- Email: [email protected]
Practice Patterns in the Treatment of Acutely III Hospitalized Asthmatic Patients at Three Teaching Hospitals
Practice Patterns in the Treatment of Acutely III Hospitalized Asthmatic Patients at Three Teaching Hospitals* Variability in Resource Utilization Jennifer Daley, M.D.;t Richard I. Kopelman, M.D.; Elaine Comeau, R.N., M.S.; Leo C. Ginns, M.D., F.C.C.E; and Th011UlS H. Rossing, M.D.
Study Objective: Our objective was to determine the extent to which patterns of diagnostic and therapeutic practice differ among hospitals caring for acutely ill hospitalized asthmatic patients in a single city. Design: Our study comprised a retrospective review of the records of patients admitted to the hospital for the treatment of acute asthma. Setting: Three large teaching hospitals in Boston were the setting. Patients: One hundred twenty-seven patients between 18 and 50 years of age who were admitted to the medical services specifically for the treatment of asthma were studied. Interventions: There were no interventions. Measurements and Main Results: For this group of patients with similar histories of asthma, clinical presentation, and severity of asthma, the diagnostic tests used within 12 hours of admission and the frequency and volume of diagnostic laboratory testing throughout the admission differed significantly among the three hospitals. Spirometry, the test
.N pressure on the medical community to provide
cost-effective medically appropriate care in an era of constrained resources has mounted in recent years, examining variations in medical practice has received increasing attention. Wennberg l has advocated examining the contents of the "black box" of medical care and revealing those medical practices based on scientific evidence and those based largely on subjective "practice style factors" of physicians. Using population-based small-area analysis, Wennberg and his colleaguess" have described wide variations in *From the Division of General Medicine, Department of Medicine (Drs Daley and Kopelman), New England Medical Center (Ms Comeau), Tufts University School of Medicine; the Center for Cost-Effective Care and the Department of Medicine, Brigham and Women's Hospital, Boston (Dr Rossing); and the Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston (Dr Ginns). tPresently affiliated with the Division of General Medicine and Primary Care, Department of Medicine, Beth Israel Hospital, Harvard Medical School, Boston. Supported by grants from the Hartford Foundation to New England Medical Center and Brigham and Women's Hospital and from the Blue CrossIMassachusetts Hospital Fund for Cooperative Innovation to all three hospitals. Manuscript received August 16; revision accepted November 30.
bearing most directly on the severity of the asthmatic attack, was not used routinely as a criterion for admission or discharge at any of the hospitals. Other tests of uncertain efficacy, such as chest x-ray Sims, were used frequently at some of the hospitals. Patients at all three hospitals were treated similarly with intensive combined regimens of methylxanthines, sympathomimetics, and corticosteroids and had similar mean lengths of stay. The use of chest physical therapy, which has not yet been demonstrated to be effective in acute asthma, differed signi6cantly among the three hospitals. Conclusions: We conclude that considerable variability exists in the diagnostic evaluation qi)acutely ill hospitalized asthmatic patients in the three hospitals; little variability exists in the pharmacologic treatment of these patients. In the absence of data on outcome regarding functional improvement and reductions in morbidity, we are unable to recommend a preferred pattern of practice from this study. (Chest 1991; 100:51-6)
the per capita rate of inpatient hospitalization for surgical and medical procedures in similar populations across the United States and Europe. Exploring explanations for these wide variations in per capita use of medical services continues to be an area of active research and controversy. Professional uncertainty about the appropriate diagnosis and management of similar clinical problems," the resulting supplier-induced demand," and the availability of physician manpower" have all been proposed to account for variations in medical practice. Little attention, however, has been placed on analysis of the variation in patterns of practice among hospitals for patients admitted for the same condition or procedure. In 1983, three hospitals in Boston initiated a collaborative effort to study the variations in clinical practice among the hospitals for patients admitted for similar illnesses or procedures. Procedures and conditions included in the study were inguinal herniorrhaphy, femoral-popliteal-tibial bypass graft, acute myocardial infarction, cholecystectom}', total hip replacement, upper gastrointestinal bleeding, appendectomy, laminectomy, and acute CHEST I 100 I 1 I JUL~ 1991
51
asthma. In this report, we present the results of the study to compare the patterns of practice in the treatment of acutely ill patients admitted to the three hospitals for acute exacerbations of asthma. Our hypothesis was that the assessment and treatment of these patients could vary among similar hospitals in a single cit)'. MATERIALS AND METHODS
The studied population was identified from patients discharged with a diagnosis of asthma between October 1, 1982, and June 30, 1984, from the medical services of three Boston teaching hospitals (New England Medical Center, Brigham and Women's Hospital, and Massachusetts General Hospital). These hospitals will not be identified further; they will be referred to as hospitals A, B, and C. Four hundred patients aged 18 to 50 years with a diagnosis of asthma on discharge were identified through retrospective review of computerized discharge abstract data and record review These patients had been classified into one of three diagnosis-related groups (DRGs): DRG 88 (chronic obstructive pulmonary disease); and diagnosis-related groups (DRGs) 96 and 97 (bronchitis and asthma). Some 282 charts (71 percent) were identified randomly for chart review and data analysis. Records were available for review in 170 (60 percent) of the 282 cases. Unavailable charts were not present in the medical record departments on multiple occasions, presumably because these patients made frequent visits to the hospitals' outpatient facilities or emergency rooms. Of the available charts, 43 were excluded after chart review; either because the patients were noted to have other diseases involving the lungs or pleura or because they were not admitted specifically for the treatment of asthma. By these criteria, the charts of 127 patients aged 18 to 50 years who were admitted to the three medical services for the treatment of acute asthma were identified for detailed chart review and data analysis.
RecordReview Detailed review of records was undertaken on these 127 patients to determine their demographic characteristics, clinical characteristics of asthma on admission, diagnostic evaluation (both on admission and subsequently), therapy prescribed (medication and other therapies such as chest physical therapy and intermittent positive-pressure breathing (IPPB) treabnents), and disposition at discharge. Data on the frequency of specific testing procedures per patient were gathered from the computerized billing data systems from each of the three hospitals and confirmed by detailed record review Statistical Analysis
The clinical comparability of the populations from the three hospitals was analyzed using one-way analysis of variance for continuous variables; categoric variables were compared using the Xl test with appropriate degrees of freedom. The Kruskal-Wallis test," a nonparametric analog of the analysis of variance, was used to test the hypothesis of no difference among the hospitals in the frequency of tests per patient. Values for p of greater than 0.05 were not statistically significant. RESULTS
Population The populations at the three hospitals were clinically similar, with few significant differences among the populations at each hospital (Table 1). The mean age in the studied population was 34 years; 89 (70 percent) 52
of the 127 patients were women. The only significant difference was in the number of patients who were employed, ie, 22 percent (10/45) at hospital A, 57 percent (27/47) at hospital B, and 23 percent (8135) at hospital C (p = 0.0035). Insufficient data were available in retrospective record review to determine if those unemployed patients were unable to work because of their pulmonary disease. Patients at each institution had similar histories of asthma-related contact with acute-care hospitals. There were no significant differences among the hospitals in patients with previous admissions for treatment of asthma (91 to 98 percent). Between 17 and 38 percent of the patients had been admitted for Table I-Clinical CharacteriBtic. at Presentation of Ibtients with Acute Aathma*
Data Demographics Mean age, yr (± SO) Male patients Asthma-related history Previous admission for asthma Admission for asthma 3 mo before index admission Emergency room visit for asthma in 2 wk prior to admission Clinical presentation at admission Current smoker Respirations > 301min Pulse> l00/min Cough Wheezing Shortness of breath Asthma-related medication on admission Theophylline preparation Inhaled steroid Oral steroid ~-Agonists*
Previous measurement of baseline FEV1 Baseline FEV 1 in prior 2 yr reported If measured, mean last value for FEV1,L
Hospital A (n = 45)
Hospital B (n=47)
Hospital C (n=35)
31±8 11 (24)
35±9 18 (38)
35±10 9 (26)
44 (98)
45 (96)
32 (91)
11 (24)
18 (38)
4 (11)
25 (56)
22 (47)
13 (37)
5 (11) 14 (31)
10 (21) 14 (30)
8 (23) 11 (31)
35 35 45 45
(78) (78) (100) (100)
28 43 47 47
(60) (91) (100) (100)
25 29 34 34
(71) (83) (97) (97)
38 (84)t
27 (57)t
25 (71)t
3 (7) 9 (20) 36 (SO)
5 (11) 13 (28) 28 (60)
6 (17) 5 (14) 22 (63)
16 (36)
15 (32)
12 (34)
1.29§
2.26§
1.69§
*Unless otherwise noted, data are numbers of patients; numbers within parentheses are percentages. tp =0.028. *~-Adrenergic agonists include terbutaline, albuterol inhaler, metaproterenol inhaler, and isoproterenol inhaler. §p=0.0013. Practice Patternsin Hospitalized AsthmaticPatIents (Daleyat 81)
the treatment of asthma in the three months prior to the index admission; 37 to 56 percent had been evaluated and treated for asthma in the emergency room in the two weeks prior to admission. These differences were not statistically significant. On initial evaluation for admission, there were no significant differences in clinical presentation. All of the patients (97 to 100 percent) complained of shortness of breath and were noted to be wheezing; the majority were coughing (78 to 91 percent). Approximately one third (30 to 31 percent) of the patients were tachypneic (respirations >30/min); 11 to 23 percent admitted to smoking cigarettes currently. Among the three hospitals, there was a marginally significant difference in the number of patients taking oral theophylline preparations on presentation for admission: 84 percent (38145) at hospital A; 57 percent (27/41)at hospital B; and 71 percent (25/35) at hospital C (p = 0.028). There were no significant differences in the use of inhaled or oral steroids or other asthmarelated medications on presentation among patients at the three hospitals. Use of inhaled steroids prior to admission was low at all three hospitals (7 to 17 percent), reHecting the low utilization of this class of medication at the time this study was conducted. One third of the patients at each hospital had had a measurement of FEV. obtained within the two years prior to admission recorded in their charts. Of those patients who had a baseline FEV1 recorded, there was a statistically significant difference in the mean FE VI (1.29 to 2.26 L) among the hospitals (p=0.0013); however, these data are largely uninterpretable, since the clinical circumstances under which spirometry was performed may have differed among the three hospitals. At the time of this study; peak expiratory How measurements were not used routinely in patients with asthma at any of the three hospitals in either hospitalized patients or outpatients.
Diagnostic Evaluation within 12 Hours ofArrival at Hospital
The initial diagnostic evaluation of the studied population was defined as all tests done within 12 hours of arrival at the hospital, including tests done in the emergency room prior to transfer to floor care. The number of patients receiving a range of diagnostic tests varied significantly among the hospitals (Table 2). The FEV. was measured in 47 percent (21/45) of the patients at hospital A vs 23 percent (11/41) and 26 percent (9135) at hospitals B and C, respectively (p = 0.038). A chest x-ray film was obtained in 89 percent (40/45) of the patients at hospital A vs 100 percent (47 and 35) at both hospitals Band C (p = 0.0046). There was no difference in the number ofpatients in whom a theophylline level was measured, but the level was significantly higher at hospital C (14.6 mVdl) than at hospital A (10.3 ml/dl) and hospital B (11.1 mVdl) (p=O.OOO1). Arterial blood gas levels were measured in 38 percent (11) of the patients at hospital A vs 70 percent (33) at hospital B and 80 percent (28) at hospital C (p = 0.00(1). An electrocardiogram was performed in 58 percent (26) of the patients at hospital A vs 89 percent (42) at hospital B and 83 percent (29) at hospital C (p = 0.0034). An eosinophil count was measured in none of the patients at hospital A and in one patient at hospital C vs 39 patients (83 percent) at hospital B (p = 0.00(1). There were no differences in the utilization of complete blood cell counts (CBCs), serum electrolyte levels, sputum Gram stain, culture and sensitivity, arterial lines, or cardiac monitors. Therapy for Asthma There were no significant differences in the number of patients receiving intravenous aminophylline, antibiotics, or steroids, or oral antibiotics, steroids, or (3adrenergic agonists (Table3).There were no significant
Table 2-Diagnostic Evaluation within 12 Hou,.,
ofArrival at Hospital·
Data
Hospital A (n =45)
Hospital B (n = 47)
Hospital C (n = 35)
CBC with differential Serum electrolytes, BUN, and blood glucose Eosinophil count Theophylline level Mean theophylline level, mVdl (±SD) Arterial blood gas levels Arterial blood gas analysis done in emergency room Mean Pco, on admission, mm Hg (±SD) Chest x-ray film Electrocardiogram Spirometry (FEV 1 measured during first 12 h of admission) FE VI measured in emergency room Mean FEV 1 on admission, L (±SD) Sputum culture and sensitivity Sputum Gram stain
45 (100) 43 (96) 0 44 (98) 10.3±5.9 17 (38) 15 (33) 4O±22 40 (89) 26 (58) 21 (47) 13 (29) 1.83±0.82 41 (91) 41 (91)
47 (100) 47 (100) 39 (83) 44 (94) 11.1±5.3 33 (70) 32 (68) 41± 17 47 (100) 42 (89) 11 (23) 11 (23) 1.44± 1.27 46 (98) 46 (98)
35 (100) 35 (100) 1 (3) 33 (94)
14.6±6.7 28 (SO) 28 (SO) 38± 13 35 (100) 29 (83) 9 (26) 8 (23) 1.20±0.49 31 (89) 30
(86)
p Value NS 0.0001 NS 0.007 0.0001 0.0002 NS 0.0046 0.0034 0.038 NS NS NS NS
*Unless otherwise noted, data are numbers of patients; numbers within parentheses are percentages. NS, not significant. CHEST I 100 I 1 I JUL'(, 1991
53
Table 3- Therapy Received by Patients in Population* Data Drug therapy Intravenous Aminophylline Antibiotic Steroid Oral Antibiotic Steroid ~-Adrener~c agonists Oral Injectable Inhalant Other therapy Chest physical therapy IPPB Supplemental oxygen
Hospital A (n=45)
Hospital B (11=47)
Hospital C (n=35)
44 (98) 15 (33)
43 (91) 11 (23)
36 (BO)
35
(74)
32 (91) 10 (29) 25 (71)
27 (60) 34 (76)
20 (43) 34 (72)
22 (63)
43 (96)
47 (100) (81) 47 (100)
33 (94)
32
28 (BO)t
37 (82) 45 (100) 15 (33)t 1 (2) 15 (33)
38
°
(68)t
24 (51)
17 (49)
21 (60) 33 (94)
°
17 (49)
*Data are numbers of patients; numbers within parentheses are percentages. tp = 0.003.
differences in the use of IPPB treatments or supplemental oxygen. There was a significant difference among the hospitals in the use of chest physical therapy, which consisted of percussion and deepbreathing exercises. One third (33 percent [15/45]) of the patients at hospital A received chest physical therapy vs 68 percent (32/47) and 80 percent (28/35) at hospitals B and C, respectively (p = 0.00(3). Length of Stay and Discharge Disposition
There were no significant differences in the mean length of stay (four to five days), the number of patients admitted to the intensive care unit (ICU) (one to four patients), or the mean length of ICU stay among the hospitals (three to four days). There were no deaths
among the 127 patients in the studied population; all were discharged to their home with arrangements for follow-up through a private physician or a hospital or community-based clinic. Frequency of Diagnostic Testing during Admission
There were significant differences among the hospitals in the frequency of diagnostic testing during admission for pulmonary-related tests, hematologic tests, and common serum chemistry tests (Table 4). Serum chemistry tests were usually ordered as a panel that included levels of serum electrolytes, blood urea nitrogen (BUN), and creatinine at each of the three hospitals. The mean number of treatments with chest physical therapy per patient was 4.51 at hospital C and 3.40 at hospital B vs 1.84 at hospital A (p<0.01). The mean number of chest x-ray films per patient was 1.34 and 1.23 at hospitals B and C, respectively, vs 0.73 at hospital A (p<0.01). The mean number of CBCs per patient was 2.53 and 3.11 at hospitals B and C, respectively, vs 1.40 at hospital A (p
In this study the patterns of diagnostic and therapeutic practice in the treatment of acutely ill asthmatic patients requiring hospitalization in three teaching hospitals in the same city were examined. For a group of 127 adult asthmatic subjects with similar asthma histories, clinical presentation, and severity of asthma, several interesting similarities and differences in diagnosis and treatment emerged. The diagnostic tests used within 12 hours of admission, particularly pulmonary tests such as spirometry, arterial blood gas levels, eosinophil counts, and chest x-ray film, differed
Table 4-Frequency afDiagnostic Testing during Admission Low Median, and High No. of Tests per Patient (Mean/Patient)
Tests Pulmonary-related Chest x-ray film Chest physical therapy (15-min treatment) Arterial blood gas levels Theophylline level Sputum culture Hematology CBC Prothrombin time Eosinophil count Serum chemistry BUN and creatinine Potassium Glucose
54
Hospital A (n=45) 0,1,4 0, 0, 16 0, 0, 18 0, 2, 13 0,0,6
(0.73) (1.84) (0.84) (2.89) (0.47)
Hospital B (n=47) 0, 1, 11 0, 0, 44 0,1,5 0, 4, 14 0, 1, 14
Hospital C (n =35)
p Value
(1.34) (3.40) (1.64) (4.06) (LBO)
0,1,2 0,2,37 0,1,6 0,4,17 0,2,8
(1.23) (4.51) (1.40) (3.97) (2.34)
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001
0,1,6 (1.40) 0,0,1 (0.11) 0,0,1 (0.07)
1, 2, 11 (2.53) 0,1,4 (1.04) 0,1,3 (1.19)
0,3,14 0,1,3 0,0,1
(3.11) (0.71) (0.06)
<0.03 <0.03 <0.03
0,1,8 (1.60) 0,2,8 (2.07) 0,1,8 (1.40)
0,2,7 (2.02) 1,2,7 (2.79) 0,1,5 (1.38)
0, 2.5, 12 (2.86) 1,3,14 (2.97) 0,1, 11 (2.14)
<0.01 <0.01 <0.01
Practice Patterns in Hospitalized Asthmatic PatIents (Daley at 8/)
significantly among the hospitals. Similarly, the frequency and volume of diagnostic laboratory testing throughout the admission differed among the hospitals. Despite these differences in approach to diagnosis, there were no significant differences in the pharmacologic therapy that the patients received or in the mean length of stay. This study has several considerable limitations. First, not all charts were available for retrospective review at any of the three hospitals. Second, given the moderate size of the original sample and extensive subsetting in the analysis, the power of this study to detect differences, if they exist, is diminished in the subset comparisons. The inability to find statistically significant differences among the hospitals in subsets may reflect only the small sample size and cannot be taken as evidence of similarity of the groups. Likewise, differences among the hospitals in the frequency of diagnostic tests per patient during admission could be discerned only on tests with sufficient volume to support the analysis. Third, this study does not address the issues of outcome for these asthmatic patients. Retrospectively, several critical questions regarding the efficacy of treatment could not be answered. To what extent did pulmonary function improve from admission to discharge? Did any of the patients require readmission to a hospital for treatment of asthma or complications of treatment within ten days of discharge? Did any of the patients experience an avoidable complication of their diagnostic evaluation or treatment in the hospital? To what extent was functional capacity improved as a result of treatment? Given the absence of measures of the effectiveness of the practice patterns described, no recommendation can be made with regard to a preferred efficient, effective means of treating acutely ill asthmatic patients requiring hospitalization. Several of the findings in the study deserve further comment and exploration. This study identified a group of adult asthmatic subjects already sufficiently ill to warrant admission to the hospital and observed their management in the emergency room and hospital retrospectively. No attempt was made to review the management of the patients' asthmatic attacks prior to their presentation to the emergency room. Further comparative study of the diagnosis and treatment of asthmatic patients as outpatients might reveal efficient, effective practice patterns that prevent hospitalization. Similarly, assessment of a patient's adherence to medical regimens prescribed in the office or clinic should be assessed as a factor in avoiding hospitalization; however, once severely ill enough to warrant hospitalization for asthma, these patients were treated aggressively with methylxanthines, sympathomimetics, and often corticosteroids at all three hospitals for the same mean length of stay.
The diagnostic evaluation of these acutely ill asthmatic patients on admission and during hospitalization merits discussion. In spite of the frequent ordering of other laboratory tests, simple spirometry was ordered infrequently. Thus, the laboratory test that bears the most direct relation to the severity of the asthmatic attack was not used routinely as a criterion for either admission or discharge at any of the hospitals. 10-15 The total eosinophil count, which has been shown to be useful in assessing the response of asthmatic subjects to treatment'"'" was also used infrequently in hospitals A and C, although it is possible that physicians in these hospitals derived similar information from the differential and complete white blood cell count. The variability in the extent of laboratory testing on admission among the hospitals may exist because there are virtually no data describing the yield of the routine use of these tests in the assessment of the acutely ill asthmatic patient. Similarly, given the variability in the volume of laboratory testing during hospitalization, no data demonstrate that repeated measurements of hematologic tests such as blood counts or levels of electrolytes, BUN, and creatinine contribute to the asthmatic patient's management. Most of the patients admitted to these three hospitals received a chest x-ray film on admission. The efficacy of chest x-ray films in the evaluation of acutely ill asthmatic patients has been reviewed in several studies. Chest x-ray films are rarely helpful in the outpatient management of acute asthma but are indicated if there is clinical evidence of pneumonia, pneumothorax, or a pulmonary or cardiac disorder that can be confused with asthma, such as congestive heart failure, 19 or if the patient's condition is unresponsive to bronchodilators. m In the absence of clinical indicators of pneumonia, fewer than 2 percent of the patients with acute asthmatic attacks treated in emergency rooms have chest roentgenographic abnormalities discovered that affect the course of treatment. m,21 The incidence of chest roentgenographic abnormalities that affect the course of treatment in patients with acute asthma severe enough to require admission to the hospital is unknown. As noted earlier, patients at all three hospitals were treated with intensive combined regimens of methylxanthines, sympathomimetics, and corticosteroids. A majority of patients at each hospital was treated intravenously with one or more of these medications. Although physicians have traditionally employed parenteral therapy in hospitalized asthmatic patients, data published since this study was conducted demonstrate that oral administration of aminophylline or corticosteroids is as effective as intravenous administration. 22 ,23 Oral sympathomimetic agents were prescribed frequently at all three hospitals at the time of this study, although these agents are now used less CHEST I 100 I 1 I JUL~ 1991
55
frequently, given current understanding of the role of inhaled sympathomimetics. Despite the fact that most "infectious exacerbations" of asthma are believed to be caused by non bacterial pathogens.t''-" antibiotics were frequently prescribed at each hospital, apparently without systematic attempts to establish the presence of a bacterial infection. Finally, the use of chest physical therapy varied among the hospitals. No data exist in the literature that document the effectiveness of this therapy in the treatment of acute exacerbations of asthma, despite its frequent use in the hospitals in the study. On the basis of this retrospective study of patients hospitalized with acute asthma at three hospitals in the same city, we conclude that significant differences existed in the diagnostic evaluation of acutely ill asthmatic patients. Spirometry, a sensitive test of reversible bronchospasm and a well-documented means of monitoring the effectiveness of therapy, was used relatively infrequently, while routine chest x-ray films were used in some of the hospitals, even in the absence of good evidence of their usefulness in the evaluation of acutely ill asthmatic patients. Few differences existed in the aggressive pharmacologic therapy that the patients received; in the absence of data on improvement in the functional status and on morbidity after hospitalization for these patients, we are unable to comment on the effectiveness of these therapies. For therapies of questionable effectiveness, such as chest physical therapy, used frequently in some of the hospitals, we speculate that variations in the use of these therapies may be related both to more visible service departments in the hospitals with high utilization and to physicians' patterns of practice. Based Oil this limited study in three academic health centers in the same city in the Northeast, we speculate that even greater variability could exist among hospitals in different geographic areas among community-based teaching and nonteaching hospitals. ACKNO\VLEDGMENTS: We thank Lance Peterson, R.N., EA.C., for chart ahstraction; Janet Andersen, M.S., for statistical analysis; and Dehorah Fallon. R.N., M.E H., for coordination of the study among the three hospitals.
4
5
6 7
8
9 10
11
12
13 14
15
16 17 18
19 20
21
22
REFERENCES \Vennberg J. Dealing with medical practice variations: a proposal for action. Health Alf 1984: 3:6-32 2 \Vennberg J, Gittlesohn A. Variations in medical care among small areas. Sci Am 1982; 246:120-64 3 \Vennberg J. Blowers L, Parker R, Gittelsohn AM. Changes in
56
23
24
tonsillectomy rates associated with feedback and review Pediatrics 1977; 59:821-26 McPherson K, Wennberg J, Hovind OB, Clifford E Small area variations in the use of common surgical procedures: an international comparison of New England, England, and Norway. N Engl J Med 1982; 307:1313-14 Wennberg J, Freeman JL, Culp WJ. Are hospital services rationed in New Haven or over-utilized in Boston? Lancet 1987; 8543:1185-88 Eddy D. Variations in physician practice: the role of uncertainty. Health Aff1984; 3:74-89 Wennberg J, Barnes B, ZubkoO: Professional uncertainty and the problem of supplier induced demand. Soc Sci Med 1982; 16:811-24 Roos LL. Supply, workload and utilization: a population-based analysis of surgery in rural Manitoba. Am J Public Health 1983; 73:414-21 Kruskal WH, Wallis WA. Use of ranks in one-criterion variance analysis. Ann Math Stat 1952; 47:525-40 Kelsen SG, Kelsen D~ Fleeger BF, Jones RC, Rodman T. Emergency room assessment and treatment of patients with acute asthma. Am J Med 1978; 64:622-28 Nowak RM, Gordon KR, Wroblewski DA, Tomlanovich MC, Kvale PAt Spirometric evaluation of acute bronchial asthma. J Am Coli Emerg Physicians 1979; 8:9-12 Martin TG, Elenbaas RM, Pingleton SH. Use of peak expiratory flow rates to eliminate unnecessary arterial blood gases in acute asthma. Ann Emerg Med 1982; 11:70-73 Fanta CH, Rossing TH, McFadden ER. Emergency room treatment of asthma. Am J Med 1982; 72:416-22 Nowak RM, Tomlanovich MC, Sarkar DD, Kvale PA, Anderson JA. Arterial blood gases and pulmonary function testing in acute bronchial asthma. JAMA 1983; 249:2043-46 Reed S, Diggle S, Cushley MJ, Sleet RA, Tattersfield AE. Assessment and management of asthma in an accident and emergency department. Thorax 1985; 40:897-902 Franklin W Treatment of severe asthma. N Engl J Med 1974; 290: 1469-72 Lowell FC. Clinical aspects of eosinophilia in atopic disease. JAMA 1967; 202:109-12 Horn DR, Robin ED, Theodore J, Van Kessel A. Total eosinophil counts in the management of bronchial asthma. N Engl J Med 1975;292:1152-55 Findlay LJ, Sahn SA. The value of roentgenograms in acute asthma in adults. Chest 1980; 80:535-36 Zieverink SE, Harper A~ Holden ~ Klatte EC, Brittain H. Emergency room radiography of asthma: an efficacy study. Radiology 1982; 145:27-29 Heckerling PS. The need for chest roentgenograms in adults with acute respiratory illness. Arch Intern Med 1986; 146:132124 Jonsson S, Kjartansson G, Gislason D, Helgason H. Comparison of the oral and intravenous routes for treating asthma with methylprednisolone and theophylline. Chest 1988; 94:723-26 Ratto D, Alfaro C, Sipsey J, Glovsky MM, Sharma OE Are intravenous corticosteroids required in status asthmaticus? JAMA 1988; 260:527-29 Bonner JR. The epidemiology and natural history of asthma. Clin Chest Med 1984; 5:557-65
Practice Patterns in Hospitalized Asthmatic Patients (Daley 8t aI)
Study Objective: Our objective was to determine the extent to which patterns of diagnostic and therapeutic practice differ among hospitals caring for acutely ill hospitalized asthmatic patients in a single city. Design: Our study comprised a retrospective review of the records of patients admitted to the hospital for the treatment of acute asthma. Setting: Three large teaching hospitals in Boston were the setting. Patients: One hundred twenty-seven patients between 18 and 50 years of age who were admitted to the medical services specifically for the treatment of asthma were studied. Interventions: There were no interventions. Measurements and Main Results: For this group of patients with similar histories of asthma, clinical presentation, and severity of asthma, the diagnostic tests used within 12 hours of admission and the frequency and volume of diagnostic laboratory testing throughout the admission differed significantly among the three hospitals. Spirometry, the test
.N pressure on the medical community to provide
cost-effective medically appropriate care in an era of constrained resources has mounted in recent years, examining variations in medical practice has received increasing attention. Wennberg l has advocated examining the contents of the "black box" of medical care and revealing those medical practices based on scientific evidence and those based largely on subjective "practice style factors" of physicians. Using population-based small-area analysis, Wennberg and his colleaguess" have described wide variations in *From the Division of General Medicine, Department of Medicine (Drs Daley and Kopelman), New England Medical Center (Ms Comeau), Tufts University School of Medicine; the Center for Cost-Effective Care and the Department of Medicine, Brigham and Women's Hospital, Boston (Dr Rossing); and the Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston (Dr Ginns). tPresently affiliated with the Division of General Medicine and Primary Care, Department of Medicine, Beth Israel Hospital, Harvard Medical School, Boston. Supported by grants from the Hartford Foundation to New England Medical Center and Brigham and Women's Hospital and from the Blue CrossIMassachusetts Hospital Fund for Cooperative Innovation to all three hospitals. Manuscript received August 16; revision accepted November 30.
bearing most directly on the severity of the asthmatic attack, was not used routinely as a criterion for admission or discharge at any of the hospitals. Other tests of uncertain efficacy, such as chest x-ray Sims, were used frequently at some of the hospitals. Patients at all three hospitals were treated similarly with intensive combined regimens of methylxanthines, sympathomimetics, and corticosteroids and had similar mean lengths of stay. The use of chest physical therapy, which has not yet been demonstrated to be effective in acute asthma, differed signi6cantly among the three hospitals. Conclusions: We conclude that considerable variability exists in the diagnostic evaluation qi)acutely ill hospitalized asthmatic patients in the three hospitals; little variability exists in the pharmacologic treatment of these patients. In the absence of data on outcome regarding functional improvement and reductions in morbidity, we are unable to recommend a preferred pattern of practice from this study. (Chest 1991; 100:51-6)
the per capita rate of inpatient hospitalization for surgical and medical procedures in similar populations across the United States and Europe. Exploring explanations for these wide variations in per capita use of medical services continues to be an area of active research and controversy. Professional uncertainty about the appropriate diagnosis and management of similar clinical problems," the resulting supplier-induced demand," and the availability of physician manpower" have all been proposed to account for variations in medical practice. Little attention, however, has been placed on analysis of the variation in patterns of practice among hospitals for patients admitted for the same condition or procedure. In 1983, three hospitals in Boston initiated a collaborative effort to study the variations in clinical practice among the hospitals for patients admitted for similar illnesses or procedures. Procedures and conditions included in the study were inguinal herniorrhaphy, femoral-popliteal-tibial bypass graft, acute myocardial infarction, cholecystectom}', total hip replacement, upper gastrointestinal bleeding, appendectomy, laminectomy, and acute CHEST I 100 I 1 I JUL~ 1991
51
asthma. In this report, we present the results of the study to compare the patterns of practice in the treatment of acutely ill patients admitted to the three hospitals for acute exacerbations of asthma. Our hypothesis was that the assessment and treatment of these patients could vary among similar hospitals in a single cit)'. MATERIALS AND METHODS
The studied population was identified from patients discharged with a diagnosis of asthma between October 1, 1982, and June 30, 1984, from the medical services of three Boston teaching hospitals (New England Medical Center, Brigham and Women's Hospital, and Massachusetts General Hospital). These hospitals will not be identified further; they will be referred to as hospitals A, B, and C. Four hundred patients aged 18 to 50 years with a diagnosis of asthma on discharge were identified through retrospective review of computerized discharge abstract data and record review These patients had been classified into one of three diagnosis-related groups (DRGs): DRG 88 (chronic obstructive pulmonary disease); and diagnosis-related groups (DRGs) 96 and 97 (bronchitis and asthma). Some 282 charts (71 percent) were identified randomly for chart review and data analysis. Records were available for review in 170 (60 percent) of the 282 cases. Unavailable charts were not present in the medical record departments on multiple occasions, presumably because these patients made frequent visits to the hospitals' outpatient facilities or emergency rooms. Of the available charts, 43 were excluded after chart review; either because the patients were noted to have other diseases involving the lungs or pleura or because they were not admitted specifically for the treatment of asthma. By these criteria, the charts of 127 patients aged 18 to 50 years who were admitted to the three medical services for the treatment of acute asthma were identified for detailed chart review and data analysis.
RecordReview Detailed review of records was undertaken on these 127 patients to determine their demographic characteristics, clinical characteristics of asthma on admission, diagnostic evaluation (both on admission and subsequently), therapy prescribed (medication and other therapies such as chest physical therapy and intermittent positive-pressure breathing (IPPB) treabnents), and disposition at discharge. Data on the frequency of specific testing procedures per patient were gathered from the computerized billing data systems from each of the three hospitals and confirmed by detailed record review Statistical Analysis
The clinical comparability of the populations from the three hospitals was analyzed using one-way analysis of variance for continuous variables; categoric variables were compared using the Xl test with appropriate degrees of freedom. The Kruskal-Wallis test," a nonparametric analog of the analysis of variance, was used to test the hypothesis of no difference among the hospitals in the frequency of tests per patient. Values for p of greater than 0.05 were not statistically significant. RESULTS
Population The populations at the three hospitals were clinically similar, with few significant differences among the populations at each hospital (Table 1). The mean age in the studied population was 34 years; 89 (70 percent) 52
of the 127 patients were women. The only significant difference was in the number of patients who were employed, ie, 22 percent (10/45) at hospital A, 57 percent (27/47) at hospital B, and 23 percent (8135) at hospital C (p = 0.0035). Insufficient data were available in retrospective record review to determine if those unemployed patients were unable to work because of their pulmonary disease. Patients at each institution had similar histories of asthma-related contact with acute-care hospitals. There were no significant differences among the hospitals in patients with previous admissions for treatment of asthma (91 to 98 percent). Between 17 and 38 percent of the patients had been admitted for Table I-Clinical CharacteriBtic. at Presentation of Ibtients with Acute Aathma*
Data Demographics Mean age, yr (± SO) Male patients Asthma-related history Previous admission for asthma Admission for asthma 3 mo before index admission Emergency room visit for asthma in 2 wk prior to admission Clinical presentation at admission Current smoker Respirations > 301min Pulse> l00/min Cough Wheezing Shortness of breath Asthma-related medication on admission Theophylline preparation Inhaled steroid Oral steroid ~-Agonists*
Previous measurement of baseline FEV1 Baseline FEV 1 in prior 2 yr reported If measured, mean last value for FEV1,L
Hospital A (n = 45)
Hospital B (n=47)
Hospital C (n=35)
31±8 11 (24)
35±9 18 (38)
35±10 9 (26)
44 (98)
45 (96)
32 (91)
11 (24)
18 (38)
4 (11)
25 (56)
22 (47)
13 (37)
5 (11) 14 (31)
10 (21) 14 (30)
8 (23) 11 (31)
35 35 45 45
(78) (78) (100) (100)
28 43 47 47
(60) (91) (100) (100)
25 29 34 34
(71) (83) (97) (97)
38 (84)t
27 (57)t
25 (71)t
3 (7) 9 (20) 36 (SO)
5 (11) 13 (28) 28 (60)
6 (17) 5 (14) 22 (63)
16 (36)
15 (32)
12 (34)
1.29§
2.26§
1.69§
*Unless otherwise noted, data are numbers of patients; numbers within parentheses are percentages. tp =0.028. *~-Adrenergic agonists include terbutaline, albuterol inhaler, metaproterenol inhaler, and isoproterenol inhaler. §p=0.0013. Practice Patternsin Hospitalized AsthmaticPatIents (Daleyat 81)
the treatment of asthma in the three months prior to the index admission; 37 to 56 percent had been evaluated and treated for asthma in the emergency room in the two weeks prior to admission. These differences were not statistically significant. On initial evaluation for admission, there were no significant differences in clinical presentation. All of the patients (97 to 100 percent) complained of shortness of breath and were noted to be wheezing; the majority were coughing (78 to 91 percent). Approximately one third (30 to 31 percent) of the patients were tachypneic (respirations >30/min); 11 to 23 percent admitted to smoking cigarettes currently. Among the three hospitals, there was a marginally significant difference in the number of patients taking oral theophylline preparations on presentation for admission: 84 percent (38145) at hospital A; 57 percent (27/41)at hospital B; and 71 percent (25/35) at hospital C (p = 0.028). There were no significant differences in the use of inhaled or oral steroids or other asthmarelated medications on presentation among patients at the three hospitals. Use of inhaled steroids prior to admission was low at all three hospitals (7 to 17 percent), reHecting the low utilization of this class of medication at the time this study was conducted. One third of the patients at each hospital had had a measurement of FEV. obtained within the two years prior to admission recorded in their charts. Of those patients who had a baseline FEV1 recorded, there was a statistically significant difference in the mean FE VI (1.29 to 2.26 L) among the hospitals (p=0.0013); however, these data are largely uninterpretable, since the clinical circumstances under which spirometry was performed may have differed among the three hospitals. At the time of this study; peak expiratory How measurements were not used routinely in patients with asthma at any of the three hospitals in either hospitalized patients or outpatients.
Diagnostic Evaluation within 12 Hours ofArrival at Hospital
The initial diagnostic evaluation of the studied population was defined as all tests done within 12 hours of arrival at the hospital, including tests done in the emergency room prior to transfer to floor care. The number of patients receiving a range of diagnostic tests varied significantly among the hospitals (Table 2). The FEV. was measured in 47 percent (21/45) of the patients at hospital A vs 23 percent (11/41) and 26 percent (9135) at hospitals B and C, respectively (p = 0.038). A chest x-ray film was obtained in 89 percent (40/45) of the patients at hospital A vs 100 percent (47 and 35) at both hospitals Band C (p = 0.0046). There was no difference in the number ofpatients in whom a theophylline level was measured, but the level was significantly higher at hospital C (14.6 mVdl) than at hospital A (10.3 ml/dl) and hospital B (11.1 mVdl) (p=O.OOO1). Arterial blood gas levels were measured in 38 percent (11) of the patients at hospital A vs 70 percent (33) at hospital B and 80 percent (28) at hospital C (p = 0.00(1). An electrocardiogram was performed in 58 percent (26) of the patients at hospital A vs 89 percent (42) at hospital B and 83 percent (29) at hospital C (p = 0.0034). An eosinophil count was measured in none of the patients at hospital A and in one patient at hospital C vs 39 patients (83 percent) at hospital B (p = 0.00(1). There were no differences in the utilization of complete blood cell counts (CBCs), serum electrolyte levels, sputum Gram stain, culture and sensitivity, arterial lines, or cardiac monitors. Therapy for Asthma There were no significant differences in the number of patients receiving intravenous aminophylline, antibiotics, or steroids, or oral antibiotics, steroids, or (3adrenergic agonists (Table3).There were no significant
Table 2-Diagnostic Evaluation within 12 Hou,.,
ofArrival at Hospital·
Data
Hospital A (n =45)
Hospital B (n = 47)
Hospital C (n = 35)
CBC with differential Serum electrolytes, BUN, and blood glucose Eosinophil count Theophylline level Mean theophylline level, mVdl (±SD) Arterial blood gas levels Arterial blood gas analysis done in emergency room Mean Pco, on admission, mm Hg (±SD) Chest x-ray film Electrocardiogram Spirometry (FEV 1 measured during first 12 h of admission) FE VI measured in emergency room Mean FEV 1 on admission, L (±SD) Sputum culture and sensitivity Sputum Gram stain
45 (100) 43 (96) 0 44 (98) 10.3±5.9 17 (38) 15 (33) 4O±22 40 (89) 26 (58) 21 (47) 13 (29) 1.83±0.82 41 (91) 41 (91)
47 (100) 47 (100) 39 (83) 44 (94) 11.1±5.3 33 (70) 32 (68) 41± 17 47 (100) 42 (89) 11 (23) 11 (23) 1.44± 1.27 46 (98) 46 (98)
35 (100) 35 (100) 1 (3) 33 (94)
14.6±6.7 28 (SO) 28 (SO) 38± 13 35 (100) 29 (83) 9 (26) 8 (23) 1.20±0.49 31 (89) 30
(86)
p Value NS 0.0001 NS 0.007 0.0001 0.0002 NS 0.0046 0.0034 0.038 NS NS NS NS
*Unless otherwise noted, data are numbers of patients; numbers within parentheses are percentages. NS, not significant. CHEST I 100 I 1 I JUL'(, 1991
53
Table 3- Therapy Received by Patients in Population* Data Drug therapy Intravenous Aminophylline Antibiotic Steroid Oral Antibiotic Steroid ~-Adrener~c agonists Oral Injectable Inhalant Other therapy Chest physical therapy IPPB Supplemental oxygen
Hospital A (n=45)
Hospital B (11=47)
Hospital C (n=35)
44 (98) 15 (33)
43 (91) 11 (23)
36 (BO)
35
(74)
32 (91) 10 (29) 25 (71)
27 (60) 34 (76)
20 (43) 34 (72)
22 (63)
43 (96)
47 (100) (81) 47 (100)
33 (94)
32
28 (BO)t
37 (82) 45 (100) 15 (33)t 1 (2) 15 (33)
38
°
(68)t
24 (51)
17 (49)
21 (60) 33 (94)
°
17 (49)
*Data are numbers of patients; numbers within parentheses are percentages. tp = 0.003.
differences in the use of IPPB treatments or supplemental oxygen. There was a significant difference among the hospitals in the use of chest physical therapy, which consisted of percussion and deepbreathing exercises. One third (33 percent [15/45]) of the patients at hospital A received chest physical therapy vs 68 percent (32/47) and 80 percent (28/35) at hospitals B and C, respectively (p = 0.00(3). Length of Stay and Discharge Disposition
There were no significant differences in the mean length of stay (four to five days), the number of patients admitted to the intensive care unit (ICU) (one to four patients), or the mean length of ICU stay among the hospitals (three to four days). There were no deaths
among the 127 patients in the studied population; all were discharged to their home with arrangements for follow-up through a private physician or a hospital or community-based clinic. Frequency of Diagnostic Testing during Admission
There were significant differences among the hospitals in the frequency of diagnostic testing during admission for pulmonary-related tests, hematologic tests, and common serum chemistry tests (Table 4). Serum chemistry tests were usually ordered as a panel that included levels of serum electrolytes, blood urea nitrogen (BUN), and creatinine at each of the three hospitals. The mean number of treatments with chest physical therapy per patient was 4.51 at hospital C and 3.40 at hospital B vs 1.84 at hospital A (p<0.01). The mean number of chest x-ray films per patient was 1.34 and 1.23 at hospitals B and C, respectively, vs 0.73 at hospital A (p<0.01). The mean number of CBCs per patient was 2.53 and 3.11 at hospitals B and C, respectively, vs 1.40 at hospital A (p
In this study the patterns of diagnostic and therapeutic practice in the treatment of acutely ill asthmatic patients requiring hospitalization in three teaching hospitals in the same city were examined. For a group of 127 adult asthmatic subjects with similar asthma histories, clinical presentation, and severity of asthma, several interesting similarities and differences in diagnosis and treatment emerged. The diagnostic tests used within 12 hours of admission, particularly pulmonary tests such as spirometry, arterial blood gas levels, eosinophil counts, and chest x-ray film, differed
Table 4-Frequency afDiagnostic Testing during Admission Low Median, and High No. of Tests per Patient (Mean/Patient)
Tests Pulmonary-related Chest x-ray film Chest physical therapy (15-min treatment) Arterial blood gas levels Theophylline level Sputum culture Hematology CBC Prothrombin time Eosinophil count Serum chemistry BUN and creatinine Potassium Glucose
54
Hospital A (n=45) 0,1,4 0, 0, 16 0, 0, 18 0, 2, 13 0,0,6
(0.73) (1.84) (0.84) (2.89) (0.47)
Hospital B (n=47) 0, 1, 11 0, 0, 44 0,1,5 0, 4, 14 0, 1, 14
Hospital C (n =35)
p Value
(1.34) (3.40) (1.64) (4.06) (LBO)
0,1,2 0,2,37 0,1,6 0,4,17 0,2,8
(1.23) (4.51) (1.40) (3.97) (2.34)
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001
0,1,6 (1.40) 0,0,1 (0.11) 0,0,1 (0.07)
1, 2, 11 (2.53) 0,1,4 (1.04) 0,1,3 (1.19)
0,3,14 0,1,3 0,0,1
(3.11) (0.71) (0.06)
<0.03 <0.03 <0.03
0,1,8 (1.60) 0,2,8 (2.07) 0,1,8 (1.40)
0,2,7 (2.02) 1,2,7 (2.79) 0,1,5 (1.38)
0, 2.5, 12 (2.86) 1,3,14 (2.97) 0,1, 11 (2.14)
<0.01 <0.01 <0.01
Practice Patterns in Hospitalized Asthmatic PatIents (Daley at 8/)
significantly among the hospitals. Similarly, the frequency and volume of diagnostic laboratory testing throughout the admission differed among the hospitals. Despite these differences in approach to diagnosis, there were no significant differences in the pharmacologic therapy that the patients received or in the mean length of stay. This study has several considerable limitations. First, not all charts were available for retrospective review at any of the three hospitals. Second, given the moderate size of the original sample and extensive subsetting in the analysis, the power of this study to detect differences, if they exist, is diminished in the subset comparisons. The inability to find statistically significant differences among the hospitals in subsets may reflect only the small sample size and cannot be taken as evidence of similarity of the groups. Likewise, differences among the hospitals in the frequency of diagnostic tests per patient during admission could be discerned only on tests with sufficient volume to support the analysis. Third, this study does not address the issues of outcome for these asthmatic patients. Retrospectively, several critical questions regarding the efficacy of treatment could not be answered. To what extent did pulmonary function improve from admission to discharge? Did any of the patients require readmission to a hospital for treatment of asthma or complications of treatment within ten days of discharge? Did any of the patients experience an avoidable complication of their diagnostic evaluation or treatment in the hospital? To what extent was functional capacity improved as a result of treatment? Given the absence of measures of the effectiveness of the practice patterns described, no recommendation can be made with regard to a preferred efficient, effective means of treating acutely ill asthmatic patients requiring hospitalization. Several of the findings in the study deserve further comment and exploration. This study identified a group of adult asthmatic subjects already sufficiently ill to warrant admission to the hospital and observed their management in the emergency room and hospital retrospectively. No attempt was made to review the management of the patients' asthmatic attacks prior to their presentation to the emergency room. Further comparative study of the diagnosis and treatment of asthmatic patients as outpatients might reveal efficient, effective practice patterns that prevent hospitalization. Similarly, assessment of a patient's adherence to medical regimens prescribed in the office or clinic should be assessed as a factor in avoiding hospitalization; however, once severely ill enough to warrant hospitalization for asthma, these patients were treated aggressively with methylxanthines, sympathomimetics, and often corticosteroids at all three hospitals for the same mean length of stay.
The diagnostic evaluation of these acutely ill asthmatic patients on admission and during hospitalization merits discussion. In spite of the frequent ordering of other laboratory tests, simple spirometry was ordered infrequently. Thus, the laboratory test that bears the most direct relation to the severity of the asthmatic attack was not used routinely as a criterion for either admission or discharge at any of the hospitals. 10-15 The total eosinophil count, which has been shown to be useful in assessing the response of asthmatic subjects to treatment'"'" was also used infrequently in hospitals A and C, although it is possible that physicians in these hospitals derived similar information from the differential and complete white blood cell count. The variability in the extent of laboratory testing on admission among the hospitals may exist because there are virtually no data describing the yield of the routine use of these tests in the assessment of the acutely ill asthmatic patient. Similarly, given the variability in the volume of laboratory testing during hospitalization, no data demonstrate that repeated measurements of hematologic tests such as blood counts or levels of electrolytes, BUN, and creatinine contribute to the asthmatic patient's management. Most of the patients admitted to these three hospitals received a chest x-ray film on admission. The efficacy of chest x-ray films in the evaluation of acutely ill asthmatic patients has been reviewed in several studies. Chest x-ray films are rarely helpful in the outpatient management of acute asthma but are indicated if there is clinical evidence of pneumonia, pneumothorax, or a pulmonary or cardiac disorder that can be confused with asthma, such as congestive heart failure, 19 or if the patient's condition is unresponsive to bronchodilators. m In the absence of clinical indicators of pneumonia, fewer than 2 percent of the patients with acute asthmatic attacks treated in emergency rooms have chest roentgenographic abnormalities discovered that affect the course of treatment. m,21 The incidence of chest roentgenographic abnormalities that affect the course of treatment in patients with acute asthma severe enough to require admission to the hospital is unknown. As noted earlier, patients at all three hospitals were treated with intensive combined regimens of methylxanthines, sympathomimetics, and corticosteroids. A majority of patients at each hospital was treated intravenously with one or more of these medications. Although physicians have traditionally employed parenteral therapy in hospitalized asthmatic patients, data published since this study was conducted demonstrate that oral administration of aminophylline or corticosteroids is as effective as intravenous administration. 22 ,23 Oral sympathomimetic agents were prescribed frequently at all three hospitals at the time of this study, although these agents are now used less CHEST I 100 I 1 I JUL~ 1991
55
frequently, given current understanding of the role of inhaled sympathomimetics. Despite the fact that most "infectious exacerbations" of asthma are believed to be caused by non bacterial pathogens.t''-" antibiotics were frequently prescribed at each hospital, apparently without systematic attempts to establish the presence of a bacterial infection. Finally, the use of chest physical therapy varied among the hospitals. No data exist in the literature that document the effectiveness of this therapy in the treatment of acute exacerbations of asthma, despite its frequent use in the hospitals in the study. On the basis of this retrospective study of patients hospitalized with acute asthma at three hospitals in the same city, we conclude that significant differences existed in the diagnostic evaluation of acutely ill asthmatic patients. Spirometry, a sensitive test of reversible bronchospasm and a well-documented means of monitoring the effectiveness of therapy, was used relatively infrequently, while routine chest x-ray films were used in some of the hospitals, even in the absence of good evidence of their usefulness in the evaluation of acutely ill asthmatic patients. Few differences existed in the aggressive pharmacologic therapy that the patients received; in the absence of data on improvement in the functional status and on morbidity after hospitalization for these patients, we are unable to comment on the effectiveness of these therapies. For therapies of questionable effectiveness, such as chest physical therapy, used frequently in some of the hospitals, we speculate that variations in the use of these therapies may be related both to more visible service departments in the hospitals with high utilization and to physicians' patterns of practice. Based Oil this limited study in three academic health centers in the same city in the Northeast, we speculate that even greater variability could exist among hospitals in different geographic areas among community-based teaching and nonteaching hospitals. ACKNO\VLEDGMENTS: We thank Lance Peterson, R.N., EA.C., for chart ahstraction; Janet Andersen, M.S., for statistical analysis; and Dehorah Fallon. R.N., M.E H., for coordination of the study among the three hospitals.
4
5
6 7
8
9 10
11
12
13 14
15
16 17 18
19 20
21
22
REFERENCES \Vennberg J. Dealing with medical practice variations: a proposal for action. Health Alf 1984: 3:6-32 2 \Vennberg J, Gittlesohn A. Variations in medical care among small areas. Sci Am 1982; 246:120-64 3 \Vennberg J. Blowers L, Parker R, Gittelsohn AM. Changes in
56
23
24
tonsillectomy rates associated with feedback and review Pediatrics 1977; 59:821-26 McPherson K, Wennberg J, Hovind OB, Clifford E Small area variations in the use of common surgical procedures: an international comparison of New England, England, and Norway. N Engl J Med 1982; 307:1313-14 Wennberg J, Freeman JL, Culp WJ. Are hospital services rationed in New Haven or over-utilized in Boston? Lancet 1987; 8543:1185-88 Eddy D. Variations in physician practice: the role of uncertainty. Health Aff1984; 3:74-89 Wennberg J, Barnes B, ZubkoO: Professional uncertainty and the problem of supplier induced demand. Soc Sci Med 1982; 16:811-24 Roos LL. Supply, workload and utilization: a population-based analysis of surgery in rural Manitoba. Am J Public Health 1983; 73:414-21 Kruskal WH, Wallis WA. Use of ranks in one-criterion variance analysis. Ann Math Stat 1952; 47:525-40 Kelsen SG, Kelsen D~ Fleeger BF, Jones RC, Rodman T. Emergency room assessment and treatment of patients with acute asthma. Am J Med 1978; 64:622-28 Nowak RM, Gordon KR, Wroblewski DA, Tomlanovich MC, Kvale PAt Spirometric evaluation of acute bronchial asthma. J Am Coli Emerg Physicians 1979; 8:9-12 Martin TG, Elenbaas RM, Pingleton SH. Use of peak expiratory flow rates to eliminate unnecessary arterial blood gases in acute asthma. Ann Emerg Med 1982; 11:70-73 Fanta CH, Rossing TH, McFadden ER. Emergency room treatment of asthma. Am J Med 1982; 72:416-22 Nowak RM, Tomlanovich MC, Sarkar DD, Kvale PA, Anderson JA. Arterial blood gases and pulmonary function testing in acute bronchial asthma. JAMA 1983; 249:2043-46 Reed S, Diggle S, Cushley MJ, Sleet RA, Tattersfield AE. Assessment and management of asthma in an accident and emergency department. Thorax 1985; 40:897-902 Franklin W Treatment of severe asthma. N Engl J Med 1974; 290: 1469-72 Lowell FC. Clinical aspects of eosinophilia in atopic disease. JAMA 1967; 202:109-12 Horn DR, Robin ED, Theodore J, Van Kessel A. Total eosinophil counts in the management of bronchial asthma. N Engl J Med 1975;292:1152-55 Findlay LJ, Sahn SA. The value of roentgenograms in acute asthma in adults. Chest 1980; 80:535-36 Zieverink SE, Harper A~ Holden ~ Klatte EC, Brittain H. Emergency room radiography of asthma: an efficacy study. Radiology 1982; 145:27-29 Heckerling PS. The need for chest roentgenograms in adults with acute respiratory illness. Arch Intern Med 1986; 146:132124 Jonsson S, Kjartansson G, Gislason D, Helgason H. Comparison of the oral and intravenous routes for treating asthma with methylprednisolone and theophylline. Chest 1988; 94:723-26 Ratto D, Alfaro C, Sipsey J, Glovsky MM, Sharma OE Are intravenous corticosteroids required in status asthmaticus? JAMA 1988; 260:527-29 Bonner JR. The epidemiology and natural history of asthma. Clin Chest Med 1984; 5:557-65
Practice Patterns in Hospitalized Asthmatic Patients (Daley 8t aI)