White Paper on Validation of detection of Abnormality via. Spandan ECG

Nitin Chandola

Posted on - July 16, 2019

Eligibility of Subject studied -

  • The test subject was undefined and of arbitrary choice.
  • The test cases studied are 153, held at the Fortis hospital in Dehradun.
  • The test subject consent was taken prior to taking their test via SPANDAN ECG.

  • The objective of the study
  • To check whether the Spandan detects the Abnormality in Lead 2.
  • Eligibility of Spandan to perform with respect to 12 lead Gold standard ECG.
  • Validating the AHA standards to evaluate Abnormality detectionby Specificity and Sensitivity.

  • Introduction
    Abnormality detection is the important feature of Spandan ECG, Hence its validation study was conducted to validate the specificity and sensitivity of results produced by the Spandan ECG with respect to 12 lead Gold standard. AHA describes the specificity and sensitivity as the correct method to evaluate the accuracy of the results provided by any smartphone based ECG machine.
    The Spandan ECG and results were evaluated and tested with respect to the 12 lead gold standard ECG. The machine that is used here for validation of the results is GEMAC2000, which was used for testing in the health camp organized by Fortis Hospital and Sunfox Technologies in Dehradun.

    Methodology
    The evaluation study performed was under the protocol set by the Spandan technical team, which was unbiased and partial and involve the involvement of Cardiologist and a physician. The Interpretation of Abnormality by Spandan was taken as the characteristic attribute for the validation study. The study was aimed to achieve the minimum deviation in the lead 2 of Spandan ECG and GEMAC2000 Gold standard.
    The tests were performed in the subjects with consideration of the following:
  • The subject consent must be taken for the test.
  • There must be not the difference of five minutes between the test taken by Spandan ECG and GEMAC2000.
  • The cases with baseline wander are excluded
  • Cases with motion artifacts are excluded from results.
  • Cases with wrinkled skin are excluded.
  • People with chest pain are included in the study.
  • Normal cases are also included.
  • The test must be taken by 12 lead Gold standard at first.
  • The following are the combinations used as True test and False test

  • Spandan/12 leadGoldstandard Abnormal/NSR Abnormal/DEFECT Normal/NSR Normal/Defect
    Abnormal/NSR true true true false
    Abnormal/Defect true true true false
    Normal/NSR true true false true
    Normal/Defect false true true true

    Here are the steps to be followed by the researcher for the validation as shown in the flow chart:-


    Results and Discussions
    The test performed will be evaluated with a human error of 5% will be taken into consideration. The Deviation in the 153 subjects is shown in the figure below. Through which there is an accuracy of 98.5 percent in interpreting the Abnormality and normality in the cases.

    Total cases True Detection False detection
    153 150 3
    Accuracy 98.5% (With Human Error)



    The figure below shows the data of the tested subject’s normality. The ability to detect the normal and abnormality Via. chest lead 2 is near to 98.5%.The results were comprised of false positives, which was the vantage point for our work in designing the algorithms for Spandan ECG
    As per the AHA, there isnocomputerprogram that can provide the correct interpretation as a skillful physician, hence the challenge was to bring the accuracy as near to the 12 lead Gold standard ECG. The patients were divided into two groups.
    Group 1 – Patients with Known CVD or any Abnormality
    Group 2 – Patients with known Normality
    This consideration was taken due to the in-depth analysis of the accuracy of Spandan in delivering better and accurate results as the 12 lead Gold standard.

    12 Lead Gold standard Abnormal Cases Normal Cases Borderline cases
    Number of cases 91 54 8




    The accuracy is calculated by using the following formula:-

    % Accuracy = (Result via.12 lead Goldstandard-Result via Spandan)/(Results via 12 lead Goldstandard) x100


    The data with the following combinations of results were neglected and reclassified as normal or abnormal:-
  • Borderline ECG in Spandan and Borderline ECG in 12 lead Gold standard is True result.
  • Borderline ECG in Spandan and Abnormal ECG in 12 lead Gold standard is a true result.
  • Borderline ECG in Spandan and Normal ECG in 12 lead Gold standard is the true result.
  • Abnormal ECG in Spandan and Borderline ECG in 12 lead Gold standard is a true result.
  • Normal ECG in Spandan and Borderline ECG in 12 lead Gold standard is the true result.

  • Hence For Normality detection in Group 2 via. Spandan ECG refers to the following table, which includes cases with 54 normal subjects,and 4 cases of Borderline subjects as interpreted by 12 lead Gold standard.

    Case detected True Detection False Detection
    Number of cases 57 1
    Accuracy 98.3%


    For Abnormality, detection is Group 1 via. Spandan ECG refers to the following table, which includes the cases of 91 abnormal subjects, and 2 Borderline subjects as interpreted by 12 lead gold standard.

    Case detected True Detection False Detection
    Number of cases 90 3
    Accuracy 96.77%


    The Two Borderline ECG cases were 100% accurate in detection and Interpretation. Hence, the accuracy of the test overall was 98.5 %.

    Conclusion and Scope
  • The Spandan ECG is 98.5% accurate as compared to the 12 lead Gold standard (GEMAC2000) in Interpreting the similar results which are under permissible limits of +-5% error by medical standards.
  • The Spandan ECG meets the Glasgow measurement accuracy standards for QT Interval.
  • Spandan is suitable for detection of heart abnormalities as there are No False positives.
  • Can be used by Cardiac patients and Elderly people for monitoring and prevent detection.
  • Recommended for daily check-ups at home.
  • Recommended for Small clinics and Skilful Physicians.