MDPhD Features: EBM Cards
MDPhD transforms clinical abstracts into evidence based medicine cards. You will notice that the ebm cards are colour coded AND filled with data visualizations. The cards are designed for fast and easy consumption of key evidence based medicine insights. MDPhD is refreshed with the latest and most relevant ebm cards daily!
Colour coding is based on the methodology or design used in the study:
Clinical Practice (Reviews/Case Reports/Guidelines) is pink
Diagnostic Trial is red
Genetic Trial is purple
Meta-Analysis is blue
Observational Trial is yellow
Randomized Control Trial is green
Please note the data visualizations in the cards above. They include the relevant statistical outcomes. MDPhD graphs point estimates, confidence intervals and p-values when available to give the user a quick summary of the results. The user can cross reference the graphs with the text in the card to quickly identify how the predictor is related to the response variable. The graphs also allow the user to appreciate the magnitude of the result, the width of the confidence interval and how far away the result is from the null hypothesis (red baseline).
As you collect the ebm cards into your topic folders, quickly scan for clinically significant reproducible results in a variety of clinical settings and studies. This is often more informative than just analysis-paralysis of a single study. Focus on the big picture as you combine clinical trials with your clinical experience, expert opinion, physiological and biological mechanisms, resources and patient values to come up with clinical decisions.
We are constantly trying to improve our algorithms and data visualizations to help you consume information fast and be current like never before. If you notice any errors in the colour coding or data visualizations please help us by tapping “Report Error” (slide card to left). In addition, please use the ‘Contact Support’ button in Settings to give us feedback or ask questions.
Dr. Sanjeev Singwi
Our summaries are called ‘ebm cards’ and are designed for fast easy consumption by the healthcare professional.
Your Home tab gets updated with new clinical summarierelevant
As users navigate through MDPhD’s A recurrent theme s Clinical or biomedical research is slow to get to the patient bedside. This critical research is essential to practicing evidence based medicine.
Physicians face many significant barriers when integrating published medical research into their daily decision analysis.
The barriers to navigating the clinical research include information overload, time constraints, cost of information and expertise in critical appraisal. Assessing quality and validity of evidence can be challenging.
As a result evidence takes long detours through conferences, expert reviews and guidelines or online textbooks before reaching Physicians , several years after publication date. And most of the evidence sits idly waiting to be discovered.
When best practices are not implemented at the bedside in a timely way, quality of patient care and healthcare costs are impacted.
The null hypothesis or H0 states that there is no difference between samples of data. With discrete variables that means the ratio between both samples is 1. For example the mortality across groups is the same. With continuous variables the difference between both samples is 0. For example the sBP across groups is the same.
Scene 9: We then measure how far away the trial data or H1 is from H0.
Scene 10: Generally in medicine if the data in H1 is more than 2SD away from the data in H0 then it is sufficiently different to reject the H0.
2SDs, 95% CI and p value <0.05 are all acceptable measures of distance between H0 and H1 in Medicine.
Other disciplines like Physics requires H1 to be 4 or 5 SD away from the H0 to be statistically significant!
In Diagnostic tests choosing a threshold means makings tradeoff between sensitivity and specificity. For example moving the threshold to the right increases specificity and decreases sensitivity.
Scene 12: Most of the time however both sensitivity and specificity are important and can be expressed using one statistic known as the area under the ROC curve. A perfectly accurate test has an area of 1.
Scene 13: Confidence Intervals give us information about
statistical significance and clinical significance.
Scene 14. Confidence intervals are used to measure distance from the H0. When the CI overlaps a ratios of 1 or a differences of 0 the result is said to be statistically insignificant.
Scene 15. Confidence intervals help clinicians decide on clinical significance. For example a statistically significant result that is really close to the null hypothesis, may be clinically insignificant. Your patient likely falls somewhere in the range of values in the CI and you have to determine if your intervention will make a clinically significant benefit to there health.