clinical trials

Survival Functions

706050403020100

3.00-censored

1.00-censored

P=.003

Multivariate analysis

Kaplan Meier

Forest plot

NACCMA Metanylysis

Phase III trialINTERPRETATION

Level of evidence

Clinical trials

Bio-medical statisticsSheh Rawat

Survival Functions

706050403020100

3.00-censored

1.00-censored

P=.003

Kaplan Meier

Forest plot

NACCMA Metanalysis

Phase III trial

Level of evidence

Clinical trials

Bio-medical statisticsDr Sheh Rawat

Survival Functions

706050403020100

3.00-censored

1.00-censored

P=.003

Kaplan Meier

Forest plot

NACCMA Metanalysis

Phase III trial

Level of evidence

Clinical trials

Primary And secondary end points

Informed consent

Randomisation

Survival Functions

706050403020100

3.00-censored

1.00-censored

P=.003

Kaplan Meier

Forest plot

NACCMA Metanalysis

Phase III trial

Level of evidence

Clinical trials

DNB:Diplomate in National Board

DNB:Did Not Bat

Rising sun

Good Morning

Or did anyone of you think it was sunset?

Good Evening!

You can prove anything with statistics.

So act with caution in interpreting, and integrity while presenting data

An Overview

• Clinical trials and Levels of evidence

• Definitions and terminology

• Collection of data and creating database

• How to analyze ?

• The need• Types of trials• Levels of evidence• Eligibility criteria• Informed consent• Benefits and possible risks of participating in a clinical

trial• Randomisaton/blind/double blind/multicentric• Defining end points (primary and secondary)• Statistical methods• Interim analysis• Publication

The need of clinical trials

• “Where the value of a treatment, new or old, is doubtful, there may be a higher moral obligation to test it critically than to continue to prescribe it year-in-year-out with the support merely of custom or wishful thinking.”

FHK Green

Clinical trials are experiments to determine the value of treatments.

• There are two key components to the experimental approach.

• First, results rather than plausible reasoning are required to support conclusions.

• Second, experiments should be prospectively planned and conducted under controlled conditions to provide definitive answers to well-defined questions.

Types of Trials

• Prevention trials • Screening trials • Diagnostic trials study tests or

procedures that could be used to identify cancer

• Treatment trials • Quality-of-life (also called supportive

care) trials• Genetics studies

Who sponsors clinical trials?

• The National Cancer Institute (NCI) and other parts of the National Institutes of Health (NIH),

• the Department of Defense,

• the Department of Veterans Affairs, sponsor and conduct clinical trials

What are eligibility criteria, and why are they important?

• Each study’s protocol has guidelines for who can or cannot participate in the study. These guidelines, called eligibility criteria, describe characteristics that must be shared by all participants.

• The criteria differ from study to study.

• They may include age, gender, medical history, and current health status.

• Eligibility criteria for treatment studies often require that patients have a particular type and stage of cancer.

ERBITUX + RT in locally advanced SCCHNPatient inclusion criteria

• Measurable disease

• Pathologically demonstrated SCC of the oropharynx,

hypopharynx, or larynx

• Stage III or IV disease with an expected survival of

>12 months

• Medically able to withstand a course of definitive RT

• Karnofsky PS > 60%

• No evidence of distant metastatic disease

Bonner et al. N Eng J Med 2006;354:567-578

EXAMPLE

Enrolling participants with similar characteristics helps to ensure that the results of the trial will be due to what is under study and not other factors.

In this way, eligibility criteria help researchers achieve accurate and meaningful results.

These criteria also minimize the risk of a person’s condition becoming worse by participating in the study.

What is informed consent?

Informed consent is a process by which people learn the important facts about a clinical trial to help them decide whether to participate.

This information includes details about what is involved, such as the purpose of the study, the tests and other procedures used in the study, and the possible risks and benefits.

• In addition to talking with the doctor or nurse, people receive a written consent form explaining the study.

• People who agree to take part in the study are asked to sign the informed consent form.

However, signing the form does not mean people must stay in the study. People can leave the study at any time—either before the study starts or at any time during the study or the follow-up period.

The informed consent process continues throughout the study. If new benefits, risks, or side effects are discovered during the study, the researchers must inform the participants. They may be asked to sign new consent forms if they want to stay in the study.

Institutional Ethics Committee (IEC)/Institutional Review Board (IRB)

ETHICAL REVIEW PROCEEDURES

BASIC RESPONSIBILITIES

1) To protect the dignity,rights and well being of the potential research participants

2) To ensure that universal ethical values and international scientific standards are expressed in terms of local community and customs

3) To assist in the development and the education of a research community responsive to local health care requirements

COMPOSITION

1)Chairperson

2) 1-2 Basic Medical Scientists

3) 1-2 Clinicians from various institutes

4) One Legal expert or retired judge

5) One Social scientist/ Representative of non-governmental voluntary agency

6) One Philosopher /Ethicist

7)One lay person from the community

8) Member Secretary

REVIEW PROCEDURES

The ethical review should be done through formal meetings and should not resort to decisions through circulation of proposals

Where do clinical trials take place?doctors’ offices

clinics

cancer centers

community hospitals

clinics

veterans’ and military hospitals

cancer centers

community hospitals

clinics

veterans’ and military hospitals

cities and towns across the country

and in other countries.

multicentrc

Where do clinical trials take place?

EXAMPLE

Clinical trials and Levels of evidence

• Phase I: To define and to characterize the new treatment to set the basis for later investigations of efficacy and superiority. eg. Establishment of MTD, toxicity profile, anti tumor activity, basic clinical pharmacology and recommendation of doses for phase II studies.

• For non life threatening diseases: Conducted on human volunteers,

• For life threatening diseases (cancer, HIV): Conducted on patients.

ERBITUX + RT in locally advanced SCCHN

Phase I study of anti-epidermal growth factor receptor antibody cetuximab in combination

with radiation therapy in patients with advanced head and neck cancer

Robert F, Ezekiel MP, Spencer SA, Meredith RF, Bonner JA, Khazaeli MB, Saleh MN, Carey D, LoBuglio AF,

Wheeler RH, Cooper MR, Waksal HW

Robert et al. J Clin Oncol 2001;19:3234-3243

EXAMPLE

ERBITUX + RT in locally advanced SCCHN Study design

Robert et al. J Clin Oncol 2001;19:3234-3243

Previously untreated patients with SCCHN, stage III or IV, or recurrent, not resectable

for curative intent n=16

ERBITUX initial dose (100, 200, 400, or 500 mg/m2) followed by 7 weekly maintenance

doses (100, 200, or 250 mg/m2) +

RT: conventional (70 Gy, 2 Gy / d) or hyperfractionated (76.8 Gy, 1.2 Gy bid)

Follow until disease progression

EXAMPLE

• The dose levels themselves are commonly based on a modified Fibonacci series. The second level is twice the starting dose, the third level is 67% greater than the second, the fourth level is 50% greater than the third, the fifth is 40% greater than the fourth, and each subsequent step is 33% greater than that preceding it.

Traditional phase I trials have three limitations:

• They sometimes expose too many patients to subtherapeutic doses of the new drug.

• The trials may take a long time to complete.

• They provide very limited information about interpatient variability and cumulative toxicity.

accelerated titration designs -ph I trials contd..

• New trial designs have been developed to address these problems.

• One new class of designs, accelerated titration designs, permit within-patient dose escalation and use only one patient per dose level until grade 2 or greater toxicity is seen. Doses are titrated within patients to achieve grade 2 toxicity.

• Accelerated titration designs appear to be effective in reducing the number of patients who are undertreated, speeding the completion of phase I trials, and providing increased information.

accelerated titration designs -ph I trials contd..

• Phase II: used to screen new regimens for activity and to decide which ones to be tested further. Usually designed with 2 or more stages of accrual, allowing early stopping due to inactivity of the regimen.

• Phase III: Randomized trials where the outcome is survival or, time until an adverse event.

• Phase IV trials are conducted to further evaluate the long-term safety and effectiveness of a treatment.

• They usually take place after the treatment has been approved for standard use. Several hundred to several thousand people may take part in a phase IV study.

• These studies are less common than phase I, II, or III trials.

• People who participate in a clinical trial work with a research team. Team members may include doctors, nurses, social workers, dietitians, and other health professionals. The health care team provides care, monitors participants’ health, and offers specific instructions about the study.

What are some of the benefits of taking part in a clinical trial?

• Participants have access to promising new approaches that are often not available outside the clinical trial setting.

• The approach being studied may be more effective than the standard approach.

• Participants receive regular and careful medical attention from a research team that includes doctors and other health professionals.

What are some of the benefits of taking part in a clinical trial?

• Participants may be the first to benefit from the new method under study.

• Results from the study may help others in the future.

What are some of the possible risks associated with taking part in a clinical trial?

• New drugs or procedures under study are not always better than the standard care to which they are being compared.

• New treatments may have side effects or risks that doctors do not expect or that are worse than those resulting from standard care.

• Participants in randomized trials will not be able to choose the approach they receive.

What are some of the possible risks associated with taking part in a

clinical trial?

• Health insurance and managed care providers may not cover all patient care costs in a study.

• Participants may be required to make more visits to the doctor than they would if they were not in the clinical trial.

• The need• Types of trials

• Eligibility criteria• Informed consent• Benefits and possible risks of participating in a clinical trial• Levels of evidence • Randomisaton/blind/double blind/multicentric• Defining end points (primary and secondary)• Statistical methods• Interim analysis• Publication

Clinical trials and Levels of evidence: ASCO guidelines (JCO;17:1999)

• Level 1: Meta-analysis of multiple, well designed, controlled studies. Randomized trials- high powered ( low false+, low false - errors) (Forest plot, L`Abbe plots-odds ratio)

• Level II: At least 1 well designed experimental study. Randomized Trials having low power (hg false + and /or false –ve errors)

Metaanalysis

• A metaanalysis is a quantitative summary of research in a particular area.

• It is distinguished from the traditional literature review by its emphasis on quantifying results of individual studies and combining results across studies.

Metaanalysis-Contd..

• Key components of this approach are:• to include only randomized clinical trials, • to include all relevant randomized clinical

trials that have been initiated, regardless of whether they have been published,

• to exclude no randomized patients from the analysis, and

• to assess therapeutic effectiveness based on the average results pooled across trials

• Including all relevant randomized trials that have been initiated in a geographic area (e.g., the world, or the Americas and Europe) represents an attempt to avoid publication bias.

• Avoiding exclusion of any randomized patients also functions to avoid bias.

• Assessing therapeutic effectiveness based on average pooled results is an attempt to make the evaluation on the totality of evidence rather than on extreme isolated reports.

12% benefit in overall survival:

EXAMPLE

Metanalysis

• 1: Cochrane Database Syst Rev. 2004;(2):CD001774. Neoadjuvant chemotherapy for locally advanced cervix cancer.Neoadjuvant Chemotherapy for Cervical Cancer Meta-Analysis Collaboration (NACCCMA)

• OBJECTIVES: This systematic review and individual patient data (IPD) meta-analysis aimed to assess the effect of neoadjuvant chemotherapy in two comparisons: (1) neoadjuvant chemotherapy followed by radical radiotherapy compared to the same radiotherapy alone; and

• (2) neoadjuvant chemotherapy followed by surgery compared to radical radiotherapy alone.

Eur J Cancer. 2003 Nov;39(17):2470-86. NACT-SX

RT Alone

NACT-RT

RT Alone

EXAMPLE

Metanalysis (NACCCMA)• MAIN RESULTS: In the first comparison, we obtained data from 18 trials

and 2074 patients. When all trials were considered together, a high level of statistical heterogeneity suggested that the results could not be combined indiscriminately. A substantial amount of heterogeneity was explained by separate analyses of groups of trials.

• Trials using chemotherapy cycle lengths shorter than 14 days (HR = 0.83, 95% CI = 0.69 to 1.00, p = 0.046) or cisplatin dose intensities greater than 25 mg/m2 per week (HR = 0.91, 95% CI =0.78 to 1.05, p = 0.20) tended to show an advantage for neo adjuvant chemotherapy on survival.

• In contrast, trials using cycle lengths longer than 14 days (HR =1.25, 95% CI = 1.07 to 1.46, p = 0.005) or cisplatin dose intensities lower than25 mg/m2 per week (HR = 1.35, 95% CI = 1.11 to 1.14, p = 0.002) tended to show a detrimental effect of neo adjuvant chemotherapy on survival.

• In the second comparison, data from 5 trials and 872 patients were obtained. The combined results from all trials (HR = 0.65, 95% CI = 0.53 to 0.80, p = 0.0004) indicate da highly significant reduction in the risk of death with neo adjuvant chemotherapy, but there were some differences between trials in their design and results.

NACT-RT Vs RT Alone

NACT-Surgery Vs RT Alone

EXAMPLE

• In calculating average treatment effects, a measure of difference in outcome between treatments is calculated separately for each trial. For example, an estimate of the logarithm of the hazard ratio can be computed for each trial. A weighted average of these study-specific differences then is computed, and the statistical significance of this average is evaluated.

Study name Statistics for each study Odds ratio and 95% CI

Odds Lower Upper ratio limit limit

Kelly, 1964 0.590 0.096 3.634Hedrin, 1980 0.464 0.201 1.074Leigh, 1962 0.394 0.076 2.055Novak, 1992 0.490 0.088 2.737Saint, 1998 1.250 0.479 3.261Pilbean, 1936 0.129 0.027 0.605Day, 1960 0.313 0.054 1.805Kelly, 1966 0.429 0.070 2.620Singh, 2000 0.718 0.237 2.179Stewart, 1994 0.143 0.082 0.250

0.328 0.233 0.462

0.01 0.1 1 10 100

Favours Tx Favours Pbo

Impact of Treatment on Mortality

Meta Analysis

the point estimate is represented by a square.

EXAMPLE

• This approach to metaanalysis requires access to individual patient data for all randomized patients in each trial. It also requires collaboration of the leaders of all the relevant trials and is very labor-intensive. Nevertheless, it represents the gold standard for metaanalysis methodology.

• the metaanalysis may be useful for answering important questions about a class of treatments that the individual trials cannot address reliably .

Metaanalysis is not an alternative to properly designed and sized randomized clinical trials.

Randomisation

ERBITUX + RT in locally advanced SCCHN Study design

Patients with measurable locally advanced SCCHN (stratified by KPS;node+/0;T1-3/4; RT regimen)

RT as before + ERBITUX initial 400 mg/m2 2-h

infusion then 250 mg/m2 1-h infusion weekly for at least 7 doses

Follow until disease progression or up to 5 years

Randomization

RT once or twice daily or concomitant

boost for 7 - 8 weeks

EXAMPLE

ERBITUX + RT for larynx preservation Study design

Bonner et al. J Clin Oncol 2005;23(16S):Abstract 5533 Updated information presented at ASCO 2005

Subgroup analysis

Randomized

Patients with stage III / IV SCC of larynx and hypopharynx (stratified by KPS;node+/-;T1-3/4; radiation regimen)

Radiotherapy once or twice daily or

concomitant boost for 6 - 7 weeks

Radiotherapy as before + ERBITUX initial 400 mg/m2

2-h infusion then 250 mg/m2 1-h infusion weeklyfor at least 7 doses

n=78 n=93

EXAMPLE

Randomisation-contd..

• There is generally differential bias in the selection of patients to be treated resulting from judgments by the physicians, self-selection by the patients, and differences in referral patterns.

• There may be bias in treatment ineligibility rates. Current patients sometimes are excluded from analysis for not meeting eligibility criteria, not receiving "adequate" treatment, refusing treatment, or committing a major protocol violation.

• The control group, on the other hand, generally contains all the patients.

• There may be differences in the distribution of known and unknown prognostic factors between the controls and the current treatment group.

• Often, there is inadequate information to determine whether such differences are present, and current known prognostic factors may not have been measured for the controls.

• Randomization does not ensure that the study will include a representative sample of all patients with the disease, but it does help to ensure an unbiased evaluation of the relative merits of the two treatments for the types of patients entered.

At what point in time is randomisation done?

• Randomization of a patient should be performed after the patient has been found eligible and has consented to participate in the trial and to accept either of the randomized options.

• A truly random and nondecipherable randomization procedure should be used and implemented by calling a central randomization office staffed by individuals who are independent of participating physicians.

Power of a trial

• The probability of obtaining a statistically significant result when the treatments differ in effectiveness is called the power of the trial.

• As the sample size and extent of follow-up increases, the power increases.

• The power depends critically, however, on the size of the true difference in effectiveness of the two treatments.

• Generally, one sizes the trial so that the power is either .80 or .90 when the true difference in effectiveness is the smallest size that is considered medically important to detect.

Forest Plot

Forest plot

0.328 0.233 0.462

0.01 0.1 1 10 100

Meta Analysisthe confidence interval (CI) for each study is represented by a horizontal line

0.328 0.233 0.462

0.01 0.1 1 10 100

Meta Analysis

the point estimate is represented by a square.

0.328 0.233 0.462

0.01 0.1 1 10 100

Meta Analysis

The size of the square corresponds to the weight of the study in the meta-analysis. ; this is the Mantel-Haenszel weight.

The confidence interval for totals are represented by a diamond shape. The pooled estimate is marked with an unfilled diamond that has an ascending dotted line from its upper point. Confidence intervals of pooled estimates are displayed as a horizontal line through the diamond; this line might be contained within the diamond if the confidence interval is narrow.

• The graph is a forest plot where the confidence interval (CI) for each study is represented by a horizontal line and the point estimate is represented by a square. The size of the square corresponds to the weight of the study in the meta-analysis. The confidence interval for totals are represented by a diamond shape. The scale used on the graph depends on the statistical method. Dichotomous data (except for risk differences) are displayed on a logarithmic scale. Continuous data and risk differences are displayed on a linear scale.

How to interpret the Forest plot?

Intervention group does

better than control

OR = 1.0(no

effect

worse than control

i. Probably a small study, with a wide confidence interval, crossing the line of no effect (OR = 1). Unable to say if the intervention worksii. Probably a small study, wide confidence interval , but does not cross OR = 1; suggests intervention works but weak evidenceiii. Larger study, narrow confidence interval: but crosses OR = 1; no evidence that intervention worksiv. Large study, narrow confidence intervals: entirely to left of OR = 1; suggests intervention worksv. Small study, wide confidence intervals, suggests intervention is detrimentalvi. Meta-analysis of all identified studies: suggests intervention works.

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

better than control

OR = 1.0(no

effect

worse than control

Clinical trials and Levels of evidence ASCO guidelines (JCO;17:1999)

• Level III: Well designed, quasi-experimental studies

such as non randomized, controlled, single group. Pre-post, cohort, time or matched case control series.

• Level IV: Well designed non experimental studies.

• Level V: Case reports and clinical examples.

What level of evidence is “In my experience” ?

Hard evidence

Evidence based medicine

ERBITUX + RT in locally advanced SCCHNStudy endpoints

Primary endpoint– locoregional control (absence of locoregional

disease progression at scheduled follow-up visits)

– Secondary endpoints– overall survival– progression-free survival– Safety (Mucositis, Dysphagia, Radiation

dermatitis, Weight-loss, Asthenia, Xerostomia, Acne-like rash,Infusion reaction)

ERBITUX + RT in locally advanced SCCHN Patient and disease characteristics

CharacteristicsRT only

% (n=213)

ERBITUX + RT

% (n=211)Gender: M / F 79 / 21 81 / 19Median age 58 years 56 yearsKarnofsky PS:

60 - 80% / 90 - 100%

33 / 66 30 / 70

Primary tumor site: Oropharynx 63 56 Hypopharynx 13 17 Larynx 24 27AJCC stage: III/IV 24 / 76 26 / 74Tumor stage: T1-3 / T4

69 / 31 70 / 29

Node stage: N0 / N+ 18 / 82 20 / 80

ERBITUX + RT improves locoregional control over RT alone in locally advanced SCCHN (1)

EfficacyRT only

% (n= 213)

ERBITUX + RT

% (n= 211)

P value/ *Hazard

ratioMedian locoregional control (months)

14.9 24.4 0.005

By site of primary tumor (months):

Oropharynx 23.0 49.0 0.61*Larynx 11.9 12.9 0.69*Hypopharynx 10.3 12.5 0.92*

By disease stageStage III 16.2 38.9 0.69*Stage IV 13.5 20.9 0.73*

Locoregional control rate

2-year 41 50

Primary End point

ERBITUX + RT prolongs survival over RT alone

in locally advanced SCCHN (1)

Efficacy RT only % (n= 213)

ERBITUX + RT

% (n= 211)

P value

Median follow-up

54 months 54 months

Median overall survival

29.3 months

49 months 0.03

Three-year survival

45 55 0.05

ERBITUX + RT prolongs survival over RT alone in locally advanced SCCHN (2)

Secondary end point

ERBITUX does not increase acute RT-induced toxicity in locally advanced SCCHN

Selected or relevant grade 3-5 adverse events reported

RT only %

(n=212)

ERBITUX + RT

% (n=208)

Mucositis 52 56Dysphagia 30 26Radiation dermatitis

Weight loss 7 11Asthenia 5 4Xerostomia 3 5Acne-like rash 1 17Infusion reaction - 3

Yet another Secondary end point

ERBITUX + RT for larynx preservation

Improved preservation of larynx with the addition of cetuximab to radiation for cancers of the larynx

and hypopharynx

Bonner J, Harari P, Giralt J, Baselga J, Shin DM, Cohen R, Jassem J, Azarnia N, Molloy P, Ang K

Subgroup analysis of study by Bonner et al. J Clin Oncol 2004;22(14S):Abstract 5507

EfficacyRT only

% (n= 213)

ERBITUX + RT

% (n= 211)

P value/ *Hazard

ratioMedian locoregional control (months)

14.9 24.4 0.005

By site of primary tumor (months):

Oropharynx 23.0 49.0 0.61*Larynx 11.9 12.9 0.69*Hypopharynx 10.3 12.5 0.92*

By disease stageStage III 16.2 38.9 0.69*Stage IV 13.5 20.9 0.73*

Locoregional control rate

2-year 41 50

ERBITUX + RT for larynx preservation Patient characteristics

Patient characteristics RT only % (n=78)

ERBITUX+RT % (n=93)

Median age (years) 61 59 Gender: M / F 79 / 21 80 / 20

Primary tumor site

Hypopharynx

Larynx

AJCC stage III / IV 28 / 72 38 / 62

ERBITUX treatment (n=91)Median duration of treatment (weeks)

Median number of infusions 8

ERBITUX + RT improves the rate of larynx preservation compared with RT

Efficacy RT only % (n=78)

ERBITUX+RT % (n=93)

Locoregional control:

Median (months)

One-year

Two-year

Overall survival:

Median (months)

Two-year survival

Three-year survival

Laryngeal preservation:

Two-year

Three-year

ERBITUX + RT improves the rate of larynx preservation compared with RT

Efficacy RT only % (n=78)

ERBITUX+RT % (n=93)

Locoregional control:

Median (months)

One-year

Two-year

Overall survival:

Median (months)

Two-year survival

Three-year survival

Laryngeal preservation:

Two-year

Three-year

ERBITUX + RT for larynx preservation Conclusions

• The combination of ERBITUX and RT improves

laryngeal preservation in patients with laryngeal

or hypopharyngeal carcinomas compared with

RT alone

• ERBITUX did not increase the RT-induced

adverse events

Definitions

• Mean, mode, median• Statistical significance (P value)• Variance, SD, SE and• Confidence Interval• Age-Adjusted Rate• DFS, OS, Event, Censored cases• Kaplan Meier/Life table methods• Cox Regression• Odds ratio (Forest and L`Abbe plots)

Definitions and terminology

• A discipline concerned with treatment of numerical data derived from groups of individuals

• To summarize our experience so that we and other people can understand its essential features.

• To use summary to make estimates or predictions about what is likely to be the case in other (perhaps future) situations.

Definitions and terminology

• Descriptive : are methods used to summarize or describe our observations.

• Inferential: Use of observations as a basis for making estimates or predictions (going beyond the fact).

Definitions and terminology Mean, mode, median

• Mean: The sum of all the observations divided by no. of observations.

5,10,15,15,20,25,100,150,1000. Mean:

• Median: of a series of observations is the value of the central or middle observation when all other observations are listed in order from lowest to highest. 5,10,15,15,20,25,100,150,1000 Median:

• Mode: Most frequently occurring value in the series. 5,10,15,15,20,25,100,150,1000

• Mode:

Definitions and terminology Mean, mode, median

• Mean: The sum of all the observations divided by no. of observations.

5,10,15,15,20,25,100,150,1000. Mean:148.5

• Median: of a series of observations is the value of the central or middle observation when all other observations are listed in order from lowest to highest. 5,10,15,15,20,25,100,150,1000 Median: 20

• Mode: Most frequently occurring value in the series. 5,10,15,15,20,25,100,150,1000

• Mode:15

Definitions and terminology: Statistical significance

• If 2 means differ to more than twice the value of SE of the difference, it is said to be statistically significant i.e. more than is likely to have arisen by chance (1 in 20=.05).

• An understanding of statistical significance (p value ) requires understanding of terms like Variance, Standard deviation, Std error.

Definitions

• Mean, mode, median• Statistical significance (P value)• Variance, SD, SE and• Confidence Interval• Age-Adjusted Rate• DFS, OS, Event, Censored cases• Kaplan Meier/Life table methods• Cox Regression

So, the rabbit sat on the ground below the crow, and rested.

A crow was sitting on a tree, doing nothing all day.

A small rabbit saw the crow, and asked him, "Can I also sit like you and do nothing all day long?”The crow answered: "Sure, why not.”

All of a sudden, a fox appeared,

Jumped on the rabbit... and ate it.

Moral of the story is….

To be sitting and doing nothingyou must be sitting very, very high up.

Definitions and terminology Variance, SD, SE and P value

20 observations of Systolic BP in men

Deviation of each observation from the mean (128)

Square of each Deviation from the mean

98 -30 900

160 +32 1024

136 +8 64

128 0 0

130 +2 4

114 -14 196

123 -5 25

134 +6 36

128 0 0

107 -21 441

123 -5 25

20 observations of Systolic BP in men

Deviation of each observation from the mean (128)

Square of each Deviation from the mean

125 -3 9

129 +1 1

132 +4 16

154 +26 676

115 -13 169

126 -2 4

132 +4 16

136 +8 64

130 +2 4

Sum 2560 0 3674

Mean Sq deviation= 3674/20=183.7

• Variance= mean squared deviation=183.7

Total no. 20 -1 (n-1)i.e.19=193.4

• Std. Deviation=Sq root of variance=13.91

• A large standard deviation means frequency distribution is widely spread out from the mean.

• What is the Utility of Std Deviation?

It enables us to test whether the observed diff. between 2 such means are more than would be likely to have arisen by chance.

• Std. Error: of any statistical value is a measure of the SD that that value would show in taking repeated samples from the same universe of observations. It shows how much variation might be expected to occur merely by chance. Std. Error of diff. between 2 means=

• (S.D.1)2 + (S.D.2)2

• n1 n2

• Statistical significance: If 2 means differ to more than twice the value of SE of the difference, it is said to be statistically significant i.e more than is likely to have arisen by chance (1 in 20=.05).

• This calculation sets a standard of judgment which is constant from one person to another.

• A “ significant” answer does not prove that the difference is real; “chance” is still a possible explanation (though unlikely).

• A “ not significant” answer does not tell us that Group A does not differ from group B. It tells that “chance” may easily be a reason for that difference.

• “ p value” does not give the reason for significance or non significance.

• It is only by planning and foresight which can ensure comparable groups of patients and only in such circumstances can we infer that the significant difference between groups is more likely to be due to the specific treatment than to any other factor.

Confidence Interval

• A range of values that has a specified probability of containing the estimated rate or trend of interest.

• The 95% (p-value = 0.05) and 99% (p-value = 0.01) confidence intervals are the most commonly used.

• If an estimated annual percentage change (APC) is -2.44 with a 95% confidence interval of (-2.83, -2.05), then we are 95% confident that the actual APC is between a decrease of -2.83% and a decrease of 2.05%. Inversely, there is still a 5% chance that the actual APC is not in the confidence interval (between the upper and lower confidence limits)

• Confidence intervals are generally much more informative than are significance levels. A confidence interval for the size of the treatment difference provides a range of effects consistent with the data. The significance level tells nothing about the size of the treatment effect because it depends on the sample size. However, it is the size of the treatment effect, as communicated by a confidence interval, that should be used in weighing the costs and benefits of clinical decision making.

Definitions and terminology survivals

• Disease free survival (DFS,PFS).

• Overall Survival (OS): calculated from date of registration or start of treatment.

• Cause specific survivals.

• Systems available: BMDP,SAS,SPSS

• Methods: Life table, Kaplan Meier, Cox regression.

“Event” Vital status

• DFS

• Recurrence, lost to f/up with disease

• OS

• Death, lost to F/up with disease, status unknown (worst case scenario),etc.

Vital status

• Alive;tumor free;no recurrence• Alive;tumor free;after recurrence• Alive with persistent, recurrent, or metastatic

disease• Alive with primary tumor• Dead;tumor free• Dead with cancer• Unknown;lost to follow up• Completeness of follow up is crucial in any study

of survival time to exclude bias in data

Definitions and terminology survivals

• Disease free survival (DFS,PFS).

• Overall Survival (OS): calculated from date of registration or start of treatment.

• Cause specific survivals.

• Systems available: BMDP,SAS,SPSS

• Methods: Life table, Kaplan Meier, Cox regression.

Kaplan Meier curve

Survival Functions

706050403020100

3.00-censored

1.00-censored

Kaplan Meier curve

Survival Functions

706050403020100

3.00-censored

1.00-censored

Life table method

Kaplan Meier curve

Survival Functions

706050403020100

3.00-censored

1.00-censored

Kaplan Meier method• This kind of data usually includes some

censored cases. Censored cases are cases for which the second event isn’t recorded (for example, people still alive disease free at the end of the study).

Kaplan Meier method

• The Kaplan-Meier procedure is a method of estimating time-to-event models in the presence of censored cases.

• It is based on estimating conditional probabilities at each time point when an event occurs and taking the product limit of those probabilities to estimate the survival rate at each point in time.

• Censored cases can happen for several reasons:

1.for some cases, the event simply doesn’t occur before the end of the study;

2.for other cases, we lose track of their status sometime before the end of the study;

3.still other cases may be unable to continue for reasons unrelated to the study

Collectively, such cases are known as censored cases, and they make this kind of study inappropriate for traditional techniques such as t tests or linear regression.

A wife is a wife,

No matter, who the hell you are!!

Cox Regression

• Like Life Tables and Kaplan-Meier survival analysis, Cox Regression is a method for modeling time-to-event data in the presence of censored cases.

• However, Cox Regression allows you to include predictor variables (covariates) in your models, allowing you to assess the impact of multiple covariates in the same model.

Intention-to-Treat Analysis

• One of the important principles in the analysis of phase III trials is called the intention-to-treat principle.

• This indicates that all randomized patients should be included in the primary analysis of the trial. For cancer trials, this has often been interpreted to mean all "eligible" randomized patients. Because eligibility requirements sometimes are vague and unverifiable by an external auditor, excluding "ineligible" patients can itself result in bias.

Intention-to-Treat Analysis-contd..

• However, excluding patients from analysis because of treatment deviations, early death, or patient withdrawal can severely distort the results.

• Often, excluded patients have poorer outcomes than do those who are not excluded.

• Investigators frequently rationalize that the poor outcome experienced by a patient was due to lack of compliance to treatment, but the direction of causality may be the reverse.

Intention-to-Treat Analysis –contd..

• In randomized trials, there may be poorer compliance in one treatment group than the other, or the reasons for poor compliance may differ.

• Excluding patients, or analyzing them separately (which is equivalent to excluding them), for reasons other than eligibility is generally considered unacceptable.

• The intention-to-treat analysis with all eligible randomized patients should be the primary analysis.

Intention-to-Treat Analysis –contd..

• If the conclusions of a study depend on exclusions, then these conclusions are suspect.

• The treatment plan should be viewed as a policy to be evaluated.

• The treatment intended cannot be delivered uniformly to all patients, but all eligible patients should generally be evaluable in phase III trials.

Interim analysis

• It has become standard in multicenter clinical trials to have a data-monitoring committee review interim results, rather than having the monitoring done by participating physicians.

• This approach helps to protect patients by having interim results carefully evaluated by an experienced group of individuals and helps to protect the study from damage that ensues from misinterpretation of interim results.

Interim analysis-contd..

• Generally, interim outcome information is available to only the data-monitoring committee.

• The study leaders are not part of the data-monitoring committee, because they may have a perceived conflict of interest in continuing the trial.

• The data-monitoring committee determines when results are mature and should be released.

• These procedures are used only for phase III trials.

What happens when a clinical trial is over?

• the researchers look carefully at the data before making decisions about the meaning of the findings and further testing.

• After a phase I or II trial, the researchers decide whether to move on to the next phase, or stop testing the agent or intervention because it was not safe or effective.

The results of clinical trials are often published in peer-reviewed, scientific journals.

• Peer review is a process by which experts review the report before it is published to make sure the analysis and conclusions are sound.

• If the results are particularly important, they may be featured by the media and discussed at scientific meetings before they are published.

• Once a new approach has been proven safe and effective in a clinical trial, it may become standard practice.

Publication bias

• An additional factor to consider is that of publication bias, which denotes the preference of journals to publish positive rather than negative results. A negative result may not be published at all, particularly from a small trial. If it is published, it is likely to appear in a less widely read journal than it would if the result were positive.

• Publication bias

These observations emphasize that results in the medical literature often cannot be accepted at face value.

Publication bias

• It is essential to recognize that "positive" results need confirmation, particularly positive results of small studies, before they can be believed and applied to the general population.

Where can people find more information about clinical trials?

• People also have the option of searching for clinical trials on their own. The clinical trials page of the NCI's Web site, located at http://www.cancer.gov/clinicaltrials/ on the Internet,

• Another resource is the NIH's ClinicalTrials.gov Web site. ClinicalTrials.gov lists clinical trials sponsored by the NIH, other Federal agencies, and the pharmaceutical industry for a wide range of diseases, including cancer and other conditions. This site can be found at http://clinicaltrials.gov/ on the Internet.

This happens only in india

Thankyou

How to tabulate &collect in what form?

• Strings

(Names)

• Numericals

(Sex, types of protocols, stage, race,etc.)

Status DFS

1:Residual

2:LFU with disease

3:LFU without disease

4 :disease Free

Status OS

1:Dead due to disease

2:LFU with disease

3:LFU without disease

4:Alive with disease

5:Alive disease free

6:dead other cause

Hands on SPSS

• Understanding variables.

• Entering the data.

• Analysis of data.

Conclusion

• It is a means of coming to conclusions in the face of

uncertainty

“Act with caution”

You can prove anything with statistics.

There are lies, dammed lies, and statistics.

Figures don’t lie, but liars use figures.

Act with Integrity.

A little bird was flying south for the winter.

It was so cold, the bird froze and fell to the ground in a large field.

While it was lying there, a cow came by

and dropped some dung on it.

As the frozen bird lay there in the pile of cow dung, it began to realise how warm it was. The dung was actually thawing him out!

He lay there all warm and happy, and soon began to sing for joy.

A passing cat heard the bird singing and came to investigate.

Following the sound, the cat discovered the bird under the pile of cow dung, and promptly dug him out and ate him!

PURR....

The morals of this story are:

1) Not everyone who drops shit on you is your enemy.2) Not everyone who gets you out of shit is your friend.3) And when you're in deep shit,keep your mouth shut

Don’t ask me 2oo many ?s.I’m not a statistician

Thank you. Have a nice day

clinical trials

Education

clinical trials - uc davis health › cancer ›...

biocapax clinical trials

1 clinical trials digital pathology in clinical trials...

clinical trials: cost coverage and implications for the...

open access to clinical trials through prospective clinical...

building quality into clinical trials -...

clinical trials of health information technology ... · pdf...

understanding clinical trials

the new clinical trials regulation to improve clinical...

th clinical trials - d2cax41o7ahm5l.cloudfront.net · •...

unit2 clinical trials

clinical trials india

outsourcing clinical trials survey · outsourcing clinical...

clinical trials

intelligent clinical trials - deloitte us · intelligent...

nci guidelines for auditing clinical trials … guidelines...

sas clinical trials

praventin clinical trials

minority participation in cancer clinical trials...

cancer clinical trials the basic workbook · pdf filethe...