Sources and References

Online resources — PubMed/MEDLINE (US National Library of Medicine); Cochrane Library; Lexicomp; Micromedex; CPS (Compendium of Pharmaceuticals and Specialties); Health Canada drug product database; FDA drug labelling database; Embase; CINAHL; DynaMed; UpToDate

Chapter 1: The Landscape of Drug Information

What Is Drug Information?

Every clinical decision that a pharmacist makes rests, explicitly or implicitly, on information about drugs — their mechanisms, their efficacy, their adverse effects, their interactions, their proper use in special populations. The ability to identify the right information, retrieve it efficiently, evaluate its quality, and apply it to an individual patient’s situation is therefore not a peripheral competency but the intellectual engine of pharmaceutical care. This course builds the foundation for that competency.

Drug information is a broad term encompassing any piece of knowledge relevant to the selection, dosing, monitoring, or outcome of drug therapy. It ranges from the molecular — the precise binding affinities of a receptor agonist — to the epidemiological — the relative risk of gastrointestinal bleeding in users of non-selective non-steroidal anti-inflammatory drugs versus selective COX-2 inhibitors. Pharmacists are called upon to respond to drug information questions from patients, caregivers, nurses, physicians, and other health professionals across an enormous range of topics, from the mundane (“can this patient crush this tablet?”) to the highly specialized (“what dose adjustment is appropriate for this antibiotic in a patient on continuous veno-venous hemofiltration?”).

The discipline of drug information practice is built on a structured approach that recognizes the diversity of sources available and the critical importance of evaluating their quality before applying their conclusions. A practitioner who simply retrieves the first answer they encounter — without understanding its provenance, its methodology, or its applicability to the specific patient — may do harm rather than good. The systematic framework introduced in this chapter, and elaborated throughout the course, transforms drug information retrieval from an unsystematic search into a rigorous clinical act.

The Hierarchy of Drug Information: Primary, Secondary, and Tertiary Sources

Drug information is conventionally classified into three tiers that reflect the proximity of the source to original research data and the extent to which the information has been processed, synthesized, or evaluated before reaching the reader.

In practice, efficient drug information retrieval begins with tertiary resources — which can rapidly answer many straightforward questions — and escalates to primary literature when the tertiary source is insufficient, outdated, or when the question requires a nuanced, evidence-based analysis. The practitioner who understands this hierarchy uses resources at the appropriate level of depth for each question rather than uniformly consulting the most or least rigorous source available.

The Systematic Approach to Drug Information Requests

Effective drug information practice follows a structured methodology that ensures the practitioner fully understands the question being asked before beginning a search and provides a complete, actionable response that directly addresses the requester’s underlying need.

The process begins with background data gathering — collecting contextual information about the patient and the clinical situation that frames the drug information question. A question about the safety of a medication during pregnancy, for example, requires knowing the stage of gestation, the indication for the drug, whether alternatives exist, and the severity of the condition being treated. Without this background, the practitioner cannot determine which aspects of the drug information are most clinically relevant or how to frame their response.

The next step is question categorization and reformulation. Drug information questions can be organized into categories — pharmacokinetics, adverse effects, drug interactions, dosing, compatibility, teratogenicity, therapeutic use, mechanism of action, identification — and recognizing the category directs the search to the most appropriate resources. Reformulating an ambiguous or colloquially phrased question into a precise, structured inquiry (analogous to a PICO framework in evidence-based medicine) sharpens the search strategy.

With a well-defined question, the practitioner selects appropriate resources based on question type, urgency, available sources, and the practitioner’s familiarity with the resource’s scope and limitations. After conducting a systematic search, the practitioner evaluates the evidence retrieved using critical appraisal skills — assessing study design, methodology, results, and applicability. Finally, the practitioner formulates and communicates a response that directly answers the question, acknowledges limitations, and recommends monitoring or follow-up as appropriate.

Chapter 2: Key Drug Information Resources

Drug Compendia and Point-of-Care Tools

The Compendium of Pharmaceuticals and Specialties (CPS), published by the Canadian Pharmacists Association, is the principal reference for Canadian drug product information. The CPS contains Health Canada-approved product monographs (the Canadian equivalent of US prescribing information), as well as clinical monographs and other reference material. The product monograph is a standardized document that pharmaceutical manufacturers must file with Health Canada as part of the drug approval process; it constitutes the authoritative statement of a drug’s approved indications, contraindications, warnings, precautions, adverse reactions, drug interactions, dosing information, and pharmacology. Pharmacists are expected to be familiar with the product monograph as the definitive labelling document for any drug they dispense.

Lexicomp is a comprehensive clinical drug information database widely used in hospital and ambulatory pharmacy settings. It covers drug-drug interactions, drug-disease interactions, drug-food interactions, pharmacokinetic parameters, dosing in organ dysfunction, use in pregnancy and lactation, adverse effects, monitoring parameters, and patient counselling information. Lexicomp’s drug interaction database rates interactions by severity and reliability and is one of the most frequently consulted tools for interaction screening at the point of care.

Micromedex provides similar clinical drug information with particular strength in toxicology and poisoning management (through its POISINDEX component), IV drug compatibility data (through King Guide and Trissel’s), and drug evaluation monographs that synthesize evidence on efficacy and safety. The Evidence-Based Medicine section of Micromedex contains systematic reviews and meta-analyses curated for clinical relevance.

UpToDate and DynaMed represent the “clinical decision support” category of tertiary resources. These continuously updated, evidence-graded databases synthesize primary literature into clinical topic summaries with explicit grading of recommendations. UpToDate in particular is extensively used by physicians and pharmacists alike and provides the practitioner with a curated, readable synthesis of what is currently known about a clinical topic, including drug therapy recommendations.

Searching Secondary Literature: MEDLINE and Beyond

MEDLINE is the flagship bibliographic database of the US National Library of Medicine, indexing more than 25 million references to journal articles in the biomedical sciences. It is freely accessible through the PubMed interface (pubmed.ncbi.nlm.nih.gov) and forms the backbone of systematic literature searching in pharmacy and medicine. Understanding how to search MEDLINE effectively — using controlled vocabulary, Boolean operators, field tags, and filters — is a fundamental skill for evidence-based pharmacy practice.

MeSH (Medical Subject Headings) is the controlled vocabulary used by MEDLINE indexers to tag articles with standardized terms that describe their content. Searching with MeSH terms captures articles about a concept regardless of the specific words the authors used, improving both sensitivity and specificity compared with free-text searching alone. A search for articles about aspirin’s antiplatelet mechanism should combine the MeSH term “Aspirin” with “Platelet Aggregation Inhibitors” rather than relying on text word variations alone.

Boolean operators — AND, OR, NOT — allow the searcher to combine sets of results. AND narrows the search (articles must contain both concepts), OR broadens it (articles may contain either concept, used to capture synonyms and related terms), and NOT excludes a concept. A well-constructed search strategy for a drug information question typically combines a drug concept (MeSH term plus synonyms joined by OR) with a clinical concept (condition, outcome, or population) joined by AND, then applies limits such as publication date, study design, language, or human studies.

Embase (Excerpta Medica Database) complements MEDLINE with stronger coverage of European journals, pharmacology literature, and conference proceedings. It uses its own controlled vocabulary (EMTREE) and is particularly valuable for comprehensive systematic review searches and for questions involving adverse drug reactions, where Embase’s coverage is superior to MEDLINE’s. The Cochrane Library contains the Cochrane Database of Systematic Reviews — considered the highest-quality systematic reviews in healthcare — alongside the Database of Abstracts of Reviews of Effects and CENTRAL.

Chapter 3: Research Methodology and Study Design

Epidemiological Foundations

Pharmacy’s evidence base draws heavily on epidemiological research, which provides the methods for studying the distribution and determinants of health and disease in populations. Pharmacoepidemiology applies epidemiological methods to the study of drug use and drug effects in large populations, generating the kind of real-world evidence that randomized trials cannot always provide.

Understanding measures of association is fundamental to interpreting pharmacoepidemiology literature. The relative risk (RR) compares the incidence of an outcome in an exposed group to its incidence in an unexposed group: an RR of 2.0 means the exposed group has twice the risk of the outcome. The odds ratio (OR) approximates the relative risk in case-control studies and rare-disease cohort studies; it is the ratio of the odds of exposure in cases to the odds of exposure in controls. The hazard ratio (HR) is generated by survival analysis methods (most commonly the Cox proportional hazards model) and approximates the relative risk in time-to-event analyses where follow-up time varies across participants.

Randomized Controlled Trials: Design and Interpretation

The randomized controlled trial is the most powerful tool for establishing causal relationships between interventions and outcomes, and the ability to critically appraise an RCT report is an essential pharmacy competency. Critical appraisal of an RCT asks a structured set of questions about validity, results, and applicability.

Regarding internal validity, the reader asks whether allocation was truly random and whether that randomization was concealed from participants and investigators enrolling patients — concealment is distinct from randomization and is crucial for preventing selection bias in trial enrollment. Whether blinding was maintained throughout the study and whether the analysis was conducted on an intention-to-treat basis (analyzing all participants according to their randomized assignment regardless of adherence or crossover) are further validity assessments. Intention-to-treat analysis preserves the prognostic balance created by randomization and produces the most conservative estimate of treatment efficacy; per-protocol analyses, restricted to those who completed treatment as specified, may overestimate efficacy and are vulnerable to attrition bias.

Assessing results involves understanding the primary endpoint and whether it is a clinically meaningful outcome or a surrogate, calculating and interpreting confidence intervals (the 95% confidence interval for a relative risk that excludes 1.0 represents statistical significance at the conventional alpha of 0.05), and evaluating statistical significance versus clinical significance. A result may be statistically significant — P < 0.05 — yet clinically trivial if the effect size is small; conversely, a large treatment effect may fail to reach statistical significance in an underpowered study.

Evaluating applicability requires judging whether the study population, intervention, comparator, and outcome are sufficiently similar to the patient at hand that the results can reasonably be extrapolated. RCTs frequently exclude elderly patients, pregnant women, patients with significant comorbidities, and patients taking multiple concurrent medications — precisely the populations most commonly encountered in pharmacy practice.

Observational Study Designs

Observational studies are indispensable in pharmacoepidemiology because they can study outcomes that occur too rarely for RCTs, follow patients for longer than is practically feasible in trials, and capture the effectiveness of drugs as they are actually used in heterogeneous real-world populations.

In a prospective cohort study, participants are enrolled before they develop the outcome of interest and followed forward in time, with exposure status determined at enrollment or during follow-up. Because data are collected as events occur, prospective cohorts minimize recall bias. The Nurses’ Health Study and the UK Biobank are canonical examples of large prospective cohorts that have generated landmark findings about drug exposures and long-term outcomes.

In a retrospective cohort study, the cohort is assembled from historical records, with exposure and outcome data already existing when the study is designed. Retrospective cohorts can be executed more rapidly and cheaply but are limited by the quality and completeness of existing records.

Case-control studies identify individuals who have experienced the outcome of interest (cases) and compare their prior drug exposure history to that of individuals without the outcome (controls) who are matched on key characteristics. Because they work backward from outcome to exposure, case-control studies are particularly efficient for studying rare outcomes and are the workhorse of post-marketing pharmacovigilance for rare adverse drug reactions. The key vulnerability of case-control studies is recall bias — the tendency of cases to remember and report prior exposures more completely or differently than controls — and selection bias in choosing appropriate controls.

Cross-sectional studies measure exposure and outcome simultaneously in a defined population at a single point in time. They estimate prevalence rather than incidence and cannot establish temporality — a fundamental requirement for causal inference. Cross-sectional studies are useful for describing patterns of drug use in a population but cannot determine whether observed associations between drug exposure and disease states are causal.

Systematic Reviews, Meta-Analyses, and Clinical Practice Guidelines

Systematic reviews are secondary research studies that comprehensively identify, select, and critically appraise primary studies on a clearly defined question, synthesizing their findings in a transparent and reproducible manner. A rigorous systematic review protocol specifies the PICO elements (Population, Intervention, Comparator, Outcome), the databases to be searched, inclusion and exclusion criteria, data extraction methods, and quality appraisal tools before any searching begins.

When appropriate, systematic reviews incorporate meta-analysis — the statistical pooling of results from individual studies to generate a weighted overall estimate of the treatment effect. The forest plot is the standard graphic representation of a meta-analysis, displaying the point estimate and confidence interval from each included study alongside the pooled estimate. The size of each study’s box in a forest plot represents its weight in the pooled analysis (typically based on the inverse of the variance of its estimate). Statistical heterogeneity — variation in true treatment effects across studies beyond what would be expected from chance alone — is assessed using the I² statistic; values above 50% to 75% suggest substantial heterogeneity that may preclude meaningful pooling or require subgroup analysis to explain.

Clinical practice guidelines represent the downstream application of systematic review methodology to generate actionable recommendations for clinical practice. The GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework, now widely adopted by Canadian and international guideline bodies, rates the quality of evidence underlying each recommendation on a four-level scale — high, moderate, low, very low — and grades the strength of recommendations as strong or conditional (weak) based on the balance of benefits and harms, patient values, feasibility, and resource use.

Chapter 4: Biostatistics for Drug Information Practice

Descriptive Statistics and Data Distribution

Biostatistics provides the mathematical tools for summarizing data from research studies and drawing valid inferences about populations from sample observations. A basic literacy in biostatistics is indispensable for critically reading pharmacy literature.

Continuous data that is approximately normally distributed is summarized by the mean (arithmetic average) and standard deviation (SD), which describes the spread of the data around the mean. For non-normally distributed (skewed) data, the median and interquartile range (IQR, the range from the 25th to the 75th percentile) are more appropriate because they are not distorted by extreme values. The distinction between mean and median becomes practically important when interpreting, for example, hospital length of stay data, which is typically right-skewed by a small number of very long stays.

Categorical data — data expressed as counts or proportions — is summarized using frequencies and percentages. Comparisons of categorical outcomes between groups employ chi-square tests (for large samples) or Fisher’s exact test (for small samples or sparse cell counts).

Hypothesis Testing and Statistical Inference

The framework of null hypothesis significance testing underlies most inferential statistics in pharmacological research. The null hypothesis (H₀) posits no difference or no association between groups; the alternative hypothesis (H₁) posits a difference or association. The P value represents the probability of obtaining results at least as extreme as those observed, assuming the null hypothesis is true. A P value below the conventional threshold of 0.05 is taken as evidence against the null hypothesis, leading to its rejection. However, the P value does not measure the size of an effect, the probability that the null hypothesis is true, or the probability that a finding will replicate.

Type I error (false positive, alpha) is the probability of rejecting the null hypothesis when it is actually true — i.e., concluding there is an effect when there is none. The conventional alpha of 0.05 means accepting a 5% chance of a false positive. Type II error (false negative, beta) is the probability of failing to reject the null hypothesis when it is actually false — i.e., missing a real effect. Power (1 − beta) is the probability of correctly detecting a true effect; most trials are designed with 80% to 90% power. Sample size calculations balance these error rates against the expected effect size and variability.

Confidence intervals complement P values by providing a range of plausible values for the true population parameter, given the observed data. A 95% confidence interval for a relative risk of 1.4 (95% CI: 1.1 to 1.8) indicates that we can be 95% confident the true relative risk lies between 1.1 and 1.8. The width of the confidence interval reflects the precision of the estimate — a narrow interval indicates high precision (typically from a large sample), while a wide interval indicates low precision. Importantly, the confidence interval communicates both statistical significance (does it exclude the null value?) and clinical significance (does the entire interval represent clinically meaningful effects?).

Special Statistical Concepts in Drug Research

Certain statistical concepts arise with particular frequency in drug information contexts and deserve dedicated attention. Number needed to treat and number needed to harm, introduced in Chapter 3, translate relative risk estimates into a format that directly communicates clinical impact.

Survival analysis methods — Kaplan-Meier curves and Cox proportional hazards models — are used when the outcome of interest is the time to an event and follow-up time varies across participants. The Kaplan-Meier curve displays the cumulative probability of surviving event-free over time in each treatment group, with a log-rank test used to assess whether the survival curves differ significantly. The Cox model generates hazard ratios that can be adjusted for multiple confounding variables simultaneously, making it the most commonly used multivariable method for time-to-event data in clinical trials and pharmacoepidemiological studies.

Subgroup analyses — analyses of treatment effects within pre-defined subpopulations — are legitimate scientific tools when pre-specified and when the trial is adequately powered for them, but they are prone to false-positive findings when conducted post-hoc on multiple subgroups. The test for interaction (effect modification) assesses whether the treatment effect truly differs between subgroups, and a significant interaction P value (typically < 0.05 with correction for multiple comparisons) is required before concluding that a subgroup result differs meaningfully from the overall trial result.

Chapter 5: Critical Appraisal of Drug Information

Appraising Randomized Controlled Trials

Systematic critical appraisal of an RCT requires evaluating three domains: validity (did the study use methods that minimize bias?), results (what was the treatment effect and how precisely was it estimated?), and applicability (can the results be applied to the specific patient or population of interest?).

The Cochrane Risk of Bias tool (RoB 2) provides a structured framework for assessing bias in RCTs across six domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, selection of the reported result, and overall bias. Trials assessed as having low risk of bias across all domains provide the strongest evidence for treatment decisions.

Particular attention should be paid to how the primary endpoint was defined and measured. An outcome that relies on subjective assessment — pain scores, quality of life ratings — is vulnerable to performance bias if blinding was incomplete. Composite endpoints that combine multiple outcomes of differing clinical importance can be misleading if the composite is driven primarily by less severe or less meaningful component events.

Appraising Observational Studies

The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist provides standards for reporting cohort, case-control, and cross-sectional studies, and its items can be repurposed as an appraisal framework. Key concerns in observational pharmacoepidemiology include confounding by indication — the tendency for patients selected for a particular drug to differ systematically in their baseline health status from those not selected, making it difficult to separate drug effects from the effects of the underlying condition — and immortal time bias, which arises when the period between cohort entry and drug exposure is incorrectly counted as time exposed to the drug, artificially inflating the apparent benefit of the treatment.

Propensity score methods are increasingly used to reduce confounding in observational studies. The propensity score is the probability of receiving the treatment of interest, conditional on observed baseline covariates; matching or stratifying on the propensity score creates comparison groups with similar distributions of measured confounders, approximating what randomization would achieve. However, propensity score methods can only control for measured confounders, and unmeasured confounding remains a fundamental limitation of observational research.

Artificial Intelligence in Drug Information: Opportunities and Limitations

The emergence of large language models (LLMs) and AI-powered clinical decision support tools has introduced a new category of drug information resource that defies easy classification within the primary-secondary-tertiary hierarchy. Systems such as ChatGPT, Google Gemini, and specialized medical AI tools can generate fluent, apparently authoritative responses to complex drug information questions with remarkable speed. For pharmacy students and practitioners, understanding the capabilities and limitations of these systems is increasingly important.

LLMs are trained on vast corpora of text data and generate responses by predicting statistically likely word sequences given a prompt — they do not “know” facts in the way a database does, nor do they access current literature during inference unless augmented with retrieval tools. As a result, LLMs are susceptible to hallucination — generating confident-sounding but factually incorrect statements, including fabricated drug names, non-existent studies, and erroneous dosing information. LLMs have a training cutoff date beyond which they lack knowledge of newly approved drugs, safety warnings, or updated clinical guidelines. They cannot reliably perform precise quantitative calculations. And they reflect the biases present in their training data, which may include outdated or geographically restricted clinical perspectives.

Despite these limitations, LLMs demonstrate genuine utility for certain drug information tasks: generating structured summaries of mechanisms of action, identifying potential drug-drug interaction mechanisms to investigate further in validated databases, and assisting with communication tasks such as drafting patient education materials. The practitioner’s role is to use these tools as a starting point for inquiry rather than as an authoritative endpoint, always verifying AI-generated drug information against validated primary or tertiary sources before applying it to patient care.

Chapter 6: Specialized Drug Information Topics

Drug Information in Pregnancy and Lactation

Determining the safety of drug therapy during pregnancy and breastfeeding is among the most clinically challenging and emotionally consequential drug information tasks a pharmacist performs. The ethical constraints on enrolling pregnant women in clinical trials mean that most drugs lack robust human safety data for prenatal or lactational exposure, and practitioners must navigate a landscape of limited, sometimes contradictory evidence.

Teratogenicity refers to the capacity of an agent to cause structural or functional abnormalities in developing offspring. The developing embryo is particularly vulnerable during the period of organogenesis — approximately gestational days 17 through 60 — when the major organ systems are forming. Exposures during this critical window carry the greatest risk for major structural malformations, though effects on growth, neurological development, and functional maturation can occur throughout gestation. The thalidomide catastrophe of the late 1950s and early 1960s, which caused phocomelia (limb reduction defects) in thousands of infants whose mothers took thalidomide as a sedative and antiemetic during the first trimester, permanently shaped both regulatory attitudes toward drug safety in pregnancy and the design of teratogenicity evaluation programs.

The Motherisk program at The Hospital for Sick Children in Toronto (now succeeded by successor programs) provided decades of evidence-based counselling to pregnant and breastfeeding women, and its work exemplifies the kind of specialized drug information service that relies on systematic review of primary literature, pharmacokinetic modelling, and epidemiological data synthesis. Key Canadian resources for teratogenicity and lactation include the MotherToBaby database (US) and LactMed (US National Library of Medicine), which provides evidence-based information on drug transfer into breast milk and infant safety.

Drug-Drug Interaction Assessment

Drug interactions occur when one drug alters the pharmacokinetics or pharmacodynamics of another, producing either a clinically meaningful increase or decrease in drug effect. As the number of concurrent medications increases — as it does throughout the population of elderly, chronically ill patients most frequently encountered in pharmacy practice — the probability of clinically significant interactions rises nonlinearly.

Pharmacokinetic drug interactions may affect any of the ADME processes. The most clinically important are metabolic interactions mediated through the CYP450 system. An inhibitor of CYP3A4 — such as clarithromycin, itraconazole, or grapefruit juice — reduces the metabolic clearance of a CYP3A4 substrate, raising its plasma concentrations and potentially producing toxicity. An inducer — such as rifampin, carbamazepine, or St. John’s Wort — upregulates CYP3A4 expression, accelerating substrate metabolism and potentially causing therapeutic failure. The clinical significance of a metabolic interaction depends on the pharmacokinetic characteristics of the substrate (its fraction metabolized by the affected enzyme, its therapeutic index) and the potency and selectivity of the interacting drug.

Pharmacodynamic drug interactions arise when two drugs affect the same physiological system, producing additive, synergistic, or antagonistic effects. The combination of two CNS depressants — such as an opioid and a benzodiazepine — produces additive respiratory depression, accounting for the disproportionate contribution of this combination to opioid overdose deaths. The combination of two QT-prolonging drugs dramatically increases the risk of torsades de pointes beyond what either drug would produce alone.

Systematic interaction assessment involves using validated databases (Lexicomp, Micromedex, Clinical Pharmacology) to identify known interactions, classifying interactions by severity and quality of evidence, assessing the clinical context (is the interaction avoidable? is monitoring feasible?), and communicating findings to prescribers and patients in a manner that supports informed decision-making.