{"id":631,"date":"2023-09-11T08:32:31","date_gmt":"2023-09-11T08:32:31","guid":{"rendered":"https:\/\/rdhs.oliviahospital.lk\/?page_id=631"},"modified":"2023-09-11T08:34:08","modified_gmt":"2023-09-11T08:34:08","slug":"statistics","status":"publish","type":"page","link":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/statistics\/","title":{"rendered":"STATISTICS"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t

WHAT IS MEDICAL STATISTICS<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

INTRODUCTION :
<\/span><\/span><\/strong>M<\/span>ost practicing physicians view medical statistics as a\u00a0complex mathematical topic alien to biological science\u00a0and medicine. Statistical jargons create a sense of fear and\u00a0a compulsive response to avoid delving into this arena.\u00a0But most medical publications cite, quote and depend\u00a0on is statistical terms, which we need to understand as\u00a0an end user. Understanding a medical publication, be it\u00a0a drug trial, a case report, an epidemiological study or a\u00a0meta-analysis, needs knowledge of medical statistics. The\u00a0intention of this article is to simplify statistical terms, so\u00a0that the reader can differentiate good robust publications\u00a0from statistically weak ones.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

STATISTICAL TERMS<\/span><\/strong><\/p>

Probability<\/span><\/strong>
Probability (p) is the most commonly used term in\u00a0statistics. Probability is an indicator of how much of theresult (outcome) can happen just by chance. If we toss a\u00a0coin, it has a 50% chance of showing a head or a tail. If\u00a0we continue to toss hundreds of time, the chance of it showing head or tail tends to be about half (50%). This\u00a0can be mathematically put as 50\/100 = 5\/10 = 0.5.\u00a0This means that a probability value of (p) = 0.5\u00a0indicates that there is 50% probability that the result may\u00a0be happening just because of a chance.<\/span><\/p>

In similar sense, different p-values would mean the\u00a0following:
p = 0.05 = 5\/100 = 1\/20 chance of the result being\u00a0accidental
p = 0.01 = 1\/100
p = 0.001 = 1\/1000
To be statistically significant, we need to look for a\u00a0p-value of at least less than 0.05.
Statistically speaking:
p < 0.05\u2014statistically significant
p < 0.01\u2014highly significant
p < 0.001\u2014very highly significant
Always look at the sample size, effect size and confidence interval before giving a judgment on p-value alone.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

Sample<\/span> Size and Effect Size<\/span><\/strong>
p-value is inversely proportional to effect size and sample\u00a0size. This means if we are testing the effects of a new\u00a0blood pressure medicine is reducing blood pressure\u00a0by 20 mm Hg, but was tested in only 100 patients, the\u00a0p-value may be significant but the sample size is too small\u00a0to be significant. In similar lines, a drug reducing blood\u00a0pressure by merely 1 mm Hg may return a significant p\u00a0value, when tested in 1 million population but the effect\u00a0size is too small to be really worthwhile statistically.\u00a0To avoid statistical error, today, the statistician\u00a0calculates the minimum sample size that would be\u00a0required to show a difference of results depending on\u00a0the prevalence of the event rate in the natural course of\u00a0the disease in the population being studied.<\/span><\/p>

Sample Size Calculation<\/span><\/strong>
To decide on the sample size of a trial, the following\u00a0factors need to be considered:
Sample size calculation
\u2022 Effect size
\u2022 Significance level
\u2022 Power of a trial<\/span><\/p>

Effect Size<\/span><\/strong>
If a \u2018old\u2019 drug A is effective in 30% population and \u2018new\u2019\u00a0drug B in 40%, the effect size is calculated as:
B\u2013A\u00a0 =>\u00a040\u201330 = 10% (absolute)\u00a0and 30\/40 = 75% (relative)<\/span><\/p>

Significance Level<\/span><\/strong>
If we are looking at 5% significance level, we have a\u00a0p-value of 0.05. If a study is not significant but gives an\u00a0impression of being significant (generally because of\u00a0borderline p-value), it is called a type 1 or alpha error.<\/span><\/p>

Power<\/span><\/strong>
If in a trial, the number of sample is small, a significant\u00a0trial may look statistically nonsignificant. This is called\u00a0beta error. A repeat trial with a larger sample size is likely\u00a0to correct such error. Power of a trial is 1-b. The usual\u00a0power of a trial is kept at 80% (0.8).<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

Odds and Odds Ratio<\/span><\/strong>
The term odds mean a disease or effect happening vs\u00a0not happening.\u00a0Supposing that 10 out of 100 patients of acute myo-cardial infarction (MI) would die, the odds are 10 will\u00a0die and 90 will live.\u00a0So the odds are 10\/90 = 0.11 (happens\/not happen).
Now a medical paper says that there is a new drug\u00a0ABC shows benefit in reducing death rate of MI. On being\u00a0treated with the new drug, only 2 out of 100 acute MI\u00a0cases died. This means 2 dies and 98 live.\u00a0So odds for this new treatment are 2\/98 = 0.02.
Odds ratio = control odds\/treatment odds = 0.11\/0.02 = 5.5
This means treatment with this new drug reduces\u00a0chance of death by 5.5 times.
<\/span>
Risk and Risk Ratio<\/span><\/strong>
Risk is a similar term that means disease or effect out of\u00a0the entire population. In the same example of the heart\u00a0attack above, the risk of death is 10 out of 100 (not 10\/90\u00a0as in odds).
Risk of death in MI 10\/100 = 0.10 (happens\/total)
Similarly, risk after being administered ABC has a\u00a0risk of 2\/100 = 0.02.
Risk ratio = 0.10\/0.02 = 5 (very close to odds ratio)
Note that, in most cases, odds ratio and risk ratio is\u00a0close.
Now consider, unlike the example of MI, a disease has\u00a0a mortality of 90% (90 out of 100 die). In such a scenario,\u00a0the odds would be 90\/10 (died 10 survived) returning a\u00a0value of odds of 9, while the risk ratio would be 90\/100 =\u00a00.9. So, dichotomy between odds and risk indicate high\u00a0event rate in control group and this may corrupt a study.<\/span><\/p>

S<\/strong>tandard Deviation<\/b><\/span>
Any physiological parameter will have variation from\u00a0people to people. If we take sufficient\u00a0number of people\u00a0and plot their values, they tend to follow a normal or\u00a0nominal distribution (a bell-shaped curve). The central\u00a0line in the curve is the mean (akin to arithmetic average).\u00a0As we go farther away from the mean, the degree of\u00a0deviation (from mean) increases. One standard deviation\u00a0(SD) covers 68.2% of data spread from mean. Two SD\u00a0covers more, up to 95.4% data and 3 SD covers almost a\u00a0whooping 99.7% data.<\/span><\/p>

How to use SD Intelligently<\/span><\/strong>
You read a paper, which says that, after a stroke, the\u00a0patients have stayed in hospital for 8 \u00b1 3 (p < 0.01). Look\u00a0pretty impressive? The number \u00b13 denotes 1 SD and\u00a0that is 68.2% patient stayed in the hospital from 5 to 11\u00a0days (8 \u00b1 3).\u00a0Now, multiply 3 \u00d7 3 SD to get the 3 SD values (covering\u00a099.7%). This value is 9. This means the patients would\u00a0have stayed in the hospital from \u20131 (minus 1) day to 17\u00a0(8 \u00b1 9) days. This is mathematical impossibility. So, this\u00a0study SD is unacceptable.<\/span><\/p>

Confidence Interval<\/span><\/strong>
Confidence interval (CI) is the interval which indicates the\u00a0lowest and the highest chances of the occurrence in the\u00a0particular trial. In simple terms, a confidence interval of\u00a018 to 34 means that the next time you do the same study\u00a0the result may be as low as 18 or as high as 34.\u00a0<\/span><\/p>

Study Example<\/span><\/strong>
How to interprete a Clinical Trial<\/span><\/strong>
A clinical trial tells us whether a new therapy is better\u00a0than older one. If so it changes our clinical practice\u00a0pattern. So, we need to carefully look into the results to\u00a0decide how much impact it would have on our clinical\u00a0practice tomorrow. Most new studies try to show that\u00a0the effect of the study medicine is better than the present\u00a0standard of care, some try to prove noninferiority.\u00a0Many noninferiority trials try to show whether the\u00a0drug or treatment is \u2018not worse\u2019 compared to the present\u00a0gold standard, but do not ask whether it is better or not.\u00a0This means it is checking one end (single tail) of the\u00a0standard bell curve and not both the ends (double tails). A\u00a0single tailed trial needs a higher p-value to be significant\u00a0(at least p < 0.01).<\/span><\/p>

TYPES OF TRIALS<\/span><\/strong>
While observational studies\u00a0are important in deciding prevalence of a disease\u00a0or its complications, interventional studies which are randomized and placebo controlled are the most robust.<\/span><\/p>

Study Design<\/span><\/strong>
There may be several types of clinical trial depending on\u00a0what exactly we need to try. Today, the cardiovascular\u00a0(CV) event rates (death, MI, stroke) of most diseases are\u00a0so low that to detect a difference made by a \u2018new\u2019 drug\u00a0may be quite low in absolute numbers. This necessitates\u00a0the need for larger trials with more power to detect even\u00a0a mild benefit of the \u2018new\u2019 drug.<\/span><\/p>

Randomized Trials<\/span><\/strong>
Active (new drug) group is allocated randomly to avoid\u00a0bias. In a nonrandomized trial, a new drug may be\u00a0given only to patients with milder disease, resulting\u00a0in an erroneous interpretation that the drug is better.
Randomization avoids that bias.<\/span><\/p>

Placebo Controlled<\/span><\/strong>
Today, all patients need to be given standard of care for\u00a0diseases. The active new drug group is given the new\u00a0drug over and above the standard therapy, making the\u00a0standard of therapy as placebo.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

Multicentric<\/span><\/strong>
To make trials larger in number, multiple centers collaborate together to get larger numbers. Multicentric trials\u00a0also ensure a homogenous mixture of patients of different\u00a0ethnicity, different socioeconomic backgrounds, making\u00a0it easy to decide whether this therapy is consistent among\u00a0all segments of population.
<\/span>
Blinded<\/span><\/strong>
Blinding makes the patients and the treating physicians\u00a0blind to what therapy they are on. The analysis of the\u00a0trials is done by analysts and statisticians who are also\u00a0blinded.<\/span><\/p>

End Points<\/span><\/strong>
These are prespecified prior to the beginning of the\u00a0trial. End points can be primary and secondary. Most\u00a0CV trials have primary end points as a combination of\u00a0death, nonfatal MI and nonfatal stroke. Non-prespecified\u00a0end points, analyzed at a later date after the trial (as an\u00a0after thought) are called posthoc analysis which has less\u00a0statistical power of prediction.
<\/span>
PROBE Design<\/span><\/strong>
The downside of large RCTs are the cost. One way of\u00a0reducing cost is to do an open trial but with endpoints\u00a0blinded. These trials are called prospective randomized\u00a0open-labelled end-point blinded (PROBE) trial. Here,\u00a0the endpoints are blinded, giving it power to analyze\u00a0significance, but many consider it inferior to RCTs.
<\/span>
CONCLUSION<\/span><\/strong>
Understanding the basics of medical statistics help the\u00a0clinician in making important judgments and decisions.\u00a0It helps the clinician to make evidence-based changes in\u00a0the practice, or junk it depending on the merits. A step-by-step approach by looking at ARR, p-value, NNT and\u00a0confidence interval can separate statistically robust data.<\/span><\/p>

REFERENCES<\/span><\/strong>
1. Essential Evidenced-based Medicine, Dan Mayer, Cambridge\u00a0Medicine. 2nd ed.
2. Harris M, Taylor G, Dunitz M. Medical Statistics Made Easy.
3. Magnello E, Loon BV. Statistics, A graphic Guide, Icon Books,\u00a0UK.
4. Nair T. Medical Statistrics and Clinical Trials. 1st ed. Wiley\u00a0Blackwell.
5. Sanjay K, Diamond G. Trial and error: how to avoid commonly encountered limitations of published clinical trials. J Am Coll Cardiol.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

WHAT IS MEDICAL STATISTICS INTRODUCTION :Most practicing physicians view medical statistics as a\u00a0complex mathematical topic alien to biological science\u00a0and medicine. Statistical jargons create a sense of fear and\u00a0a compulsive response to avoid delving into this arena.\u00a0But most medical publications cite, quote and depend\u00a0on is statistical terms, which we need to understand as\u00a0an end user. Understanding […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-631","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/pages\/631","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/comments?post=631"}],"version-history":[{"count":0,"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/pages\/631\/revisions"}],"wp:attachment":[{"href":"https:\/\/healthdept.ep.gov.lk\/rdhs-kl\/wp-json\/wp\/v2\/media?parent=631"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}