever. Patients who received antibiotics before hospital arrival had a reduced
odds of undergoing blood culture collection (Table 1). The percentages of patients undergoing blood culture collection
within 24 hours of hospital arrival increased according to PSI risk classes:
61% for those in risk class II, 62% for those in risk class III, 66% for those
in risk class IV, and 72% for those in risk class V (P<.001
for trend).
Patients treated in hospitals with a limited teaching status had a higher
odds of blood culture collection than those treated in hospitals designated
as nonteaching. A higher registered nurse–bed ratio and a larger hospital
size were independently associated with blood culture collection within 24
hours of arrival (Table 2). Direct
admission to the hospital floor, urgently and electively, was associated with
a decreased odds of blood culture collection. A higher annual emergency department
volume of visits was also independently associated with increased performance
of blood culture collection (Table 2).
Timely blood culture collection was less likely to occur in hospitals in the
South and more likely to occur in hospitals in the Northeast.
COMMENT
This analysis reveals a striking number of associations between patient
and hospital characteristics and performance of 2 important initial processes
of care for elderly patients hospitalized with pneumonia. A common thread
among many of these patient and hospital characteristics is that they may
influence the provision of timely care by affecting the ability of health
care providers to make prompt diagnostic and therapeutic decisions and to
execute them. Recognition of the positive and negative potential influences
of these characteristics may help efforts to improve the delivery of key hospital
services.
The ability to speedily reach a correct diagnosis and decide on treatment
is strongly influenced by patients' presenting signs and symptoms. As found
in this study, patients with certain abnormal vital signs consistent with
a more severe infection had increased odds of receiving antibiotics in a timely
fashion. Vital sign abnormalities, particularly fever, were also associated
with successfully obtaining blood cultures within 24 hours of hospital arrival.
In contrast, the diminished ability to communicate symptoms may account for
the reduced odds of receiving timely initial antibiotic therapy among patients
with cerebrovascular disease. Communication difficulties have been cited also
as a possible explanation for the lower quality of health care provided to
minority populations in the United States,13
as observed in this study.
A history of treatment with antibiotics within 48 hours before hospital
arrival was associated with an increased odds of timely antibiotic treatment.
This finding is not surprising given that prior treatment implies that the
diagnosis of a respiratory tract infection was already made and alleviates
delays incurred by diagnostic decision making at the hospital. Furthermore,
these pretreated patients had lessened odds of undergoing blood culture collection,
presumably because of the expected lower yield of blood cultures from patients
already receiving antibiotics.
Previous studies14-16
have shown that minority populations receive fewer medical services. This
study reveals that, for community-acquired pneumonia, they receive less timely
initial hospital care as well. The utility of race as a variable, however,
must be qualified by acknowledgment of the subjective method by which it is
defined in medical records.17 Its association
with other variables, such as socioeconomic status and cultural and societal
factors, not evaluated in this study may also contribute to its links with
specific outcomes.
Among the few characteristics examined that held significant positive
associations for timely antibiotic administration and blood culturing was
the number of registered nurses per bed. This finding buttresses concerns
about the impact of curtailed use of registered nurses on quality of patient
care.18-20 Higher
nurse staffing may facilitate patient triage, assessment, and treatment. In-depth
knowledge of each hospital's staffing arrangements, including use of nurse
surrogates, is necessary to determine where in the care of patients with pneumonia
the impact of higher nurse-bed ratios occurs.
Among the hospital characteristics examined, major teaching status as
indicated by membership in the Council of Teaching Hospitals was associated
with a lower likelihood of administering antibiotics in a timely fashion.
Affiliation with a medical school or the presence of a residency without membership
in the Council of Teaching Hospitals was associated with a better likelihood
of collecting blood cultures. These findings may be partially attributable
to a tendency, observed by 2 of us (J.M.F. and T.P.M.), of some emergency
medicine physicians to perform diagnostic tests but leave antibiotic selection
to the admitting resident physicians and, thereby, incur a delay in the initiation
of therapy. As previously noted, the differences in care between teaching
and nonteaching hospitals may reflect differences in organization and delivery
of care rather than exclusively the involvement of house staff.21
Admission to bigger and busier hospitals, as defined by number of beds
and higher bed occupancy rates, was negatively associated with timely antibiotic
delivery. For blood culture collection, almost the opposite was observed:
increased hospital size favored performance, and bed occupancy rate held no
significant association. This split between the odds of performance of diagnostic
and therapeutic processes of care as related to hospital size and occupancy
rate is attributable, no doubt, to numerous variables, such as drug retrieval
factors and time to transport patients to the ward for the initiation of antibiotic
therapy. In contrast to this observation of inconsistent quality of care associated
with large hospital size is the finding by Keeler et al22
that quality of care improves with increasing size among 5 disease states
analyzed, including pneumonia.
Patients arriving during the 3 PM to 11 PM shift were significantly
more likely to receive antibiotics in a timely fashion compared with those
arriving during the other shifts. One explanation may be that performance
of processes of care declines during times of peak clinical demand, as noted
with prescribing accuracy.23 Indeed, more patients
admitted with pneumonia arrived during the 7 AM to 3 PM interval than either
of the other 2 shifts. More factors must be involved, however, given that
during the 11 PM to 7 AM interval when the least number of patients with pneumonia
arrived, the odds of performance of antibiotic administration were also reduced.
Pursuit of this link between clinical workload and performance of processes
of care will require exact data on staffing, complete patient load, and disease
acuity that is beyond the scope of this study. Variation in performance of
processes of care with time of presentation may be due to patient factors
and not just hospital practices.
Variations in processes and outcomes of care according to hospital location
are well recognized in the medical literature.24-26
Metropolitan hospitals were less likely than nonmetropolitan hospitals to
administer initial antibiotics in a timely fashion. This finding is consistent
with the findings by Lave et al27 pertaining
to rural vs urban hospital variations regarding other pneumonia care practices.
Although the size of the data set and its selection by state sampling preclude
more detailed geographic analyses, the finding that performance of both processes
of care lags in the South suggests some systematic effect that warrants further
investigation.
The implications of our findings for those seeking to improve the quality
of care for hospitalized patients with pneumonia are multiple. Clinicians
should be aware that variations in the delivery of care may be attached to
specific patient characteristics and should explicitly examine their own practices
for evidence of such. Where suboptimal communication may be a factor, earlier
attainment of chest x-ray films to evaluate for pneumonia may speed diagnostic
decision making.
Inconsistencies in care associated with the presence of house staff
may be amenable to corrective actions. In the case of timely initial antibiotic
administration, for example, the first dose can be administered by physicians
stationed in the emergency department, the entry point for most hospitalized
patients with pneumonia, rather than wait for house staff to begin treatment.
The storage of antibiotics in the emergency department avoids the delays in
shipping from the pharmacy that larger hospitals especially may encounter.
For other aspects of staffing, such as registered nurse–bed ratios,
hospitals should recognize and track the impact of staffing changes on the
accomplishment of time-sensitive processes of care. Tracking patient flow
with attention to time of day of presentation, patient volume, and the corresponding
staff may also uncover explanations for lags in care. The decreased odds of
performance of blood cultures among patients bypassing the emergency department
and directly being admitted to the hospital wards may be ameliorated by the
institution of standing orders or clinical pathways that call for blood culture
collection in all patients with pneumonia.
Although this study was limited to the examination of the performance
of only 2 processes of care, the 2 chosen pertain to much of what hospitals
and clinicians provide early in the care of patients with pneumonia and, hence,
seem fairly reflective of the quality of initial care. Remeasurement of the
nationwide performance of these processes of care from 1997 to 1999 reveals
similar performance of antibiotic administration within 8 hours of hospital
arrival (median, 79%) and improved performance of blood culture collection
within 24 hours of hospital arrival (median, 82%) compared with this data
set. Changes in performance over time are to be expected given the increased
attention devoted to quality of care in general and these quality indicators
in particular.28 Indeed, CMS is promoting as
quality indicators antibiotic delivery within 8 hours of arrival and blood
culture collection (modified to blood culture collection before antibiotic
administration) in its HCQIP evaluation of pneumonia care among Medicare-insured
patients. These 2 processes of care join choice of proper empirical antibiotic
therapy and predischarge evaluation for pneumococcal and influenza vaccination
as CMS's pneumonia care quality indicators. The Joint Commission on Accreditation
of Healthcare Organizations,29 moreover, is
considering using these processes of care as core performance measures in
its assessment of hospitals. Health care workers seeking to improve care for
their patients with pneumonia and to meet CMS and Joint Commission on Accreditation
of Healthcare Organizations expectations may be well served by using the findings
of this study in their analysis of their hospital's performance of these processes
of care.
AUTHOR INFORMATION
Accepted for publication August 9, 2001.
This study was supported in part by a grant from the Polly Annenberg
Levee Charitable Trust, St Davids, Pa (Dr J. M. Fine).
This study was presented in abstract form at the Association for Health
Services Research 15th Annual Meeting, Washington, DC, June 22, 1998.
The analyses in this study were performed under contract 500-96-P549,
entitled "Utilization and Quality Control Peer Review Organization for the
State of Connecticut," sponsored by the CMS, US Department of Health and Human
Services. The content of this publication does not necessarily reflect the
views or policies of the US Department of Health and Human Services nor does
the mention of trade names, commercial products, or organizations imply endorsement
by the US government. The authors assume full responsibility for the accuracy
and completeness of the ideas represented. This article is a direct result
of the HCQIP initiated by the CMS, which has encouraged identification of
quality improvement projects derived from analysis of patterns of care and,
therefore, required no special funding on the part of this contractor.
Corresponding author: Jonathan M. Fine, MD, Section of Pulmonary
and Critical Care Medicine, Norwalk Hospital, 34 Maple St, Norwalk, CT 06856
(e-mail: jonathan.fine{at}norwalkhealth.org).
From the Section of Pulmonary and Critical Care Medicine, Norwalk Hospital,
Norwalk, Conn (Dr J. M. Fine); Qualidigm, Middletown, Conn (Drs J. M. Fine
and Meehan, Mr Galusha, and Ms Petrillo); the Division of General Internal
Medicine, Department of Medicine, and the Center for Research on Health Care,
University of Pittsburgh, Pittsburgh, Pa (Dr M. J. Fine); and the Department
of Medicine, Yale University School of Medicine, New Haven, Conn (Dr Meehan).
REFERENCES
| |
1. Centers for Diseases Control and Prevention. Pneumonia and influenza death rates: United States, 1979-1994. MMWR Morb Mortal Wkly Rep. 1995;44:535-537.
PUBMED
2. Niederman MS, McCombs JS, Unger AN, et al. The cost of treating community-acquired pneumonia. Clin Ther. 1998;20:820-837.
FULL TEXT
|
ISI
| PUBMED
3. Jencks SF, Wilensky GR. The health care quality improvement initiative: a new approach to quality
assurance in Medicare. JAMA. 1992;268:900-903.
FULL TEXT
|
ISI
| PUBMED
4. Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA. 1997;278:2080-2084.
ABSTRACT
5. Gleason PP, Meehan TP, Fine JM, et al. Associations between initial antimicrobial therapy and medical outcomes
for hospitalized elderly patients with pneumonia. Arch Intern Med. 1999;159:2562-2572.
FREE FULL TEXT
6. Connecticut Peer Review Organization. Medicare Quality Indicator System: Pneumonia Module
Final Report. Middletown: Connecticut Peer Review Organization; 1998.
7. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37-46.
FULL TEXT
|
ISI
8. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired
pneumonia. N Engl J Med. 1997;336:243-250.
FREE FULL TEXT
9. Annual Survey of Hospitals Data Base, 1994. Chicago, Ill: American Hospital Association; 1995.
10. A Guide to State and Local Census Geography. Princeton, NJ: Association of Public Data Users, US Dept of Commerce;
1993.
11. Landwehr J, Pregibon D, Shoemaker A. Graphical methods for assessing logistic regression models. J Am Stat Assoc. 1984;79:61-63.
12. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, NY: John Wiley & Sons Inc; 1989.
13. King TE, Brunetta P. Racial disparity in rates of surgery for lung cancer. N Engl J Med. 1999;341:1231-1233.
FREE FULL TEXT
14. Kahn KL, Pearson ML, Harrison ER, et al. Care for black and poor hospitalized Medicare patients. JAMA. 1994;271:1169-1174.
ABSTRACT
15. Buckle JM, Horn SD, Oates VM, Abbet H. Severity of illness and resource use differences among white and black
hospitalized elderly. Arch Intern Med. 1992;152:1596-1603.
ABSTRACT
16. Ford ES, Cooper RS. Racial/ethnic differences in health care utilization of cardiovascular
procedures: a review of the evidence. Health Serv Res. 1995;30:237-252.
ISI
| PUBMED
17. LaVeist TA. Beyond dummy variables and sample selection: what health services researchers
ought to know about race as a variable. Health Serv Res. 1994;29:1-15.
ISI
| PUBMED
18. Oermann MH, Huber D. Patient outcomes: a measure of nursing's value. Am J Nurs. 1999;99:40-47.
19. Blegen MA, Vaughn T. A multisite study of nurse staffing and patient occurrences. Nurs Econ. 1998;16:196-203.
ISI
| PUBMED
20. Implementing Nursing's Report Card: A Study of RN Staffing, Length
of Stay and Patient Outcomes. Washington, DC: American Nurses Association; 1997.
21. Rosenthal GE, Harper DL, Quinn LM, Cooper GS. Severity-adjusted mortality and length of stay in teaching and nonteaching
hospitals. JAMA. 1997;278:485-490.
ABSTRACT
22. Keeler EB, Rubinstein LV, Kahn KL, et al. Hospital characteristics and quality of care. JAMA. 1992;268:1709-1714.
ABSTRACT
23. Lesar TS, Lomaestro BM, Pohl H. Medication-prescribing errors in a teaching hospital. Arch Intern Med. 1997;157:1569-1576.
ABSTRACT
24. O'Connor GT, Quinton HB, Traven ND, et al. Geographic variation in the treatment of acute myocardial infarction. JAMA. 1999;281:627-633.
FREE FULL TEXT
25. Wennberg JE, Freeman JL, Culp WJ. Are hospital services rationed in New Haven or over-utilized in Boston? Lancet. 1987;1:1185-1189.
FULL TEXT
|
ISI
| PUBMED
26. Perrin JM, Homer CJ, Berwick DM, et al. Variations in rates of hospitalization of children in three urban communities. N Engl J Med. 1989;320:1183-1187.
ABSTRACT
27. Lave JR, Fine MJ, Sankey SS, et al. Hospitalized pneumonia: outcomes, treatment patterns, and costs in
urban and rural areas. J Gen Intern Med. 1996;11:415-421.
ISI
| PUBMED
28. Jencks SF, Cuerdon T, Burwen DR, et al. Quality of medical care delivered to Medicare beneficiaries: a profile
at state and national levels. JAMA. 2000;284:1670-1676.
FREE FULL TEXT
29. Joint Commisssion on Accreditation of Healthcare Organizations. Plan for introducing joint commission hospital care measure requirements. Available at: http://www.jcaho.org/oryx_frm.html. Accessed
January 22, 2002.
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES
|
Impact of antibiotic guideline compliance on duration of mechanical ventilation in critically ill patients with community-acquired pneumonia.
Shorr et al.
Chest 2006;130:93-100.
ABSTRACT
| FULL TEXT
Quality of care for patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease.
Lindenauer et al.
ANN INTERN MED 2006;144:894-903.
ABSTRACT
| FULL TEXT
Accuracy of Administrative Data for Identifying Patients With Pneumonia
Aronsky et al.
American Journal of Medical Quality 2005;20:319-328.
ABSTRACT
Guidelines for the Treatment of Community-acquired Pneumonia: Predictors of Adherence and Outcome
Menendez et al.
Am. J. Respir. Crit. Care Med. 2005;172:757-762.
ABSTRACT
| FULL TEXT
Understanding variation in quality of antibiotic use for community-acquired pneumonia: effect of patient, professional and hospital factors
Schouten et al.
J Antimicrob Chemother 2005;56:575-582.
ABSTRACT
| FULL TEXT
Is Pneumonia Still the Old Man's Friend?
Rozzini et al.
Arch Intern Med 2003;163:1491-1492.
FULL TEXT
Estimated frequency of nursing facility-acquired pneumonia?
Metersky et al.
Am. J. Respir. Crit. Care Med. 2003;167:1287-1288.
FULL TEXT
|
