Research Design and Literature Evaluation 1

Convience sample

"Haphazard" sampling The investigator decides a certain criterion to be evaluated and chooses the subjects at will Quicker and easier to do Difficult to make judgements about entire population Ex. "Person on the street" opinion polls and website polls

Journal club presentation statisitics

"Meat" of the study - incorporate your knowledge Be able to identify and explain intention-to-treat verses per-protocol, power analysis, appropriate statistical tests, type of data, etc

Cross-sectional study

"Snapshot" study Exposure of interest, disease of interest, and covariates are measured from random sample at single point in time Data is collected through surveys, databases, charts, etc Ex. -rofecoxib (vioxx) and sudden death -smokers current health status -teacher evaluations

Case Control Study Table 10.2-8 results from a case-control study examining the association of cutaneous melanoma and the use of sun beds and sunlamps Exposure. Yes. No Cases. 67. 210 Controls. 41. 242 Determine the odds ratio

(A/b)/(c/d) (67/201)/(41/242) = 1.88

Descriptive statistics

*describes* Summarizes a anergy of observations (frequency, central tendency, spread) facilitating discussions S about distribution Bowling score Batting average GPA Can be displayed in a scatter plot, frequency distribution table, histogram Categories include measures of central tendency and measures of dispersion (or variability) Concerned with presentation, organization, and summarization of data Summarizes a set of observations (frequency, central tendency, spread) facilitating discussions about distributions

In clinical practice the provision of drug information involved the

*efficient retrieval, evaluation, and communication of medication information* to assist in care decisions, develop evidence-based recommendations, and improve patient outcomes

Using clinical pharmacology

*good for inactive ingredients* Reports allows you to see drug interaction reports, IV compatibility, product comparison reports, and adverse reaction reports Find by allows you to search NDC, indication, classification, adverse reaction, and contraindication/precaution Lists includes monographs and products with extra precautions such as REMS, boxed warnings, interactions with grapefruit juice, do not crush, narrow therapeutic index, etc Resources provides a drug identifier, lab reference values, clinical calculators, drug class overviews, VIS handouts, etc Patient education under resources provides counseling sheets that are searchable by drug product or by condition

Confidence interval interpretation

*increasing* the number of subjects/patients *narrows* (or decreases) the confidence interval *decreasing* the number of subjects/patients *widens* (or increases) the confidence interval All values within the confidence interval are *statistically the same*, meaning that the true population mean lies somewhere within the interval *If confidence intervals overlap, the two groups are not significantly even if the means are different*

Inferential statistics

*infers* Conclusions or generalizations made about a population from a study of that group Who is the best basketball player of all time? How many homeless people live on the streets of Chicago? Utilize descriptive statistics to make generalizations Require random sampling of population, meaning each member of population of interest has equal chance of being chosen for the sample

Advantages of cohort study

- Investigates outcomes from rare exposures - study determines disease incidence rate - exposure is known at beginning of study - prospective data collection

Advantages of cross-sectional study

- cost and time efficient - measure current health status - hypothesis generating

A study yo are evaluating claims to have a 90% power and a level of significance set at 0.025. Which of the following represents the likelihood of type 1 error? 0.025 0.05 0.10 0.20

0.025 From the alpha value 0.10 would be for a type 2 error

In vivo studies Use when? 1 Used to determine? 2 Models include? 3

1 -after in vitro work -short term lasts about 2 weeks to a month 2 -set dose for IND -reproductive, developmental, behavioral, cardiovascular, renal, etc 3 -rodents -rabbits -dogs -monkeys -others Gene toxicity -carcinogenesis -mutagenesis Cardiovascular toxicity -telemetry Reproductive toxicity -impairment of fertility -pregnancy -nursing Behavioral studies -Morris water maze -marble burying test -rota rod performance

In vitro studies Use to determine: 1 Studies include: 2 Endpoints: 3

1 -max tolerated dose -max feasible dose 50X multiple (product highest dose then multiply) 2 -PK study -tolerability study -toxicokinetics study -mutagenicity 3 -cell density -cell proliferation -cell death (apoptosis) -cell migration

Power

1-beta Probability of making correct decision when the null hypothesis is false (ex. Probability of detecting a difference if one actually exists) In real terms, the ability the detect difference between the groups (ideally should be as high as possible) Most studies desire able power is 80% In study design, once a researcher decides the alpha and beta levels for the study, then a *power calculation* can be performed to determine the number of samples needed (or patients to enroll)

Steps to calculating standard deviation

1. Calculate the mean 2. Get the deviations of each data point 3. Square the resulting values 4. The the resulting values from #3 together 5. Divide by total numbers less one (n-1) 6. Take the square root of the resulting value and this is your standard deviation

Major points of choosing a test

1. Look at the *independent variables* (discrete or continuous) 2. Look at the *dependent variable* (discrete or continuous) 3. Look at *number of levels (or groups)* to compare (two or more) 4. Look at the *relationship between/among groups* (paired or unpaired) 5. Look at the *characteristics* of the data (parametric or nonparametric)

Decision errors: Researchers decision v true outcome Ho is true. Ho is false Accept Ho. 1. 2 Reject Ho. 3. 4 Type 1 error (5) Type 2 error (6)

1. No error 2. Type II error 3. Type I error 4. No error 5. Inappropriately *reject* the null hypothesis, leading to a *false positive* result - causes more problems 6. Inappropriately *accept* the null hypothesis, leading to a *false negative* result

Steps in answering a drug information question

1. Secure demographics (who's calling? - direct response appropriate to requestor's position, training, and knowledge; gather contact information - practice site, address, telephone/fax number, e-mail; secure mechanism for response delivery - verbal, written, email 2. Obtain background information - why is the requestor asking for this information? (Differentiate between patient-specific or General; have any resources already been consulted?; speaking to intermediary in transfer?; find out what you missed; what prompted this question? Is this question regarding a specific patient? What is the patient's demographic data (age, weight, height, gender, allergies, history, labs), what resources have already been consulted? How soon does the requestor need a response?) 3. Determine and categorize the ultimate question - refocusing the requestor's inquiry (appropriate retrieval of background information facilitates this process; why is the requestor asking for this information?); confirm ultimate question with requestor before categorizing; categorization serves as a road map to the response 4. Develop strategy and conduct search (a search strategy ensures efficiency and completeness; utilize references related to the organization; we typically search tertiary then secondary then primary literature) 5. Perform evaluation, analysis, and synthesis (step where PharmD can display professional training; evaluate the value of the information along with the utility and reliability of the resource; analyze and interpreter data into clinical outcomes; synthesize to transmit available information into a logical and concise response) 6. Formulate and provide response (reflect on how the info may be used (restate questiosn and background info); aim to clearly inform requestor (present both sides of an arguement, succinct yet adequately comprehensive, foresee questions); be professional (format response based on requestor's knowledge, use proper grammar and spelling, avoid first person, write out uncommon abbreviations and acronyms, provide resources used, etc) 7. Conduct follow-up documentation (its just good business (demonstrates professionalism, reduced legal liability, validates value to healthcare system); availability of new info may prompt an update of original response, reimbursement)

Equation for outliers*

1.5 x IQR Add to upper quartile or minus the lower quartile

Drug X and Drug are being evaluated as clinical cures for herpes simplex. Drug X result in a cure rate of 81%, while Drug Y has a cure of 72%. What is the number needed to treat (NNT) to cure one patient with drug X? 4 9 11 22

11

Ames Assay What does it measure?

1970s Bruce Ames Test substances for ability to *cause cancer* Combines the use of mainly tissue to generate active metabolites of the substance with a *test for mutagenicity in bacteria (salmonella)* The mutagenicity of a substance is proportional to the number of colonies observed

You are measuring the temperature of 10 random patients out of the whole population of 100. How well might those 10 patient temperatures (the mean of the samples) represent the temperature of the rest of the population? Mean=99.4F SD=2.45 N=10

2.45/square root of 10 0.77 More accurate and lower if we increase the sample size

CV given mean and standard deviation Mean=24.4 SD=31.23

31.23/24.4=1.128 31.23/24.4 x 100 = 128%

Calculate the median of {1, 3, 5, 8, 10, 19}

5 and 8 are in the middle (5+8)/2 = 6.5

Median*

50th percentile Point where 50% of values fall above and 50% fall below *Less affected by outliers than mean (useful for ordinal or continuous data (especially skewed data)* Put in order from smallest to largest Find the middle value is the number is odd If the number of values is even, then the median is the average or mean of the two values in the middle Best to use if we have outliers

The internet problem with finding health information

72% of users reported looking online for health information within the past year (in 2012) Inadequate law exists regarding internet content with litter to no means of ensuring the accuracy of posted information Potential for misinformation to be disseminated

In the normal (Gaussian) distribution ___ % of the data are found within 2 standard deviations (+/-) the mean of the data? 50% 66% 95% 99%

95%

Which of the following statements is true pertaining to parametric data? Data should be normal/ordinal and normally distributed Data can be paired, matched, or correlated Variances within the groups compared are unequal 95% of the data are found within 2 standard deviations (+/-) the mean

95% of the data are found within 2 standard deviations (+/-) the mean

External validity

Ability to apply results into practice Becomes limited with flaws in internal validity

Question classification examples

Adverse drug reaction Alternative medicine/homeopathy Bioequivelence Compounding Contraindications Dictionary/medical terms Disease state Dose/dosing Drug administration Drug approvals Drug ID/availability Drug interactions Drug manufacturers Drug monographs Drug therapy Foreign drug ID Investigational drugs IV compatibility/ stability IRB Laboratory values Literature evaluation/bio stats Literature search Managed care Medical writing/style Nutrition OTC Patient counseling Pharmecutics Phamacoeconomics Pharmacokinetics Pharmacy management Pharmacy law Pharmacy practice Pharmacology Pregnancy and lactation Toxicology/positioning Vaccines Other

Independent variables: -Age -sex -APACHE II score Dependent variable: -mortality (expired/not expired) What types of data? Test to use?

Age is ratio Sex is nominal Apache score is ordinal Mortality is nominal Linear regression

Randomization

Al study subjects have an equal chance of being assigned to either the intervention or control group Subjects are eligible after meeting trial inclusion/exclusion criteria *minimizes bias* Want groups as similar in characteristics as possible (age, gender, severity of illness, etc) Control may be a 1:1 allocation, but may see 2:1, 3:1:1 etc Various methods avaliable (simple, blocked, stratified)

regression analysis

Allows a mathematical way to estimate or predict values of one variable based on known values of another variable Using X variable to predict y variable Many different methods -simple linear regression -multiple linear regression -simple logistic regression -multiple logistic regression -nonlinear regression -polynomial regression

Open vocabulary

Allows for search of any term in the title, abstract, author listing, article, etc Search will find term regardless of whether it is important to the article

Telemetry

Allows remote measurement and reporting of information -heart rate -BP -QT interval Cardiovascular toxicity

Probability established before the study (a priori) of falsely REJECTING the null hypothesis when it is really true:

Alpha level

What represents the predetermined allowable amount of type 1 error in a study?

Alpha value

Standard error of the mean (SEM)

Also called the standard deviation of the mean Estimates the standard deviationfo f sampling distributions Estimates the variation of the standard deviation of different samples within the same population *Decreases as sample size increases* *SD/square root of n*

Closed vocabulary

Also known as thesaurus (MeSH, Emtree) Restricts search to identified keywords for given article (high blood pressure would be categorized as hypertension)

Variable

An attribute that describes a person, place, thing, or idea Can vary from one entity to another

If we believe that 2 treatments are equally good then

Another design is needed perform an equivalence trial

Cell death measures

Apoptosis

Journal club presentation Application Importance

Application -discuses the application of the study into practice -consider disease state treatment issues -consider place for this treatment in pre-existing therapy Importance -discuss the degree of importance of this study to pharmacy medicine -discuss the clinical meaningfulness (external validity)

When might noninferiority studies be used?

Applications based upon similarity in areas where bioequivelence studies are not possible Products with potential safety advantage over the standard to allow a risk-benefit assessment to be made Cost-effectiveness studies Disease state areas where the use of placebo is unethical

Simple randomization

Assigned subjects according to some criteria Day of the week, subject birthday, medical record number, coin toss Not considered proper randomization since number of subjects in the groups can be imbalanced If investigators assigned all subjects with an office visits on Thursday to one group, subjects did not have an equal opportunity to be assigned to either group

Mean

Average Sum of all values divided by total number of values *Used for continuous and normally distributed data* (interval and ratio) Very sensitive to outliers (each values distance from the other ones is not taken into account

Morris water maze

Behavioral study on mice exposed throughout gestation and 2 weeks post natal Spatial learning and memory If they had increased latency period "swim time" *Test for dementia*

Goal is to find the __ evidence

Best Sources should be up-to-date and valid

Which of the following represents the predetermined allowable amount of type II error in a study? P-value 1-Beta Alpha Beta

Beta

Probability established before the study (a priori) of falsely ACCEPTING the null hypothesis when it is actually false:

Beta level

Embase

Biomedical and pharmacological scientific literature from *8500+ published journals-including MEDLINE plus more* *conference abstracts* Emtree thesaurus (*closed vocabulary*) *high cost*

Double blinding

Both investigators and subjects are not aware of the assigned groups

Triple blinding

Both investigators and subjects are not aware of the assigned groups PLUS data interpreters are are not aware of subject assignment

Disease state background

Briefly discuss therapeutic background for disease Provide scientific references for this information (disease state guidelines preferred) Include primary therapies (both drug and non drug) and second line agents Stay relevant to the article at hand (ex. In reference to the current standard of care, this is where the study fits in)

Cybermedicine

Broader concept that includes the marketing, relationship creation, advice, prescribing, and selling of pharmaceuticals and devices in cyberspace

Pharmacy typically reports to the

CEO Viewed by the Investifgational review board before it is send to the board of directors

Community dwelling female subjects 70+ years of age (n=2500) were randomly assigned to recieve either 400 IU of synthetic alpha-tocopherol (n=1239) or placebo (n=1261) by mouth on a daily basis for 7-9 years with a 12-month follow-up period. Median timeframe of intervention for each group was 8.36 years. The primary endpoint of the study was first major cardiovascular event. Secondary endpoints were the individual endpoints of the primary endpoint. During follow-up, major cardiovascular events occurred less in the subjects taking vitamin E compared to those taking placebo (47 [3.79%] vs. 52 [4.12%], respectively; 95% CI, 0.81 to 1.03). With regard to individual end points, the risk of stroke was reduced with vitamin E (RR=0.83). Vitamin E had no significant effect on the risk of fatal or non fatal MI (RR=1.02) or death from cardiovascular diseases (RR=1.01). Major bleeding occurred more frequently in the high-dose vitamin E group than in the placebo group (25 [2.01%] vs. 16 [1.27%], respectively). Fill out the chart Calculate the the following in decimal format, rounding to the nearest thousandth when applicable. Relative risk= Absolute risk reduction= Numbers needed to treat=

CHD Yes. No. Total Vitamin E. 47. 1192. 1239 Placebo. 52. 1209. 1261 RR=0.0379/0.0412=0.920 ARR=0.0412-0.0379=0.003 NNT=1/0.003=334

Locating guidelines

Can be accessed from: -organization that developed the guidelines -secondary (Medline) or tertiary resources -national guideline learning house - no longer up to date Turning research into practice Guidelines international network Cardiology (American college of cardiology, American heart association, American college of chest physcians, national heart, lung, and blood institute) Endocrinology (American association of clinical endocrinologists, American diabetes association, national osteoporosis foundation) Gastroenterology (American college of gastroenterology) Infectious diseases (infectious diseases society of America) Nephrology (kidney disease: improving global outcomes; kidney disease outcomes quality initiative) Neurology (American academy of neurology) Obstetrics (American congress of obstetricians and gynecologists) Oncology (national comprehensive cancer network) Pediatrics (American academy of pediatrics) Psychiatry *american psychiatric association; american academy of child and adolescent psychiatry) Pulmonary (national heart, lung, and blood institute; global initiation for chronic obstructive lung disease)

Discrete data

Can only take a limited number of values within a given range Nominal or ordered

Continuous data

Can take up any value within a given range Anything not discrete Includes "counting variables" like number or occurrences Interval and ratio Discrete data

Phase 1 *

Carried out in healthy volunteers Usually small 20-100 Identify metabolic and pharmacological effects of drug in humans Determine side effects associated with increasing doses

Superiority trial

Classic design Proving experiment is better than control *often phase 3 RCTs* conducted for establishing SUPERIORITY of a new drug over a placebo (or active comparator) *Show that 95% CI does not contain 0 and p < 0.05 * *Non-significant does not indicate the two treatments options are similar or guarantee a lack of difference in performance shown by the therapies* Crucial to differentiate between equivelenace and lack of evidence supporting superiority

Nominal data

Classified into groups without order and no indication of relative magnitude *Named or Naming* Gender, color, death, right/left handed, physical attributes *had a stroke or not? Died or not?* Discrete data

You are the pharmacist working at Memorial Hospital. You receive the following order for a patient in the ICU: Protonix 40mg IV Q24H The formulary states that's all orders for protonix should be substituted for the formulary drug, famotidine. Open or closed formulary

Closed

You are the pharmacist working at General Hosptial. You recieve the following order for a patient in on the med-surg floor: Simvastatin 40mg PO QD at 5 pm The formulary lists a umber of expensive agents that are not allowed to be ordered. Simvastatin is not on that list Open or closed? Positive or negative?

Closed Negative

R2

Coeffiencent of determination Describes the proportion of the variation of data presented by the dependent variable explained by the independent variable Does not provide mechanistic understanding of exactly how they are related, but provides a description of how well a straight line describes the relationship between the 2 variables If the R2 value were 0.6, then 60% of the variation in the data presented by the dependent variable can be explained by the independent variable

What is a case study?

Collection and presentation of the clinical course of a particular participant or small group

Case study

Collection and presentation of the clinical course of a particular participant or small group No control or comparison group Retrospective Aka. Case report to case series Role of case control study: -identify treatments for rare disorders -early recognition of drug toxicities and teratogenicity -dietary supplement products -hypothesis generating

Confidence intervals

Commonly seems and reported as a way to estimate the population parameter in a study based on the sample data In medical literature, 95% confidence interavals are most commonly reported

Logistic regression

Commonly used in studies to associate multiple independent variables (both continuous and discrete) with a discrete (categorical) independent variable Heavily used in epidemiology studies Can generate odds ratios using any of the independent variables (very useful)

T-test

Compares the means of two groups (not the median) to determine if they are determine if they are different from one another Continuous data and parametric Needs to have at least 20-30 data points per group Tests whether or not the sets of samples are likely to come from the same population Null hypothesis: same population Alternative hypothesis: different population *p value is calculated to determine agreement with the null hypothesis* Named for the "t statistic" which is a ratio of the hypothesized value of an estimate to the actual value taking into account a standard error Compares your value to an estimated population value (or mean of other group of values your study has deemed to be the population value; "control" vs "experimental" for example) I'd the true population mean is known (rather than estimated), a Z-test is used instead Paired to unpaired *paired if one thing affects the other Unpaired is independent samples* Is one samples value likely to predict antlers sample's value

Lexicomp

Composed of multiple databases (Lexi-drugs, pediatric and neonatal Lexi-drugs, Lexi-drugs international, natural products database, etc) Includes additional clincal links (more clinical tools and choose other clincal links) Ability to locate all medications in a class Pricing is avaliable for many medications Images are avaliable for many medications Patient educations print outs are avaliable in many languages More clincal tools > other clincal links (new drugs, ASHP drug shortages, national library of medicine, orange book, FDA recalls, dangerous abbreviations)

Reported way to estimate the population parameter in a study based on the sample data:

Confidence interval

Which of the following is a way using SEM to estimate the population parameter in a study based on the data? Effect size Confidence interval P-value Beta

Confidence interval

Chi-square goodness of fit test

Considered a one-dimensional test Test heavily used in the real world in quality control in factories and with machine work Can be gathered to compare any categorical (discrete) data gathered to expected values Ex. Are coin flips fram a certain referee really fair? Not statistically significant if p>0.05

Formulary

Contains a *list of drugs* available in an institution, reflects *clinical decisions* by the P&T committee Details info on: -dosage forms -generic/brand/chemical names -strengths -ingredients of combo products When well designed, a formulary can guide clinicians to order the safest and most cost effective medications May be divided into tiers, which differ in co-payment, reimbursement, or quality limits Considers: -dosing schedule -physician demand -storage requirements -hazardous handling risk -estimated frequency of use -convince to staff -compliance of patients -generic availability

A website poll is an example of which type of sampling? Random Convience Systematic

Convience

Correlation vs regression

Correlation describes the strength of association between data (descriptive, calculates "r" or the correlation coefficent) Regression examines the ability of one variable to predict another variable (calculates "r2" of the coefficent of determination)

Why might you choose different species to test your compound?

Cost, DNA similarity, reproductive

Which of the following study designed measures exposure of interest from a random sample at a single point in time? Controlled clinical trial Cohort study Case-Controls study Cross-sectional study

Cross-sectional study

Opportunities for DIs

DICs Medical informatics positions Managed care organizations Poison control Pharmecutical industry Academia Scientific writing and communication

Interval

Data are ranked in a specific order Consistent change in magnitude between units Zero point is arbitrary Fahrenheit Continuous data

Ratio data

Data is ranked in a specific order Consistent change between units *Zero point is significant* in that it implies a lack of or the absence of what is being measures Kelvin

If you increased the sample to 15 patients (from 10) and the SD remained the same, how would the standard error of the mean change?

Decrease

A research framework

Define (what problem am i solving? What information do i need? What else do i already know? What else do i need to know?) Locate (where can i find the information I need? Which sources are best? Where can i find them?) Select (how can i best search these sources? How will i record what I find? How will I give credit to the sources?) Organize (how should I organize the information? How can i arrange it so that others understand it? What else do i need to learn?) Use ethically (how can I shore my information with others? Who is the audience? Have I included everything? How best can i make an impact? How do I use research ethically and avoid plagiarizing? What have I learned by doing this research?) Authority is constructed and contextual Information creation as a process Information has value Research as inquiry Scholarship is a conversation Searching is strategic

Why might you choose a naked mouse versus one with fur?

Dermal studies and tumor studies

Sanford Guide

Designed for infectious disease point of care Provides treatment options for infectious diseases, antimicrobial spectra of activity, adverse effects, dosing tablets, suggested duration of therapy, etc Tables and tools drop down allows for you to select an organisms name to like to preferred agents for that pathogen

Stratified randomization

Designed to achieve similarities in baseline characteristics between groups Selected characteristics are identified (age, smoking, diabetes, etc) and used to determine which groups the subjects will be assigned *prevents significant imbalances among groups*

Function of P&T subcommittees

Determine what drugs are avaliable Decide who can prescribe specific drugs/classes Writing policy related to medication use -standard order sets -clinical pathways and guidelines -automatic substitutions Quality assurance - drug utilization review Monitor and report ADRs and med errors

Informal consensus panel

Developed during a meeting of experts Often based solely on opinions

Formal consensus panel

Developed during a more structured 2.5-day meeting buy experts Similar to informal however experts often work to reach consensus

Evidence-based guideliens

Developed over months to years Includes extensive documentation of methodology and appropriate levels of evidence for recommendations Recommendations are supported by literature

Questions for a cross-sectional study

Did investigators ensure accuracy in data collection? If a survey or questionnaire was used, was it validated? We're the inclusion and exclusion criteria clearly defined and stated? Was selection of cases clearly described?

Funding bias

Did the drug company have a hefty hand in the data analysis and results/conclusion of the study? Look for independent statisticians, disclosures, etc

Observational study designs

Disease/outcome has occurred Participants are selected and followed forward or backward Only *associations* can be formed Ex. Expert opinion/editorials, case series, case-control studies, cohort studies

Skew

Distribution that has a high number of values clustered at one end with very few values spread out towards the other end, forming a tail

Limitations of secondary resources

Do not provide quick access to information (requires search proficency on the HCP's part in order to maximize efficiency; final results are dependent upon the skill of the user) Many databases and resources are avaliable by subscription only Access to full text articles may delay process

Disadvantages of cross-sectional study

Does not test effectiveness of interventions Errors in data collections Transient effects

Cell density measures

Dose-response

Both investigators and subjects are not aware of the assigned groups

Double blind

Study design where patietns recieve two formulations - one active and one control - to ensure blinding is maintained; often when two therapies are not the same

Double dummy

Some services provided by DICs

Drug information requests Pharmacy and therapeutics committee activities Publications (newsletters, journal columns, websites) Education (in-services for health professionals, students, consumers) Medication use evaluation Investigational medication control (institutional review board (IRB) activities, information for practitioners) Literature reviews and product training

History of P&T

During formularies and committees to decide them were formed during the Revolutionary War - infection based A standard hospital system drug formulary was developed for all municipal hospitals in NYC in 1868 With infection control standards gaining hold, hospitals adopted committees to oversee policies and growing medication supplies Committee originated to help hospitals meet regulatory drug therapy standards Members researched the quality and safety of products carried in their institutions Over time, committees have evolved, and have played an increasing role in the institution

Hierarchy of evidence

Editorials, expert opinions (based on beliefs come from well renounced surgeons/physicians/practitioners/pharmacists who have spent years with a certain population and have done self study and came up with their own optioning but no rigorous trial; based off of numerous observations) Case series, case report (looking back at incidence to see if we can determine how it happened; documented report about an event) Case-control studies (find a matching control where the outcome was not present to determine if a factor was cause of outcome) Cohort studies (observations of a group of people with exposure to see if outcome happened) Randomized controlled trials (based on an intervention) Systematic reviews (guidelines, reviews multiple RCT and will include all various tiers and give a grade of evidence) As you go from expert opinion to systematic review the grade of evidence increases

Responsibilities of pharmacists regarding provisions of medication information have changed substantially

Educating and training pharmacists to adequately inform those who consume, prescribe, dispense, and administer drugs with the *goal of improving patient outcomes through the use of medical literature to support choices* Pharmacists now frequently answer DI questions, participate in evaluating a patient's drug therapy, and conduct medication use evaluation activities (community settings: counsel patients, answer medication information questions, review patient medication regimens for potential problems, manage patients' chronic diseases, etc)

Single blinding

Either investigators OR subjects are aware of the assigned groups

Watch timing bias

End of quarter surveys

AAALAC international promotes humane treatment of animals

Endorses the use of animals to advance medicine/science when there are *no non-animal alternatives*

Reject the null hypothesis

Establish that differences between groups in question cannot be explained by chance alone

Accept the null hypothesis

Establish that differences between groups in question could be related to chance alone

Rota Rod performance test

Evaluate balance, coordination, physcial condition, and motor planning Test Parkinson's medications Sleep aids may make the mice fall asleep and fall off sooner

Random sampling

Every member of the population has the *same* likelihood of being included in the sample Different subtypes of random sampling Important to define what the population of interest is before identifying which type of sampling Often difficult to do because there is always some problem *every member of the population* has equal chance of being in sample

Clinical research overview

Every new drug formulation used in a clinical trial has to first undergo rigorous laboratory tests Once results confirm the formulation is safe to be taken by humans, the drug is given to *healthy volunteers& in what are called *phase 1 clinical trials* Total process 12-15 years 1000 compounds identified by a drug company > 30 promising results > 3 clinical trials > 1 market

Systematic sampling

Every nth data element is selected as the sample from the entire possible population Requires that the list of the possible populations not be ordered in a way that induces non-random factors into the sampling Ex. For a follow-up survey, every 7th register receipt at Joe's Drug Store has a code at the bottom that the customer can use to evaluate their service

Phase 0 *

Exploratory study involving very limited human exposure to the drug, with no Therapeutic or diagnostic goals Very limited research hand they may not even come out with a drug

Clinical meaningfulness is related to

External validity BP medication only lowers 2mmhg > not too meaningful Extending life for a week but costs a lot > not too meaningful

Good resources to use for drug shortages

FDA ASHP Manufacturer

A type 2 error is another way to describe which type of result? True positive False positive True negative False negative

False negative

Reproductive toxicity

Fertility Pregnancy Birth defects Nursing Cognitive defects Coincide with phase 2 to support the clinic

Survival analysis

First used to evaluate efficacy of treatments for cancers and other diseases with high immediate mortality (with mortality as primary outcome) If the patient is still alive at the end of the study, they are counted as successful In reality, many studies have patients who leave the trial for reasons other than the primary outcome -lose contact with researchers -move away -voluntarily leave study With intention to treat analysis, all patient data is included in study analysis, even those patients who left the study without experiencing the primary outcome Solution became to "censor" data for these patietns to avoid attributing thier lack of data as a success or failure Because of the lack of data, researchers are unaware of the patient's survival

A study is examining the mortality benefit of two drugs. If the researcher's goal is to determine if the two drugs have a statistically significant difference in mortality outcomes, which test should be used? Survived. Expired A. 20. 3 B. 13. 10

Fishers exact test Because A expired has less than 5

Natural medicines

Formerly "natural standard" and "natural medicines comprehensive database" Resource avaliable on dietary supplements, natural medicines, and complementary/alternative therapies Features include monographs on natural ingredients, evidenced based information, interaction checker, pregnancy and lactation, etc

Non-inferiority trial

Frequently seen in active-controlled trails Showing experimental is not (much) worse than control Goal is to determine if a *test therapy is no worse than a standard therapy using a predefined marginal value for a primary outcome of interset* Need test therapy, standard therapy, predefined marginal value (effect size), and primary outcome of interest Adds more importance to improvement in secondary outcomes than in a traditional superiority may carry less weight Results can be expanded or extended into testing for superiority once noninferioirty is established Must establish an upper limit of the difference in treatment outcomes (maximum amount of difference allowable between groups) A detectable difference in an outcome measure between the test and active comparator treatments would be defines as delta NI

Journal watch

From publishers from the New England journal of medicine Summarizes important research and guidelines (highlights primary literature of high clinical revelance; provides a quick summary of recent articles)

Phase 3 *

Gather more info about safety and effectiveness Risk vs benefit 100s-1000s of patients

Tertiary literature

Generally well established information gathered from primary and secondary literature Textbooks, reference books, databases, monographs, review articles, guidelines, etc Confident and easily accessible Lag time between publications, author biased, and information may be incomplete General: clinical pharmacology, drug facts and comparisons, Lexicomp, micromedex, physcians' desk reference (PDR) Specialty: Shockley;s drug interactions, meyler's side effects of drugs, medications and mothers' milk, the Harriet lane handbook, natural medicines, Sanford guide, trissel's handbook and injectable drugs

Purpose of journal clubs

Goal is to explain key concepts of the article and why the articles *does to noes not matter* Recall step 5 off the modified systematic approach -evaluate: value of the information, along with the quality and reliability of the resource -analyze: interpret data into clinical outcomes -synthesize: transmit avaliable information into a logical and concise response *professional responsibility* to stay up to date on the literature

What does GRADE stand for

Grading of Recommendations, Assessment, Development, and Evaluation

GRADE

Grading of *R*ecommendations *A*ssessment, *D*evelopment and *E*valuation Created to standardize literature rating process and determination of appropriate level of evidence Literature is catagorized as *high, moderate, low, or very low* based upon study design, inconsistent results, indirectness of evidence, bias, etc and additionally as *strong or weak* depending on level of support American college of chest physcians (ACCP), world health organization (WHO), centers for disease control (CDC), etc have adopted GRADE approach

Composite endpoints

Group of endpoint measures into one Usually refers to clinical outcomes related to direct morbidity and mortality (reduction in stroke, MI, and overall death) Rationale is to measure the whole overall effect of therapy, as one specific outcome may not be deemed most important

HIPAA

Health Insurance Portability and Accountability Act of 1996 Security standards intended to protect the security of the environment in which health care information is maintained and transmitted (common identifiers (name, address, birth date, social security number), Vivian and criminal penalties exist for violating patient confidentiality Allows disclosure of patient information for treatment, payment, and health care operations Health care operations include quality management, outcomes evaluation, development of clinical guidelines, peer-review, and credentialing

Biostatistics

Help to design and interpret trials to best answer essential question

Measures of dispersion (or variability)

How are the data different Range, interquartile range (IQR), standard deviation (SD) Catagory of descriptive statistics Describes how closely (or not closely) the data cluster around the measure of central tendency

Journal club presentation strengths and limitations *preceptors love*

Identify strengths and limitations in bulletin format -do not only include the investigators assessments -think outside the ox - how would you have performed the trial? Consider internal and external validity

In non-inferiority studies, it is important to consider whether investigators recruited an adequate sample size to ensure they will not miss a small, but important, treatment effects

If the sample size is inadequate, the investigator runs the risk of a beta error > conclude treatments are equivalent when, in fact, patents given standard therapy derive benefits

Fair use

In US copyright now The doctrine that's brief excepts of copyright material may, under certain circumstances, be automated verbatim for purposes such as criticism, news reporting, teaching, and research, without the need for permission from or payment to the copyright holder

Effective provider skills

In addition to standard curricula, pharmacists working in a DI setting must be investigators Perceive, assess, and evaluate medication information and retrieve, evaluate, communicate, and apply data from published literature Lifelong learner and problem solver

In vitro

In glass Inside a test tube or Petri dish Cell tissue/culture Mutagenesis assays Carcinogenesis assays Proliferation

In vivo

In life Studies performed inside a living organism Rodents Rabbits Dogs Porcine Primates Bovine

Comparative effectiveness research

In the development of medical treatments, it is becoming *increasingly difficult* to develop more powerful drugs Pharmaceutical companies are looking for new treatments that approximate the same efficacy but *demonstrate better quality* in other aspects Trend stimulated to push for *comparative effectiveness research (CER)* A rigorous evaluation of the impact of different options that are avaliable for treating a given medical condition for a particular set of pateitns *whihc treatment of all the options is the best one for pateitns* Theoretically puts treatment options in real-world scenarios and uses different mechanisms for evaluating efficacy and appropriateness -cost may be included -treatments may be different in nature

Journal club presentation secondary endpoint

Include any pertinent secondary endpoints that were measured.

Parametric tests

Include more data and by definition are more powerful; tests But there are certain requirements that must be met Most have non-parametric tests that are analogs (meaning they evaluate the data in the same manner)

ANOVA test

Including more than 2 variables Discrete independent variable(s) Continuous dependent variable Paired or unpaired Parametric or nonparametric

You are conducting a study to compare the antipyretic of 3 drugs in 100 different people with the same infection (X, Y, and Z) What test do you use?

Independent variables (discrete, nominal, 3 variables) -drug X, Y, and Z Dependent variable (continuous, 100 samples) -time to first afebrile temp Parametric, unpaired test test for more than 1 variables is required Use anova

You are conducting a study to compare the antihypertensive efficay of 3 drugs in 100 different people (A, B, and C) What test do you use?

Independent variables (discrete, nominal, 3 variables) - drug A, B, and C Dependent variables (continuous, 100 samples) - blood pressure change (before-after) Parametric, paired test for more than 2 variables Repeated measures ANOVA test "AKA complete randomized block"

Student are classified into two groups: those who are wearing red and those are not. We would like to determine if those students are more likely to pass the test that day. what is the independent variable? What is the dependent variable? What test do we use

Independent-red Dependent-passing Use chi-square test off independence or fisher extract test if you have less than 5 in the study Set up a table with dependent variable and totals by independent variable and total

Case control study advantages

Inexpensive Shorter time-frame Smaller sample size

Independent variable

Intervention chosen by the researcher before the Trudy begins Presumed to be the cause of some effect

Major limitation of Kaplan-Meier

Investigates only one variable (group in a trial) Was this the only variable seemingly responsible for the outcome (mortality) - could there have been other factors involved Is there a method for including multiple variables?

IND

Investigational new drug

Main players in P&T committees

Investigational review board Medical executives Pharmacy

Open-label blinding (no blinding)

Investigators and subjects are aware of the assigned groups End of life or orphan drugs

Transforming data

Investigators may decide to transform the data to a lower type of data *more specific to less specific* Interval and ratio may be categorized into ordinal or nominal, but nominal cannot be reported as ordinal, interval, or ratio -we classify temperature greater than or equal to 100.4F as febrile and anything less nonfebrile...but we cannot go backward and conclude that all febrile patients have a temperature of 101F -expected high temperatures can be transformed into 60s (ordinal) or warm (nominal) -a survey using 1-5 scale is ordinal and can be transformed into nonnegative or dissatisfied ... transformed ordinal to nominal

Incorporating guidelines into practice

Involves communication of recommendations, education of healthcare professionals, and overcoming barriers to change Institutions encouraged to have guideline-driven practices Reimbursement may be a motivator Regulatory bodies (ex. Joint commission, centered for medicate and medicaid services)

Hospital practitioner asked to provide information that might be used to speed ending of a terminal patient's life

Is that what the patient wants? Is it legal?

Systematic approach to answering drug information questions

It is a checklist used to ensure quality responses in an efficent ,manner Is is applicable to almost every tuner of questions encountered in clinical practice Developed in 1975 and included five steps - modified in 1987 to include 2 additional steps (7 steps total)

Uses survival time (often censored) to estimate the proportion of people who would survive a certain length of time under certain circumstances

Kaplan-Meier

Pubmed also contains

Links to many sites providing full text articles Provides advanced search, clincal queries search filters, and special queries pages Links to related articles that may be of interest Automatic e-mailing of search updates and the ability to save records and filters for search results

Positive formulary

List of agents allowed to be dispensed in an organization

Negative formulary

List of agents which are NOT available for dispensing, usually because they are overly expensive

Facts and comparisons special features

List of medications which should not be crushed or chewed can be found under the "don't crush/chew" tab Creative clearance estimates by the Cockcroft-Gault equation can be found in interactive tools within the medicinal section of the clinical calculators tab Look alike/sound alike medications are located under the look alike/sound alike tab Information about safety recalls and new medications is located under the medications safety programs/REMS tab

Absolute risk reduction becomes smaller when event rates are

Low, even if relative risk remains constant A condition has an adverse effect rate of 50% in the general population, and a new wonder drugs take that down to 17% Relative risk reduction = 66% (50-17)/50 Absolute risk reduction = 33% (50-17) A condition with an adverse effect of 3% in the general population may experience a reduction to 1% with a new drug RRR=66% (3-1)/3 ARR=2

Objectives of P&T comittee*

Make prescribing practices cost efficient Assess clinical outcomes Improve patient safety Monitor: -adverse drug reactions -medication error monitoring -quality assurance -policy and procedures -formulary approval -wholesaler contracts -medicaid/medicare reimbursements -insurance company contracts

How does survival analysis fit in?

Many studies are designed to examine an ovtcome over the course of time Adds depth and dimension to study design In interpretation of study results, kept in mind that the effects of the intervention may not be realized until a certain time into the study (when the lines seperate)

Pharmacist assists in identification of tablets or capsules found without appropriate labeling

May be inappropriate to identify medications fora third party Considerations also apply to legal guardians of children, depending upon child's age, disease state, or medication in question

Barriers to guideline use

May not be updated regularly and may become outdated Guidelines published bu different organizations for a disease state may have conflicting information Changed to public policy Reminder to consider individual patient

For continuous data (normal or near normal) -best measure of central tendency: -Best measure of dispersion:

Mean Standard deviation

450 patients in a sample, mean SBP 115mmHg SD=5.4 mmHg Calculate 95% confidence interval

Mean +/- (1.96 x SEM) SEM = SD/square root of n SEM=5.4/square root of 450 = 0.25mmHg 95% CI=115 +/- (1.96 x 0.25) 95% CI=115+/-0.49 95% CI = 114.5-115.5

Calculating confidence intervals 95% confidence interval=

Mean plus or minus (1.96 x standard error of mean) Standard error of the mean is SD/square root of population As a result, the smaller the sample size, the larger (or wider) the confidence interval

{1, 4, 8, 10, 11, 18, 23, 45, 100} Mean: Median: Q1: Q3: IQR: Outliers:

Mean: (1+4+8+10+11+18+23+45+100)/9 = 24.4 Median: 11 - this is also known as Q2 or 50th percentile 25th percentile: median of lower 1/2 data - (4+8)/2 = 6 75th percentile = (25+43)/2 = 34 IQR 34-6 = 28 Outliers: 1.5 x 28=42; 6-42=-36; 34+42=76; 100 is an outlier

Data set: {74, 160, 0.11, 16, 8.4, 640, 62, 8, 0.7, 0.006} Mean: Median: Mode: Range: Q1: Q3: IQR: Outliers

Mean:96.9 Median: 12.2 Mode: No Mode Range: 639.994 Q1:0.7 Q3:74 IQR:73.3 Outliers 1.5x73.3=109.95 109.95+74=183.95 .7-109.95=-109.25 640 is an outlier

Repeated measures ANOVA

Meant to analyze multiple occurancecs of the same event rather than different events together -productivity of a corn field over 3 years in bushels of corn Cannot measure bushes of corn, as well as number of earths, amount of disease present, etc -only one characteristic at a time

Surrogate endpoints

Measure of treatment effect that may correlate to clinical endpoint Ex. Cholesterol levels, risk for fracture

Correlation ( or covariation)

Measures the strength of relationship between two variables, or the degree to which one variable changes with the other changes Tests vary based on type of data (Pearson vs spearman)

For discrete data (ordinal, but could be nominal) Best measure of central tendency: Best measure of dispersion:

Medical Interquartile range

MeSH

Medical Subject Headings NLM's controlled vocabulary thesaurus Descriptors are arranged in both alphabetic and hierarchical structure Continually revised Used to index articles into secondary database Retrieved from the *title and summary of the articles* Developed from medical terminologies Defines and customizes search strategy Different indexing and abstracting databases use different indexing terminologies

Statistics

Method for classifying, summarizing, and analyzing data

Case control disadvantages

Methodological issues (selection of controls; level of exposure) Retrospective data Temporal relationship

Legal aspects of drug information

Minimal standards of performance and a consistent level of competence must be assured when providing DI in the daily practice of pharmecutical care Failure to counsel or warn cases are an increasing trend in pharmacist liability Liable for incomplete information, inappropriate quality information, outdated information, inappropriate analysis/dissemination of information

Effect size

Minimum clinically relevant difference (determine a priori meaning a type of knowledge that agrees with reason and often obtained outside of experience) Selected or examined based on clincal judgement and previous literature Also involves the number or patietns The smaller the effect size that is *clinically important* the *greater* number of subjects needed to find a difference should one truly exist Bigger change can use smaller sample size The smaller the difference you are attempting to detect, the greater number of subjects needed in the sample

Bets measure of central tendency for nominal data

Mode Really not a best measure of dispersion here

Recall bias

More likely to make associations related to negative events Ex. Women diagnosed with breast cancer were asked to recall childhood diets compared to age-matched groups. All of the women's had actually completed a dietary survey years earlier before any diagnosis. Women with breast cancer recalled a diet that was much higher in fat than what they actually consumed

The Evolving P&T

More pharmacists are active, voting members Other clinicians like RN;s and respiratory therapists may rotate on Non-clinicians also have a stake: risk management, medical records, ethicists Decisions may or may not be final and implemented until approved by the institutions medical executive committee

Mode

Most common value in distribution Can be used for ordinal, continuous, nominal *best tendency for nominal data* There can be more than one mode. (Bimodal, trimodal, etc)

Normal (Gaussian) distribution

Most commonly used and referred to is the "normal" distribution Bell-shaped curve Also referred to as the Gaussian Distribution or Gaussian population Most values occur in the middle with fewer at each extreme When measuring a random variable in a large enough population, some samples appear more frequently than others When graphed on a histogram (data on horizontal axis, amount of data on vertical axis), resembles a bell *by the empirical rule: -68% are found within mean plus or minus one standard deviation -95% are found itching plus or minus 2 standard deviations -99% of the data are found within plus or minus 3 standard deviations To completely define a normally distributed data set you need the range, standard deviation, and mean*

What has a normal distribution?

Most natural phenomena or really most measurements when measured in enough patients (usually 30+) -height -blood pressure -shoe size -arm length A quick way to tell is to plot the data in a histogram

Non-formulary prescribing

Must be allowed when medications on the formulary are required -patients with rare illnesses -true allergies to formulary medications -intolerable or non-response to formulary medications -stable patient on a non-formulary critical care medication Institutions can be funded, or face non-reimbursement for inappropriate non-formulary use Should require a second signature, department approval, pharmacist consultation, etc

A study is evaluating diabetes control in a group of patients. The current sample size is 150 patients. If the number of subjects in the trial were increased to 450, how would the confidence interval change? Narrow Widen Remain the same/no change

Narrow

Sampling

Nearly impossible in most populations to sample every member of the population (as i nsurveys of the public or opinion polls Samples required to estimate the actual value Ex. Random, systematic, convience

Examples of journals

New England journal of medicine Clinical infectious diseases The journal of clinical pharmacology Pharmacotherapy

For chart: how do you determine no if given yes and total How do you get the percentages needed for RRR and stuff

No=total-yes %=yes/total

A study you are reading groups subjects by the area of the city in which they currently live. The areas are east, west, north, and south. Which type of data does this represent?

Nominal

Study design used to determine if a test therapy is no worse than a standard therapy using a predefined marginal value for a primary outcome of interest

Non-inferiority

For skewed distribution, interquartile range is more reflective of central tendency, and this type of tests are used to analyze

Non-parametric

Creative Commons

Non-profit organization headquartered in Mountain View, California, US Devoted to expanding the range of creative works available for others to build upon legally and to share

A researcher is attempting to establish a relationship between blood pressure and atmospheric pressure. What's the independent variable? What's the dependent variable? How many variables? Parametric or nonparametric?

None BP 2 Parametric

For skewed distribution, mean and median are

Not likely close (depending on degree of skew) Interquartile range more reflective of central tendency Use non-parametric tests to analyze

A study is attempting to establish a link between eating red meat and development of a certain type of cancer. The alpha value set before the study was 0.05. If the p-value calculated was 0.24, how should it be interpreted? Strongly associated Not adequately powered Not statistically significant Statistically significant

Not statistically significant

Equivalence hypothesis testing statements *

Null: difference between groups Alternative: no difference between groups

Superiority hypothesis testing statements *

Null: no difference between groups Alternative: difference between groups

Noninferiority hypothesis testing statements *

Null: treatment fails to show that it is no worse than active comparator -detectable difference in outcomes is greater than maximum allowable difference in outcomes Alternative: treatment is no worse than active comparator -detectable difference in outcomes is less than maximum allowable difference in outcomes When the upper limit of 95% CI of the detectable difference in outcomes the test treatment is less than the maximum allowable difference in outcomes is considered non inferior - relies on a one-sided test for the outcome If the upper limit of the 95% CI of detectable difference in outcomes is greater than maximum allowable difference in outcomes the test group is not non inferior to the active comparator

Number needed to harm (NNH)

Number of patients who would need to be treated to cause one adverse outcome Inverse of the absolute risk increase Drug X has a mortality rate of 27% and the control has a mortality rate of 15% (ARI=12%) NNH=1/0.12=8.3 patients You need to treat 9 patients to have one adverse outcome

An asthma study is evaluating the number of times a rescue inhaler (SABA) is being used per month by patients enrolled in the study. A patient uses her inhaler 2 times per month. Another one uses his 5 times per month. What type of data are they generating?

Number of uses do SABA can take up any value in the range of possible values Change from one value to the next a consistent change in magnitude The value of zero is the absence of use of SABA Ratio

Ad hock sub-committees of P&T

Nutrition Infectious disease Medical devices Medication safety Anti-coagulation

Observational or interventional

Observational -no intervention -associations -prospective or retrospective -rare exposures or treatment Interventional -intervention -cause and effect -prospective -mainstay therapy

Cohort study

Observing a group of people Disease free subject population and divided into groups whether they have been exposed to the factor and those who were not Subjects followed prospectively as disease state of interest is observed Aka follow-up, longitudinal, or incidence study Examples: -European prospective investigation into cancer (EPIC) -British doctors study -aplastic anemia with use of clozapine -asbestos exposure on lung cancer Measures of association -relative risk (gives an idea of the magnitude of an effect) -confidence interval (provides range in which true value for population lies; wider the confidence intervals less precise the results) -relative risk of 1 indicates no difference exists between groups; therefore, CI cannot include 1 and still maintain statistical significance

Why would you choose a rabbit for studies?

Ocular studies and if you need a lot of bloods samples becUAse they have good veins in their ears

Odds ratio

Odds of being exposed in cases (A/C)divided by odds of being exposed in controls (b/d) Outcome Exposure Event. No event Treated(exposed). A. B Control (not exposed) c. D (OR) = (a/c)/(b/d)

Risk

Often evaluated to determine benefit of treatment or drug Implies associated with prevention of negative outcome or association with a positive outcome (cure) How does the risk of a disease or a side effect compare betwee ngroups in a study *(relative risk)* or compared to nothing at all *(absolute risk)* -absolute risk (AR) -absolute risk reduction (ARR) -relative risk (RR) Relative risk reduction (RRR)

Clinical endpoints

Often refers to occurrence of disease or event Often holds more weight than surrogate Ex. Death, time of survival, heart attack, stroke, number of asthma attacks, fracture

M&Ms claims on the package to have 30% yellow candies, 25% brown, 25% green, 10%red, and 10% blue candies, you pull a random bag to test and find 34 pieces: 5 blue, 6 red, 7 green, 8 brown, 8 yellow candies. How can you test their theory? How many variables? Why are these discrete data?

One variable Discrete variable because it is nominal (just names) Chi-square goodness of fit 0.05 is set as level of significance (alpha) P did not equal 0.05 so there is less than an established chance that the variation in the back is related to the expected variation laid out

Per-protocol analysis

Only those subjects meeting the inclusion criteria and following protocol are included May lead to biased conclusions because dropouts and non-adherent participants from the two group potentially differ

Investigators and subjects are aware of the assigned groups

Open-label (no blinding)

Broadest and least Boolean operator

Or is broadest And is least because it has to have both

A researcher is examining the strength of hurricanes that hit the Atlantic coast in 2007 as measured by the Saffir-Simpson scale (category 1 - category 5). Wheat is the best measure of dispersion for this type of data?

Ordinal data set Best measure of central tendency: median Best measure for dispersion: interquartile range

Dependent variable

Outcomes measured at the end of the study Result of outcomes, or effects, of the independent variable Data are observed values and may be considered outcomes of an experiement

Calculated after the study to determine how well the collected data correlate with the null hypothesis (related to alpha)

P-value

Paired vs unpaired Researcher checks blood glucose before and after administration of investigational drug

Paired because the patients glucose value prior to administration is likely to predict the value after administration

Patients are divided into two groups in a blood pressure medication trial. Their MAP is measured before and after administration of the drug. Independent variable? Dependent variable? How many variables? Parametric requirements? Paired or unpaired? Test we use?

Paired t-test

This type of data assumes certain characteristics such as normally distributed, randomly selected samples from the population, unpaired, matched, or correlated, and equal variances within groups

Parametric data

Cumulative Index to Nursing and Allied Health Literature (CINAHL)

Part of EbscoHost Suite database Resource for nursing and allied health professionals Provides indexing for over 5492 journals with full text avaliable from 1981 to present Includes nursing conference proceedings, educational software, and health related books

Interventional study designs

Participants assigned to a group to recieve treatment of placebo (or active control) Best at establishing *cause-and-effect* Ex. Randomized controlled trials

HIPPA allows disclosure of

Patient information for treatment payment, quality management, outcomes evaluation, development of clinical guidelines, peer-review, and crednetialing

Pharmacist may recieve a question from a patient seeking to determine whether a medication is appropriate treatment for a particular medical condition

Patient may use information to validate or refute the recommendations of their healthcare provider

Consequences of plagiarism

Penalties for a violation of the college of pharmacy honor code for academic misconduct include but are not limited to a simple warning, a failing grade on an examination, a failing grade in a course, professional performance plan In most egregious cases, suspension or expulsion from the college of pharmacy could be imposed. If a violation is found to have taken place, a record of the proceedings shall be kept in the students academic file Outside of school you could lose your job and/or go to jail

Quality improvement within P&T

Perform drug use evaluations (DUE) Identify adverse drug events (ADE) Contain costs Expand clinical pharmacy opportunities Identify trends Regular reporting to medical staff Internal incident reporting strategy (forms/online) track all unusually events regarding patients Often done by tracking use of "tracer drugs" - meds used to treat the side effects of other drugs -epinephrine for allergic reactions -vitamin K for warfarin overdoses -flumazenil for benzodiazepine overdoses Shortages should be continually monitored (a joint commission requirement) -FDA has a website to list current known shortages -ASHP publishes guideline on how to deal with shortages -the drug manufacturer's website and hotline may be of use Appropriate providers should be notified (physicians, drug info center, pharmacy director, buyer, chief of medicine, ethics committee) Should be a standard agenda item at every P&T meeting

Intent to treat analysis

Performed by including all of the data, irrespective of any protocol violations, patient dropouts, reduced number of patient visits, and crossovers across treatment options May lead to biased conclusions because of protocol violators and withdrawals

Types of secondary resources

Periodicals/Journals -Clin-alert -cancer today -current contents connect -reactions weekly -Journal watch -*Journals* (ex NEJM) Databases -Cochrane library -web of science -scopus -embase -google scholar -Iowa drug information services (IDIS) -*Medline/PubMed* -*CINAHL* -*IPA*

What is a P & T

Pharmacy and Theraputics committee Oversees all aspects of drug therapy in an institution (hospital clinic, insurance plan, etc) Members include physicians, pharmacist, and administrators May form sub-committees to address focused issues

International Pharmaceutical Abstracts (IPA)

Pharmacy specific Originally started by ASHP Include *about 800 journals* *open vocabualry search format* Links to full text articles Includes majority of pharmacy specific abstracts, conference presentations, and other grey literature

Probability of detecting a difference if one actually exists; measurement of a study's ability to prevent a type II error:

Power

Factors influencing the evolution of the pharmacist;s role as a medication information provider

Prevention of adverse drug events (ADEs) Growth of information technology Focus on evidenced-based medicine and outcomes Advancement of medication therapy Knowledgeable patients

Drug information

Printed information in a reference or verbalized by an individual that pertains to the medication Any information about the drug and is often referenced (printed or verbalized) Often provided by specialists, drug infocenters, drug services, etc Term originated in early 1960s in conjunction with the first drug information center at the university of Kentucky medical center (source of drug information for staff physcian and dentists to evaluate and compare drugs)

Alpha level

Probability established (before the study (a priori)* of falsely rejecting the null hypothesis when it is really true *risk of a false positive result Usually set as 0.05 (5%; or 1 out of 20)* Type 1 data Can be one sided or two sided (positive and negative effect) One-sided alphas only attempts to detect if the differences the sample is *either* higher or lower than the predicted difference *(but not both)* A two sided alpha is designed to detect bath a higher and lower difference (stronger test in most study designs, but in some studies only one side of a difference in relevant to the researchers

Beta level

Probability established *before the study (a priori)* of falsely accepting the null hypothesis when it is actually false *risk of a false negative result Usually set at 0.2 (20%)* Type 2 error Relates to the chance that no difference will be detected when one actually exists If we said that increasing the amount of samplestake allows for a higher likelihood of accurately detecting the difference *directly related to the POWER of the study and as a result, the SAMPLE SIZE*

Ethical aspects of drug information

Professional responsibity to make moral judgements in situations where action must be taken despite uncertainty (oath of a pharmacist) Pharmacists called upon to provide drug information many encounter ethical dilemmas regarding what information, if any, should be provided Identify pertinent facts Analyze relevant points for the situation Rank or balance pertinent ethical rules and principals involved Recognize rights and responsibilities to yourself, the client, other involved individuals, and the society as a whole

Radioactive Thymidine measures

Proliferation

Relative risk reduction

Proportional reduction IV events between control and treatment groups Calculated as relative risk reduction =[control group-treatment group]/control group Or 1-risk reduction Ex. Drug X results in a 14% MI rate whereas the placebo results in a 25% MI rate RRR=(25-14)/25=.44 ot 44% risk reduction

Alternative Hypothesis (Ha)

Proposes that there is a difference between the two groups in a question above that which could have occurred by chance alone Opposite of null hypothesis *does not offer information about the type or magnitude of any difference between the groups*

Null hypothesis (Ho)

Proposes that there is no difference between the tow groups in question that cannot be explained by chance Used to assist in determining if any observed differences are due to chance alone Tested for statistical significance to determine whether the data collected from the sample are consistent with Ho

Controlled clinical trials ____ measure a difference in effect between two therapies

Prospectively Groups are similar and treated identically with the exception of the treatment being studied Subjects are assigned to the groups and monitored

Journal club presentation Other results of interest

Provide ant other results that are important Ex. Number of patients completing the trial, % of patients requiring additional therapies, mean medication dose at study conclusion, etc.

Journal club presentation Adverse drug effects/reactions

Provide any significant adverse effects (quantify) -severe bleeding occurred in 4 patients in combination group (0.2%) and 4 patients in the aspirin group (0.2%) (HR 0.94; 95% CI 0.24 - 3.79; p=0.94) -moderate bleeding occurred in 3 patient in combination group (0.1%), and 4 patients in aspirin group (0.2%) (HR 0.73, 95% CI 0.16 - 3.26; p=0.68)

Journal club presentation Primary endpoint results

Provide specific endpoints, including p-values Include percent and Rangel with confidence intervals -be able to explain and interpret results -does range include value of equality? Calculate NNT and/or NNH May include figures/graphs, just be sure to cite appropriately

Google scholar

Provides access to *full-texts journals, preprint, theses, books, abstracts, etc* *open vocabualry* Broad coverage including multiple languages and GREY LITREATURE Search results are often less accurate and less frequently updated Results displayed in related to the number of visits by users

Abstract

Provides bibliographic citation in addition to a brief description of information provided by the article

Index

Provides bibliographic citation information for easy retrieval Author, title, and citation of article

Clin-Alert

Provides health care providers with summaries of *adverse drug reactions, drug interaction, and market withdrawals* from over 100 key medical and research journals from around the world

Internal validity

Quality of study Flaws within the design reduce this and result in limited external validity

Multitude of study designs include

Randomized controlled trials, oversvations trials, laboratory experiments, epidemiological and outcomes research, etc Typically fall into two main categories (interventional and observational)

Why use observational study design?

Randomized experiment would violate ethical standards Rare exposures or treatments Epidemiological research

Ordinal data

Ranked in a specific order but no consistent level of magnitude in difference between ranks *ordered* McDonald's value menu Students classified into groups based on order of entry into classroom Distance from the center of city

Reclassifying a students numerical grade on an exam (0-100 scale) into 2 groups ("pass" or "fail) based on a score of >70 on the exam is an example of transforming data from ___ to ___

Ratio to nominal

At its core, the systematic approach is addressing four important and distinct actions of answering drug information request

Receive Research Respond Record

Advanced PubMed search

Refines the search Tailored to the researchers need Results are specific Maximizes the efficiency of researching *when performing a drug/medication search, search by the generic instead of brand name*

Coefficient of variation (CV)

Relates the mean and the standard deviation Expresses the variation as a percentage of the mean Removes units nad allows comparison of mean and SD across different sets of data *CV=SD/mean* Can also be given as a percentage SD/mean X100 Compares various entities because we remove the units

What is the most appropriate test to measure the strength of relationship between a discrete independent variable and a continuous dependent variable, with MORE THAN 2 levels of and PAIRED data?

Repeated measures ANOVA (complete randomized block)

Journal club presentation Secondary endpoint results

Report any worthwhile secondary endpoint results, following same format as primary results (p-values, 95% CI, etc)

Why would you choose a mouse for studies?

Reproductive studies

Drug information specialists

Responsible for the operation of drug information centers Retrieved, selected, evaluated, and disseminated information

Journal club presentation null hypothesis

Review null vs alternative hypothesis Watch for non-inferiority chains versus superiority

Suppose that a study only has 200 patients instead of 450 patients. SD = 5.4 mmHg 115 mean Calculate 95% confidence interval

SEM=SD/square root of n SEM=5.4/square root of 200 = 0.38 mmHg 95% CI = mean +/-(1.96xSEM) 95% CI = 115 +/- (1.96x0.38) 95% CI = 115 +/- 0.7448 95% CI = 114.2-115.7

What is a control?

Sample group where the questioned factor is absent or held constant, in order to provide a comparison

Basic PubMed search

Searches the entire database Expands search results No limits or preferences are used Enter search terms Click search to generate results Read titles/abstracts to select choices for review and retrieval

The institutional animal care and use committee (IACUC)

Self-regulating entitity that, according to US federal law must be established by institutions that use laboratory animals for research or institutional purposes to oversee and evaluate all aspects of the institutions animal care and use program

Uses of secondary resources

Serve as gateway to primary and tertiary Make literature searching more efficent Help locate publication of choice Enable customized researching Assist in finding articles or full citations through specialized features

Researchers want to know the characteristics of distributions, such as

Shape The most common value (central tendency) How spread out the values are (variability, dispersion)

How to choose an article

Should be *primary research and preferred to be published within the last year* Consider study designs -prospective trials (interventional or observational) -retrospective studies (case control, observational) Consider searching for an article from a journal with a high impact factor Ensure you have access to supplementary material, if applicable *understand the study questions, primary endpoint, study design and purpose of the article before proceeding* Follow AMA citation (may refer to trial throughout presentation by acronym; abbreviate and italicize journal names using NLM abbreviations) Funding source (remember to assess bias - role of authors, disclosures, independent statisticians, etc

Parametric data assume certain characteristics

Should be interval ratio and normally distributed Samples must be randomly selected from population and independently measured Data should not be paired, matched, or correlated Variances within the groups compared are equal Ideal data set is normally distributed If a sample of continuous data has over 30-50 data points, cna generally be assumed to have a normal distribution

Equivalence trial

Showing bioequivelence Showing experimental is equivalent to control Aims to compare an outcome measure between two interventions, assessing equivalence (within a specific range) with respect to predefine clinical criteria Ideally the patient will not detect any change in effect when replacing on drug by another *95% CI is inside the interval of therapeutic equivalence* Two therapies are equivalent if the observable difference between them lies within an established interval Ex. To establish brand/generic bioequivelence, the FDA mandates the calculated confidence interval should fall within the limit of 80-125% for ratio product averages

Either investigators or subjects are aware of the assigned groups

Single blind

PubMed Special Features

Single citation matches (used to retrieve full citation when any part of it is not avaliable) Related citations (find citations similar to those of interest) Clinical trials (used to search for information on clinical trials) Tutorials Journal in NCBI database (provides information on specific journals and journal abbreviations) Clinical queries (provides information about specific topic and searching strategy help) MeSH database (provides access to controlled vocabulary for researching)

Journal club presentation study design

State the study design, blinding, location, and randomization types

Journal club presentation investigators conclusion

State the study's conclusion (do not Copt verbatim from the study)

A study is attempting to establish a link between eating red meat and development of a colon cancer. The alpha value set before the study was 0.05. If the p-value calculated was 0.025, how should the results be interpreted?

Statistically significant

Linear regression

Straight lien tha passes through data that minimizes the sum of squared difference between the original data and the line "The least-squares regression line" Requires continuous independent and dependent variable

Type of randomization designed to achieve similarities in baseline characteristics between groups:

Stratified

Noninferioirty and equivelence trials require planning and analysis

Stress the importance of drawing a conclusion by comparing the obtained results with the pre-specified margin Use a standard vocabulary to avoid misleading conclusions Intent to treat and per protocol analyses should be performed Researchers suggest combining intent to treat and per-protocol approaches for proper interpretation of results obtained from equivelence and non-inferiority studies

Double-dummy blinding

Studies comparing different regimens, for example once daily to twice daily, will require the once daily group to take a placebo As a second dose Included when the two therapies given are not the same (ex. Different routes, frequency of administration, different formulations, etc) Patients receive two formulations - one active and one control- to ensure blinding is maintained

Selection bias

Study is only good as its sample Consider inclusion/excluding criteria - did the study only include young, healthy individuals? Often exclude young, old, and ill Consider sample selection and see if the study was truly randomized

Important to understand ___ for journal club

Study questions, primary endpoint, study design, and purpose of the article before proceeding

Observational study design

Subject group based on presence or absence of exposure or outcome No involved intervention Draws *inferences* about effect of exposure on subjects Ex. Cohort, case-control, Ross-sectional, case study

Case-control study

Subjects (cases) with outcome are recruited, matched with, and compared to similar groups of subjects Data regarding exposures are collected retrospectively and two groups are compared to identify contributors Aka case-referent, case history, retrospective, or trohoc study (cohort spelled backwards) Ex. -hormone replacement therapy in relation to breast cancer -cellular phone use associated with brain cancer -Sunbed exposure and risk of skin melanoma Measures of association -odds ratio (odds of having event in the exposed group relative to odds of having event in the control group) -OR is used because the event is known

Blocked randomization

Subjects have an equal probability of being assigned to a block

Summary of journal club presentations

Summarize key points in application and importance Remember to cite other references used within the presentation Expect about 5-10 minutes for questions Review grading rubric for grading criteria Practice presentation before hand

Guidelines

Systematic review of all evidence providing an assessment of benefits and harms of care options Assist healthcare practitioners in making informed evidence based medicine decisions Set the standard of care Three types for development: informal consensus panel, formal consensus panel, evidence-based guidelines

Negative skew

Tail towards the left "Left skew"

Positive skew

Tail towards the right "Right skew"

Phase 2 *

Test effectiveness of drug for a particular indication Several hundred people Common side effects and risks associated with the drug Typically well-controlled, closely monitored, and conducted in a relatively small number of patients

Log-rank test

Test to compare two survival curves to determine statistical significance between study groups -null hypothesis: survival is identical between the groups -alternative: survival is not identical between groups (generates p-value (because remember what a p-value is) -samples are random -subjects are not paired or matched (independent) -consistent criteria for entry and exit throughout -baseline survival rate does not change over time -censored subjects have same survival as non censored

No matter what you are trying to the null must be

That you are wrong Hence why it needs to be proven

P-value

The acceptable rate of error (type 1 error) is 1 in 20 or 5 times in 100 Leading to a value of 0.05 being significant Significance is not relevant, it is a constant A value of 0.02 is *no more significant* than a p value of 0.044

DICs and DI specialist set the stage for the beginning of

The clinical pharmacy concept Pharmacists demonstrated the ability to assume more responsibility in providing input on patient drug therapy Key tools was the *evaluations and interpretation* of clinical drug literature

Evidenced-Based Medicine (EBM)

The conscientious explicit and judicious use of current best evidence in making clinical decisions about the care of the individual patients Includes integrating clinical experience with best available external clinical evidence from systematic research Professional responsibility to stay up-to-date on the literature and current practices Clinical experience + references

Absolute risk reduction

The difference in the event rate between control group (CER) and treatment group (EER) ARR=CER-EER Ex. Drug X results in a 14% MI rate whereas he placebo results in a 25% MI rate ARR=25-14=11%

Proportional hazards assumption

The hazard is defined as the instantaneous potential per unit time for a person to experience the event of event of interest, given that the person survives up to that point Assumption states that *the hazard is proportional to that for any other individual and is independent of time*

Standard Deviation (SD or sigma)

The index of the degree of variability of data about the mean of that data Measure of how much the data vary from the mean Uses a formula that sums the squatters of the differences between each value from a sample and the mean of the sample *applicable to normal or near-normal distribution*

Normal distribution is the "cleanest: statistically because

The mean and median are very close

Number needed to treat (NNT)

The number of patients who need to be treated to prevent one bad outcome Inverse in absolute risk reduction Drug X results in a 14% MI rate whereas the placebo results in a 25% MI rate (ARR 11%) NNT=1/0.11=9 patients Also works for positive outcomes (inverse of the difference between the two groups) 1/(% benefiting from exp)-(%benefiting from control)

Percentile

The point or value at which a value is larger than X% of the other values -75th percentile lies where 75% of the other values are smaller Does not assume the population has a normal distribution Useful for skewed distributions

Relative risk or risk ratio

The ratio of the probability of developing, in a specified period of time, an outcome among those receiving the treatment of interest or exposed to a risk factor, *compared* with the probability of developing the outcome if the risk factor or intervention is not present The ratio of risk in the treated group to the risk in the control group RR=EER/CER *odds ratio* is a type of relative risk Ex. Drug X results in a 14% MI rate whereas the placebo results in a 25% MI rate RR=14%/25%=56% *a relative risk above 1 means the treatment is associated with the outcome, values below one mean the treatment is negatively associated with outcome* 0.56 RR means the treatment is negatively associated with the outcome of MI events (=good)

Hazard ratio

The relative risk of a complication based on a comparison of event rates between groups *proability that if the event in question has not already occurred, it will occur in the next time interval, divided by the length of that interval* In models it represents the likelihood of the outcome of interest in treatment versus control subjects at any given point in time At a given point in time (stop and evaluate at that point) what is the risk of the event if it has not already occurred? In a study, they are cumulative (the study progresses through time and subjects experience the primary event) Studies often use analyses that exclude people who have already had the event (usually) primary outcome at each interval (exclude people? How is this allowed?)

Are the groups different statistically? Group 1: sample mean 7.5 (95% CI 4-12) Group 2: sample mean 4.2 (95% CI 2-6)

These 2 groups are not statistically different from one another because they overlap from 4-6

Why would you choose a pig for studies?

They are like us for the most part

Disadvantage of cohort study

Time consuming Expensive Selection bias Information bias

Journal citations

Title is all lower except for proper nouns 1-6 authors you list them all More than 6 you list the first 3 and add et al

Changes in research

Traditionally new drugs/treatments needed to be proven superior to older treatments or the standard of care CER adds legitimacy to the idea that new drugs or treatments (or even older drugs/treatments) can be proven equivalent or non inferior to the standard of care

With the development of different effective medical treatments, there is an ethical dilemma associated with choosing a placebo as the comparator group

Trail participants of Phase III RCTs Have the condition being studied by usually are otherwise healthy May not characterize the population at large with multiple comorbidities Solution is to conduct equivalence or non-inferiority trials with populations under routine practice conditions Both types typically include a comparator group treated with reference intervention other than placebo

Both investors and subjects are not aware of the assigned groups PLUS data interpreters are not aware of the subject assignment

Triple blind

Community pharmacist requested by a a patient at the counter to critique another health care provider's recommendation

Try not to critique other healthcare providers (keep a good relationship unless a patient's health is at risk)

Factors going into the power calcuation

Type and classification of data (nominal, ordinal, etc) Alpha level desired One-sided or two sided The size (or magnitude) of difference that the study is attempting the detect (effect size)

Censoring data

Typically arises from studies in which the dependent variable of interest represents TIME to an event (either an actual event or death) and the study is time-limited Many cardiovascular studies measure time to first cardiovascular event (time to an event as the dependent variable) over a certain number of years (time-limited)

Medline

US national library of medicines (NLM) bibliographicdatabase of citation and abstracts Largest component of *pubmed*, a freely accessible online database of biomedical journal citations and abstracts created by NLM In 1997 Medline became free access to the public via pubmed interface Widely used by health care practitioners Citations from approximately *5600 worldwide journals* are currently indexed for medline 2,000 to 4,000 completed references are added weekly with 813500 total added in 2017 *excludes a majority of grey literature* Records are indexed with NLMs controlled vocabulary, MeSH Interfaced with PubMed, EbscoHost, first search, Ovid, and Cambridge

Copyright infringement

Unauthorized use of copyrighted work (3 year statue of limits) Impact (monetary cost, injunctions, criminal liability) Impact on your academic career (failing grade, failing course, expulsion)

Paired vs unpaired A researcher checks the blood glucose readings of every patient who visits the clinic on a certain day

Unpaired because one patients value is not likely to predict another patients value (in a mathematical sense)

Telemedicine

Use of telecommunications and interactive video technology to provide health care services to patients who are at a distance

Person coefficent

Used between two variables measured on continuous scale Parametric

Nonparametic requirements

Used in cases where the assumptions of parametric tests cannot be guaranteed if it is not parametric (ex. Normal distribution of data not expected, such as income, car accidents, or a rare disease) Small sample size limits the assumption of normal distribution of data (small samples cannot be appropriately tested for normal distribution; of the sample is large enough, it is safer to assume that the sampling is normal even if the data itself is not)

Elevate plus maze

Used to study anxiety medications

Marble burying test

Used to study anxiety medications Psychotropic and neurotrophic activity

Spearman coefficent

Used when one or both variables are measured on ordinal scale Nonparametric

Cox proportional hazards model

Uses proportional hazards regression to determine the effect of two or more variables on time to an event Can handle any number of covariates Used when researchers are concerned about group difference at baseline and related to a covariate measured on a continuous scale Allows researchers to evlauate survival data and adjust for confounders like age, severity of illness, etc

Kaplan-Meier

Uses survival times (often censored) to estimate the proportion of people who would survive a certain length of time under similar circumstances Most commonly seen analysis tool (and simplest) Allows for production of table and graph -table ="life table" -graph = "survival curve"

Traditional members of the P&T

Usually made up by medical staff representatives -ideally from a wide range of practice areas -leadership and chair are instrumental in success One or two pharmacists may be members, sometimes without voting privileges The director of pharmacist may be the only pharmacy representative Other staff pharmacists may act of consultants and attend intermittently Administrators like the CFO may weigh in on finance issues

With regard to a nominal endpoint Enter if the following odds ratios and 95% CIs are "significant" or " not significant" 1. OR 2.56 (95%CI 1.01-3.6) 2. OR 0.65 (95% CI 0.34-1.2) 3. OR 89.5 (95% CI 78.4-100) 4. OR 0.1 (95% CI 0.09-0.99)

VOE=1 1. Significant 2. Not significant 3. Significant 4. Significant

Value of equality (VOE)

Value at which the two groups are not different statistically *discrete data-VOE=1 Continuous data-VOE=0* Discrete data ex. RR 0.83 (95% CI 0.76-1.05) - not significant as the CI range crosses the VOE of 1 Continuous data ex. Sample mean 3 mmHg (95% CI -5-7) - not significant as the CI range crosses the VOE of 0

Outliers

Values grater and 1.5 x IQR (interquartile range) plus the upper quartile or values less than the lower quartile minus IQR Data is divided into 4 equal parts for quartiles

Correlation coefficent

Values range from -1 to +1 (zero represents the "null value: or abscesses of a relationship between two variables) When both variables change in the same direction (positive or negative), it is a "postive correlation" (correlation coefficent is between 0 and 1) When variables change in opposite direction, it is a "negative correlation" (correlation coefficent between zero and -1)

Why were drug information centers developed?

Vast and complex literature Increase in interdisciplinary and interprofessional communication between professions Clinical experience with drugs Evaluation of information

Blinding

Want to *minimize outside influences* on groups in order to determine intervention is causing outcome *technique in which subjects and/or investigators are unaware of who is in which group* Various types (open-label, single, double, triple) To ensure it remains intact, therapies should be exact in frequency of administration, appearance, size, taste, smell, etc

Questions for Case-Control Study

Was predisposition of disease similar in cases and controls except for exposure to the risk factor? We're cases and controls matched? Was exposure to the risk factor similar to that which would occur in the general population? Did cases and controls undergo similar diagnosis evaluations? We're investigators who assessed patients or collected data blinded to the status of the subject as a case or control? Did the investigators compare cases with several different control groups? We're 95% confidence intervals calculated?

Journal club presentation patient demographics

We're all groups well matched? Any confounding variables? (Compare baseline groups; look for any reported p-values) Report mean age, % gender, and% race (helps identify study's external validity)

Cohort study questions

We're exposed and unexposed subjects similar in terms of demographic characteristics and susceptibility to disease states? Were subjects randomized to exposure or non exposure, if possible? We're inclusion and exclusion criteria described in detail? Was the research question clearly stated? We're the same efforts to measures outcomes made in each group? We're 95% confidence intervals calculated? We're follow-up rates the same for the exposed and unexposed groups?

DI specialist confronted by an administrator, pressuring for a certain formulary recommendation that is possibly more cost-containment than evidence based.

Weigh the pros and cons of each to determine the best solution

Primary endpoint

What did the investigators measure to achieve the outcome? Can the endpoint be measured objectively (BP, LDL, serum creatinine) or is it a subject marker (pain scale, Hamilton rating scale for depression) Efforts to make subject measures more objective

Journal club presentation patient type

What type of patient was enrolled into the study? List significant patient types excluded (be concise) Was selection bias present? Explain.

Journal club presentation primary endpoint

What was measured t determine study objective? Was this an appropriate endpoint for the study question? Is this a surrogate vs clinical endpoint? Is this a subjective or objective measure?

Journal club presentation treatment

What was the drug regimen/intervention/control? Are the intervention and control groups appropriate? Include sample size for each group. Include duration of therapy.

Absolute risk increase (ARI)

When a treatment harms more people than the control ARI=experiment-control Drug X has a mortality rate of 27% and the control has a mortality rate of 15% 27-15=12%

Distribution

When data are collected on a variable, and the results are arranged from smallest to largest

What if the researcher is wrong?

When the concept of "chance" is involved, one always has to account for the possibility to be mistaken in concluding something as an outcome In research design, these can be factored into the design of type study and have mechanisms for evaluating the possibility of errors Generally the errors possible can be termed type I or type II errors, based on the resulting conclusion

When would you use SEM?

When you are evaluating the size of sample to collect and if the sample size is sufficient Generally want a small number so you are getting an accurate reading

measures of central tendency

Where does the *center* lie Mean, median, mode Category of descriptive statistics

Journal club presentation study question

Why is this study being conducted? What is the purpose/objective? Write as a question Ex. Would the combination of clopidogrel and aspirin be better at reducing the risk of recurrent stroke compared to aspirin alone?

Roel of Pharmacists on the P&T

Wiring policy and procedures Conducting quality assurance studies and surveys Creating drug monographs Evaluation of medications for formulary adoption Conducting drug utilization reviews Monitoring adverse events and med errors Communicating drug information to the P&T from pharmacy conferences and literature Communicating to staff from the P&T committee Creating online and paper versions of the formulary Assisting IT analysts in wring software for automatic drug substitution

City hospital's drug formulary is a list of meds that we're approved for use by the P&T committee. You see an order for the following med, whihc is for a patient that has just been admitted with a rare illness: Agalsidase Beta 1 mg/kg IV now and in 2 weeks This drug is not on the formulary Should this drug be allowed to be prescribed? Positive or negative

Yes - rare illness Positive

Initially healthy women working as healthcare providers 45 years of age or older were randomized to aspiring 100 mg on alternate days (n=19,934) or placebo (n=19,942) and then monitored them for 10 years for a first major cardiovascular event.secondary endpoints were the individual components of the primary endpoint. During follow up, major cardiovascular events occurred less in the subjects taking aspirin compared to those taking placebo (477 [2.39%] vs. 522 [2.62%], respectively; relative risk [RR]=0.91; 95% CI, 0.80 to 1.03). Worthy regard to individual endpoints, the risk of stroke was reduced with aspirin (RR=0.83). Aspirin had no significant effect on the risk of fatal or non fatal MI (RR=1.02) Ir death from cardiovascular causes (RR=0.95). Gastrointestinal bleeding required transfusion was more frequent in the aspirin group (12%) than in the placebo group (9%). Measures of assoicatation calculations CHD Yes. No. Total ASA Placebo RR= ARR= RRR= NNT= Interpret the following results RR=0.91; (95% CI, 0.80 to 1.03) gastrointestinal bleeeding requiring transfusion was more frequent in the aspirin group (12%) than in the placebo group (9%). Calculate the ARI and NNH Recommend?

Yes. No. Total ASA. 477. 19457. 19934 Placebo 522. 19420. 19942 RR=0.0239/0.0262=0.91 ARR=0.0262-0.039=0.0023=0.23% RRR=1-0.91%=9% NNT=1/0.023435 CI is not clincially significant because it is a discrete value and the data set includes 1. ARI=12-9=3% NNH for GI=1/0.03=33.33 Would not recommend because the data is no clinically significant

Home health care pharmacist who is asked to positively present questionably substantiated information on the efficacy of a given therapy, in order to support insurance reimbursement for a truly needle patient.

You have to have confidence in the effectiveness of the medication because you put your license on it

Survival analysis - time

You have to think about the way the way data appear when *time* is a factor in the outcome A 5 year study for a new drug to prevent ischemic stroke resulted in an overall NNT of 20 patients meaning 20 patients would need to take the drug for 5 years to prevent one ischemic stroke What if the patients had to take the drug for at least 2 years to see any benefit at all

Closed formulary

a limited number of agreed-upon agents are available for prescribing

Open formulary

any drug on the market is available for prescribing (a suggested list)

Big R research vs little r research

big R - testing a hypothesis, creating new knowledge, discovering a new drug Little r- answering a questions, fill an information need, discovering what is already out there

Drug informatics

emphasizes the use of technology in organizing, analyzing, and managing information on medication use in patients

Impact factor

measure of frequency with which the average article in a journal has been cited in a particular year

Publication bias

positive findings are more likely to be published than negative findings Ex. Antidepressants

Phase 4 *

post marketing surveillance Additional information about a drugs safety, efficacy, or optimal use

Intellectual property

refers to creations of the mind, such as inventions; literary and artistic works; designs; and symbols, names and images used in commerce. A work or invention that is the result of creativity, such as a manuscript or a design, to which one has rights and for which one may apply for a patent, trademark, etc

Interquartile range (IQR)

the difference between the first and third quartiles Roughly middle 50% Upper quartile is 75th percentile Lower quartile is 25th percentile 75th percentile - 25th percentile

Range

the difference between the highest and lowest scores in a distribution Max-min value

Copyright

the exclusive legal right, given to an originator or an assignee to print, publish, perform, film, or record literary, artistic, or musical material, and to authorize others to do the same. Applies to all tangible mediums (written works, images, sound recordings, video recordings) Published and unpublished is automatic (journal article, book, napkin) Allows holder to create, reproduce, display, distribute, and create derivative - doing ant of the above without written, expressed permission of the copyright holder is a copyright violation

Plagarism

the practice of taking someone else's work or ideas and passing them off as one's own.

Absolute risk

the risk of having a disease or condition at any point in time Related to incidence in epidemiology Ex. Syndrome X has an incidence of 1 in 200,000 persons. The absolute risk is 0.0005%

Public domain

the state of belonging or being available to the public as a whole, and therefore not subject to copyright.