Browsing by Author "zeray, Abrehet"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Impact of Clinical Data on Interpretation of Chest Radiographs
    (Addis Ababa University, 2015-10) zeray, Abrehet; Gisila, Azmera; Miller, Andrea
    BACKGROUND: The inclusion of clinical information may have unrecognized influence in the interpretation of diagnostic testing. OBJECTIVE: The objective of the study was to determine the impact of clinical history on chest radiograph interpretation. DESIGN: Prospective hospital based cross sectional study. METHODS: 10 residents, 4 second year and 6 third year residents evaluated a total of 100 frontal CXRs. Clinical information was withheld during the first interpretation. After one month the radiographs were reviewed with clinical information. The residents reported their finding using a standardized reporting tool containing different clinical parameters (i.e., pattern of infiltrate; alveolar infiltrate, interstitial infiltrate, air bronchograms, hilar adenopathy, pleural effusion etc.). SETTING: Tikur Anbesa teaching and referral Hospital RESULTS: The accuracy improved from fair (k ¼ 0.281, p-value 0.000) to moderate (k ¼0.418 , p- value 0.000) for identification of ‘‘hilar or mediastinal LAP’’, from moderate (k ¼ 0.557 , p- value -0.000 ) to substantial (k ¼ 0.705 , p-value – 0.000) for identification of pneumothorax , from fair (k ¼ 0.340 , p-value-0.000) to moderate (k ¼ 0.425 , p-value -0.000 ) for identification interstitial infiltrate with the addition of clinical history. The sensitivity increased from 23.5 % to 40%, from 44.4 % to 55.6 %, and from 46.2 to 51.3 % for hilar or mediastinal LAP, pneumothorax and interstitial infiltrate respectively with the addition of clinical history. Except for identification of the interstitial infiltrate which were found to be statistically significant, - 6 - for the rest two parameters there was an overlap of the 95 % CIs, suggesting that even though addition of clinical Hx resulted in higher sensitivity, it was not statistically significant. Although there was an increase in kappa values for Identification of ‘‘rib fracture ’’ and ‘‘pleural effusion,’’ and a decrease in the kappa value for identification of ‘‘cardiomegaly ’’, ‘‘alveolar infiltrate ’’, and “air bronchograms’’ with the addition of clinical information, there was substantial overlap of the 95% CIs, suggesting that inclusion of clinical history did not result in a statistically significant change in the reliability of these findings. CONCLUSIONS: The addition of clinical Hx resulted in improvement in the sensitivity of many of the clinical parameters, interstitial infiltrate being statistically significant. The knowledge of clinical Hx also has an impact in interpretation of CXRs for the junior residents as compared to senior residents. So residents should be advised to obtain clinical Hx whenever they read.