Chest radiography for general practitioners: Scope for change?

ClinicalRadiolo~gy(1992)46, 51-54

Chest Radiography for General Practitioners: Scope for Change? M. T. K E O G A N , A. R. P A D H A N I and C. D. R. F L O W E R

Department of Radiology, Addenbrooke ~" Hospital, Cambridge In order to derive guidelines for general practitioners on the value of chest radiography we prospectively reviewed all chest radiographs over a 10 month period. Radiographic chest examinations (frontal with or without a lateral view) were performed on 2017 patients in the radiology department or at the chest clinic. Patient details were noted, referring practice and waiting times monitored. The adequacy of the referral letter was assessed, and specific predetermined clinical findings were correlated with the radiographic findings. Results were divided into three groups (normal, abnormal with a clinically unrelated or chronic abnormality, and abnormal with a clinically related abnormality). O f these patients, 1245 (62%) had normal examinations and 460 (23%) had clinically related abnormalities. Abnormalities were rare in the 10-29 year age group, and when reassurance was sought for patients with symptoms of nonspecific chest pain or general ill health. Patients presenting with haemoptysis or with symptoms of lower respiratory infection, or of chronic airways disease often had an abnormal radiograph. A clinical diagnosis was indicated or implied in 1664 (82%) and confirmed in 876 (53%). The most frequent reason for referral was for reassurance (618 patients). Guidelines based on age and symptoms should rationalize the use of chest radiography by general practitioners. Keogan, M.T., Padhani, A.R. & Flower, C.D.R. (1992). Clinical Radiology 46, 51-54. Chest Radiography for General Practitioners: Scope for Change?

The value of direct access to certain radiological services for general practitioners is now widely accepted [1,2]. However, increasing demand for diagnostic radiological services has serious implications for the limited resources of radiology departments and a more selective use of available facilities is desirable [3]. There is also an important need to reduce the collective radiation exposure to the population as a whole, by reducing the number of unnecessary exposures to X-rays [4]. Guidelines have already been shown to promote these aims [5]. We aimed to produce guidelines for the use of chest radiographs by general practitioners, by first conducting a prospective analysis of all these examinations requested over a 10 month period. Both the indications for and the diagnostic yield from these examinations are presented here.

METHODS All chest radiographs requested by general practitioners over 10 months (December 1989 to September 1990) were assessed. Access to radiographic facilities is available either through the radiology department or via the hospital chest clinic 'X-ray only' service. In the radiology department both frontal and lateral views are obtained in most patients at presentation. A frontal view only is obtained for most follow-up examinations. At the chest clinic 100 mm radiographs are routinely used with conventional radiographs reserved for: patients with haemoptysis, those with positive Heaf tests, those obtained for immigration purposes, known occupational dust exposure, known or suspected cardiac abnormality Correspondenceto: Dr C. D. R. Flower, Department of Radiology, Addenbrooke's Hospital, Cambridge, CB2 2QQ.

or hypertension, for patients with previous abnormal films, to elucidate doubtful abnormalities (recall patients), and lastly in those with definite abnormalities. General practitioners are able to arrange radiographs in one of two ways: by a letter to the radiology department (by telephone if urgent), or by referral to the chest clinic X-ray unit, thus by-passing the consultative process. Requests to the chest clinic are made using a standard request form. This is given to the patient who makes an appointment by telephone; in this way referrals to the chest clinic may have been seen earlier. All examinations were reported by one of three radiologists (M.T.K., A.R.P., C.D.R.F.). Patient data obtained included age, sex, waiting time and clinical presentation. General practitioner details including practice location and number of partners were noted. The referral letter was scrutinized with regard to the clinical history (including a smoking history), and physical examination provided. The presumptive diagnosis (if indicated) was noted and correlated with the radiographic findings. The films were graded in three ways. First, normal. Second, abnormal with clinically unrelated or nonsignificant findings; these included such age-related features as ectatic great vessels or degenerative spinal disease, and clinically established chronic abnormalities such as chronic obstructive airways disease. The third grading, abnormal with clinically significant or related findings, depended both on the clinical information provided and on any abnormalities present on the radiograph. The waiting time between referral and examination dates was noted for the radiology department. The majority of patients referred to the chest clinic are seen within 1 week, though exactly comparable data are not available. The statistical test was the z2-test.

52

CLINICAL RADIOLOGY

100

100 -

(a)

80

80

60

60

4.0

~5

7~/

E o* 4O

~7 /, //

(D

11

Ix.

,

,.~

i

vA

I

ii

V/I

i

~5 E

//

40

/ / "

~

O-

20 "

20

0-9

40-49 60-69 >80 20-29 10-19 30-39 50-59 70-79

0-9

60-69 >80 20-29 40-49 10-19 30-39 50-59 70-79

Age group (years)

Age group (years)

Fig. 1 Proportion normal and proportion with clinically relevant abnormalities by age and sex. (a) Male. n, Normal (%); N, clinically relevant abnormality (%). (b) Female. D, Normal (%); @, related abnormality (%).

RESULTS In all, 2017 patients were examined, 593 in the radiology department and 1424 in the chest clinic. Age/Sex Distribution

The analysis with reference to age and sex is detailed in Fig. 1. One thousand and seventeen males and 1000 females were examined. The age distribution was similar: males (mean 52 years; SD 20 years) and females (mean 49 years; SD 20 years). More females were examined in the age group l0 29 years (sex ratio: 0.69); the converse was true for the age group 60 79 years (sex ratio: 1.30). A clinically relevant abnormality was noted in 460 (23%) patients overall. In general this likelihood increased with age, e.g. in the age group 70-79 years a clinical significant abnormality was noted in 132 (43%). Conversely in the age group 10 29 years a clinically significant abnormality was found in only 10 patients (3%). Abnormalities in Relation to Clinical Presentation

The frequency of significant abnormalities in relation to the principal presenting symptoms (as detailed in the referral letter) is presented in Fig. 2. Multiple symptoms were present in most patients and overlap was expected. A positive smoking history (present or previous smoker) was a strong predictor of abnormality in the presence of other symptoms, with a significant abnormality noted in 117 (33%) of 358 patients. The commonest abnormalities were those suggestive of chronic airways disease (n = 68), pulmonary consolidation (n = 13), cardiomegaly or cardiac failure (n = 19), pleural reaction ( n = 8 ) and neoplasm ( n = 12). Six patients had lung opacities of indeterminate origin. Multiple abnormalities were noted in many patients. The smoking habits of the group as a whole is unknown as this information is not routinely given with the referral letter. Symptoms of cough, wheeze, breathlessness, or a history suggestive of chronic airways disease were present in 932 patients. O f these examinations, 503 (54%) were normal and a significant abnormality was noted in 278

__1

Smoker

_ _ ]

Infection

Cough x \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ ' ~ J

Wheeze Chronic airways

disease Dyspnoea Haemoptysis Cardiovascular symptoms Hypertension -1 alone Non-specific chest pain Pre-employment • General ill-health I

0

I

100

I i 200 Total number I

I 300

i

I 400

Fig. 2 Presenting symptoms/clinical problem, n, Normal; N, clinically relevant abnormality.

(30%). The commonest abnormalities were those suggestive of chronic airways disease (n--- 114), cardiomegaly or cardiac failure (n=70), lung consolidation (n=35), pleural reaction (n = 31), lung tumour (n = 15), hilar mass or lymphadenopathy (n=10), and lung opacities of indeterminate origin (n = 8). Multiple abnormalities were again noted in m a n y patients. A clinically significant abnormality was present in 18 (33%) of 54 patients with a history of haemoptysis. Abnormalities noted were those of chronic airways disease (n--5), cardiomegaly and heart failure (n = 5), a mass or lobar collapse (n=4), a cavity (n=2), and bronchiectasis (n = 2). As above, multiple abnormalities were noted in m a n y cases. When hypertension was the only clinical indication for chest radiography, a clinically significant abnormality (usually cardiomegaly) was noted in only three (11%)

CHEST R A D I O G R A P H Y FOR

Table 1 - Non-specific chest pain

53

Waiting Times

Total

Normal or non-relevant abnormality

All patients

336

80%

20%

Non-specific chest pain as only s y m p t o m

19 l

92 %

8%

if age > 35 years Sex ratio (M : F)

30"/,, 0.95

Relevant abnormality

The majority of patients referred to the chest clinic were given appointments within 1 week of their request, while waiting times at the radiology department usually varied between 1 and 2 weeks.

Referral Patterns 100% 2.75

patients. This proportion increased to 15 patients (21%) when other symptoms were associated with hypertension. Analysis of those patients with chest pain not suggestive of a cardiac, pleural or bony origin (non-specific chest pain) is shown in Table 1. When non-specific chest pain was the main presenting symptom a clinically relevant abnormality was detected in 8% of patients all of w h o m were over 35 years and were more often male. The commonest abnormalities detected in this group were chronic airways disease (n=6), cardiomegaly or heart failure (n = 5); a neoplasm was detected in two patients. Analysis of those patients where reassurance was sought for symptoms of general ill health (fatigue, lassitude, lethargy, sweats or weight loss), in conjunction with other symptoms revealed a clinically relevant abnormality in one patient (3')'0). No abnormality was detected in 22 patients where there were no other symptoms.

Analysis of Referral Information Table 2 reviews the quality of the information available in referral letters and chest clinic forms. Only 772 patients (38%) overall had referrals that provided both an adequate clinical history and examination, and 276 (14%) patients were referred with no clinical information. An adequate referral was more likely to the radiology department than the chest clinic (66% vs 27%). A diagnosis was indicated or implied in a greater proportion of referrals to the radiology department than to the chest clinic.

Confirmation of Diagnosis A total of 33 different diagnoses were suggested. An analysis of those most frequently indicated is given in Table 3. Of referrals, 618 (31%) were for the purposes of reassurance, both for the patient and for the general practitioner. In 575 (93%) of these referrals the radiograph was normal or unchanged from previous radiographs.

Request Type Examinations were classified as follows: initial request (n = 1806), repeat request with an unhelpful initial film (n = 6), follow up of a previously detected abnormality (n = 173), and recall film, when a previous radiograph had shown a doubtful abnormality and the patient was recalled for further clarification. Of the patients recalled (n = 32) 27 had been examined at the chest clinic with 100 m m radiographs. Subsequent conventional films demonstrated a significant abnormality in 12 patients. The recall rate was higher (2%) for the chest clinic compared with the radiology department

(0.8%).

GPs

The number of patients referred to each department and the overall findings are given in Table 4. More patients were referred to the chest clinic and they were more likely to have normal radiographs. These patients were also less likely to have clinically related abnormalities. Table 5 outlines the most frequently presumed diagnosis in relation to their place of referral. Patients were more likely to be referred to the chest clinic for reassurance and for chronic airways disease. The radiology department was the more likely referral place for rib fractures and for suspected primary or secondary cancer.

Other Observations No significant abnormalities were demonstrated in 89 pre-employment radiographs obtained during the same period and 30 radiographs obtained for immigration purposes were normal. A total of 66 practices referred patients for chest radiographs. Forty-seven practices generated more than five examinations and accounted for 1946 (96%) of all radiographs requested. These 47 practices have 182 partners; the number of the first-time radiographs requested varied from 0.8 to 27.3 (mean 8.4) per partner. Some practices referred patients exclusively to one or other of the two departments and there was no discernible reason for this difference.

Discussion The percentage o f n o r m a l examinations was high (62%), and only 23% had a clinically relevant abnormality. These figures are in keeping with a previous report [6], and suggest the potential for a reduction in the number of examinations undertaken. The assessment of age/sex distribution indicates that a disproportionately large number of patients aged between 10 and 29 years are examined with a low abnormality rate, suggesting that chest radiography in this group is over-used. Patient selection may also be improved by attention to symptoms. Those who present with a lower respiratory tract infection, haemoptysis or those with evidence of obstructive airways disease are the most likely to have an abnormality; those with nonspecific symptoms of chest pain or general ill health are more likely to have normal radiographs. This analysis is in agreement with previous studies [6,7]. Adequate clinical information is accepted as important for the accurate interpretation of radiographs [8]. A smoking history, however, was provided in surprisingly few patients (24%), including those with haemoptysis (32%). The disparity in adequacy of clinical details provided to the two departments may in part be explained

54

CLINICAL RADIOLOGY

Table 2 - Referral information; history and physical examination

Adequate history alone Physical examination alone Adequate history and physical examination No history or physical examination Smoking status provided

Both departments (n = 2017)

Radiology department (n = 593)

Chest clinic (n = 1424)

911 58 772 276 483

185 (31%) 4 (1%) 390(66%) 14((3%) 137

726 (51%) 54 (4%) 382(27%) 262(18%) 346

(45%) (3%) (38%) ([4%)

P=<0.005 P=<0.005 P, NS

NS, Not significant.

Table 3 - Commonly indicated diagnosis/reason for examination

Diagnosis~reason

Indicated

Confirmed (%)

Reassurance Infection Chronic airway disease Heart failure Cardiomegaly Lung cancer Metastases Tuberculosis Rib fracture/swelling

618 489 94 80 72 89 18 20 24

575 (93) I 12 (23) 59 (63) 28 (35) 13 (18) 30 (34) 3 (17) 4 (20) l (4)

Table 4 - Comparison of referrals to the chest clinic and the radiology department

Total number Normal Related abnormality

Radiology department

Chest clinic

(%J

(%J

593 312 (53) 157 (26)

1424 933 (66) 303 (21)

Both (%) 2017 1245 (62) 460 (23)

P<0.005 P<0.05

Table 5 - Common diagnoses indicated/implied and their place of referral

Diagnosis

Chest clinic

Radiology department

Total number Reassurance Infection Chronic airway disease Asthma Cardiomegaly Cardiac failure Lung cancer Metastases Sarcoidosis Tuberculosis Rib fracture or swelling

1145 475 332 77 26 49 54 51 5 8 13 6

519 143 157 17 17 23 26 38 13 9 7 18

P < 0.005 P, NS P < 0.01 P, NS P, NS P, NS P=0.03 P < 0.0001 P = 0.03 P, NS P < 0.0001

NS, Not significant.

by the booking procedure at the chest clinic where the patient is in possession of the referral form. When no abnormality was anticipated, i.e. when reassurance was sought, usually for symptoms of nonspecific chest pain or for general ill health, the chest radiograph showed no significant abnormality in 93%. Previous studies [9] emphasized the value of a normal report to doctor and patient, but the cost in financial

terms and radiation dose call this practice into question. Lung cancer, heart failure and lower respiratory tract infection were the commonly confirmed diagnoses and represent conditions for which a chest radiograph is well justified. The discrepancy in the radiological patterns between the two departments m a y in part be accounted for by the easier access to the chest clinic. We noted that radiographs for reassurance were more likely to be referred to the chest clinic and these referrals were less likely to have clinically relevant abnormalities. The referral patterns between the two departments however was different for certain suspected diagnoses. Some of these findings are at variance with a previous report [7] which suggested that referrals would be more selective, and inappropriate referrals minimal, for a chest clinic-based radiography unit. Despite a slightly higher recall rate for the chest clinic, 100 m m radiographs seem an acceptable alternative to standard radiographs. In conclusion we suggest that there may be scope for a more selective use of chest radiography by general practitioners.

REFERENCES

1 Barton E, Gallagher S, Flower CDR, Hanka R, King RH, Sherwood T. Influence on patient management of general practitioner direct access to radiological services. British Journal of Radiology 1987;60:893 896. 2 Wright H J, Swinburne K, Inch J. The general practitioner's use of diagnostic radiology. Journal of the Royal Society ()f Medicine 1979;72:88-94. 3 Raison CA. Medical and legal aspects of the increasing demand for diagnostic radiology. Proceedings of the Royal Society of Medicine 1976;69:755-758. 4 National Radiological Protection Board. Patient Dose Reduction in Diagnostic Radiology. Documents of the National Radiation Protection Board 1990;1:3. 5 Meiring de vos P, Wells IP. The effect of radiology guidelines for general practitioners in Plymouth. Clinical Radiology 1990;42: 327 329. 6 Guyer PB, Chalmers AG. Chest radiography for general practitioners - a low yield investigation. Journal of the Royal College of General Practitioners 1983;33:477-479. 7 Simpson FG, Morrison JFJ, Cooke N J, Pearson SB. General practitioner referrals for static miniature chest radiography: indications and diagnostic yield. British Journal of Diseases q[" the Chest 1988;82:76 78. 8 Doubilet P, Herman PG. Interpretation of radiographs: effect of clinical history. American Journal of Radiology 1981;137:1055 1058. 9 Mills KA, Reilly PM. Laboratory and radiological investigations in general practice. British Medical Journal 1983;287:1033-1036.