Survey of physicians concerning the use of chest radiography in the diagnosis of pneumonia in out-patients

Departments of Medicine and Microbiology, Dalhousie University and Victoria General Hospital, Halifax, Nova Scotia Correspondence and reprints: DR TM Marrie, Victorial General Hospital, Room 5014 ACC, 1278 Tower Road, Halifax, Nova Scotia B3H 2Y9. Telephone 902-428-5553, fax 902-428-7094, e-mail tmarrie@is.dal.ca Received for publication February 23, 1996. Accepted July 30, 1996 TJ MARRIE. Survey of physicians concerning the use of chest radiography in the diagnosis of pneumonia in out-patients. Can J Infect Dis 1997;8(2):95-98.

P neumonia is a common illness affecting 10 to 12 adults per 1000 per year (1).Up to 80% of cases of pneumonia are treated at home (2).However, most of our knowledge about pneumonia is derived from studies of patients with pneumonia who are admitted to hospital (3)(4)(5)(6)(7).Indeed we have no information about the diagnosis of pneumonia in those who are treated as out-patients, such as how many receive a chest radiograph.This is a central question because at the moment it is considered to be the gold standard for pneumonia diagnosis.The objectives of this study were to determine how physicians used chest radiographs in the diagnosis of out-patient pneumonia and to determine the factors that physicians consider when deciding to order a chest radiograph.

STUDY DESIGN
A convenience sample of physicians was selected from the Nova Scotia registry.Only internists and family doctors were selected, and physicians throughout the province were sampled proportional to the population.
A 35-item questionnaire was mailed to 176 physicians.The questionnaire inquired about physician demographics; experience with out-patients with pneumonia; and decision making regarding the ordering chest radiographs for these patients (a series of characteristics were presented and the physician was asked to rank the importance of each factor from 1 to 5 with 5 being most important).In addition two skill-testing questions were included.
One hundred and two (89%) had treated patients with pneumonia as out-patients in the previous three months.Twenty-five per cent of physicians treated seven or more patients with pneumonia during this time period.Fifty-seven (49%) ordered chest radiographs on 91% to 100% of patients in whom they made a clinical diagnosis of pneumonia (Table 1).Twenty-two of the 29 (76%) internists stated they ordered chest radiographs on 91% to 100% of patients in whom they suspected pneumonia compared with 33 of 83 (40%) family physicians (P<0.001).Time since graduation seemed to be an influence because 19 of 40 (47.5%)graduates from 1980 to present; 26 of 51 (51%) graduates from 1970 to 1979; and 19 of 25 (76%) of those who graduated before 1970 ordered chest radiographs on 91% to 100% of patients with pneumonia.The factors that physicians considered in ordering chest radiographs are presented in Table 2. Clinical appearance, respiratory distress and physical findings were the most important factors in this decision.
Forty per cent of physicians indicated that they ordered follow-up chest radiographs in 91% to 100% while 23% ordered these radiographs in only 0% to 10% of patients (Table 1).Indeed in this side-by-side comparison there was little difference in the proportions of physicians who ordered initial and follow-up chest radiographs.Thirty physicians (26.3%) said they ordered initial and follow-up chest radiographs on 100% of patient suspected of having pneumonia.The factors that they considered important in the decision to do a follow-up chest radiograph are given in Table 3.All 102 physicians who replied to a question about how often they had seen a patient with carcinoma of the lung who initially presentated with pneumonia had seen such a patient at least once.
The answers to the skill-testing questions are presented in Tables 4 and 5.

Number of physicians (%) who ordered follow-up chest radiographs
Jay et al ( 13) studied 72 patients with bacteremic pneumococcal pneumonia and found that consolidation disappeared in all patients by eight to 10 weeks.Age greater than 50 years, chronic obstructive pulmonary disease and alcoholism were associated with delayed resolution.These findings indicate that 90% of the physicians who answered the question in Table 4 underestimated the time for the 62-year-old male smoker with chronic obstructive lung disease to clear his pneumonia.Mittl and co-workers ( 14) studied 37 in-patients and 44 outpatients with pneumonia.The mean time to clear pneumonia in out-patients was 3.8 weeks versus 9.1 weeks for in-patients.Nonsmokers cleared the pneumonia in five weeks and smokers in 8.4 weeks.
Sixteen per cent of physicians overestimated the time for the 22-year-old female to clear her pneumonia in our study (Table 5).
It is evident from our observations that both patient and physician factors contribute to the decision of who receives a chest radiograph for the diagnosis of out-patient pneumonia.

TABLE 2 Factors that physicians (n=114) considered important in their decision to perform a chest radiograph on a patient with pneumonia
(11)clearly impractical for screening purposes.Studies of patients with Pneumocystis carinii pneumonia (PCP) indicate that for this diagnosis the chest radiograph may be a tarnished gold standard in that 39% of patients with PCP had a normal chest radiograph(10).Diehr et al(11)tried to predict pneumonia in out-patients with acute cough.They found that 48 of 1819 (2.6%) had *Each item was rated on a scale of 1 to 5 with 1 being least and 5 being most important; rating mean ± SD