Research

Dr. Junaid Rashid arrives early with a blue-pen timeline he drew himself, and it turns out to be a cohort study. In this post of The Research Clinic, the group learns how a cohort follows an exposed and an unexposed group forward in time, how to calculate incidence and relative risk from a 2×2 table, and what a hazard ratio is. It covers prospective versus retrospective cohorts, the Framingham Heart Study, loss to follow-up, confounding and cost, and why Dr. Bushra Fatima’s short hospital-based cohort is the realistic model for a CPSP trainee.

Dr. Hammad Ali spends forty-seven minutes typing eleven references in three different styles, and his supervisor stops reading at the reference list. In this post of The Research Clinic, the group learns what a reference manager does, how EndNote, Zotero and Mendeley compare, and how to import and clean references from PubMed. It also shows the same paper in Vancouver (CPSP) and Harvard (UHS) style, how to cite while you write in Word, and the reference mistakes reviewers spot in seconds.

Dr. Bushra Fatima’s synopsis has one specific objective: “to find the outcomes.” In this post of The Research Clinic, she learns to write a one-sentence general objective, two to five measurable specific objectives, and a non-directional hypothesis only when the study compares groups. It covers the objective formula, how the verb signals your analysis, null and alternative hypotheses, one-tailed versus two-tailed tests, and the five mistakes CPSP reviewers notice first.

Dr. Hammad Ali searches for research on Google and YouTube, and Dr. Zunaira Malik drowns in tens of thousands of PubMed results. In this post of The Research Clinic, the group learns to search the literature properly: where to search, how MeSH headings and free-text [tiab] words work together, how AND, OR and NOT control a search, which filters to trust, how ChatGPT and Claude can draft a PubMed query in seconds and how Consensus gives a quick first look at the evidence (and how to check both), how to save a search and set alerts, how to keep a search log, and how to get full texts legally through PubMed Central, the HEC National Digital Library and other routes. A proper search is also how you prove your topic is novel.

A good research question can still be the wrong project. In this post of The Research Clinic, Dr. Hassan Raza learns that the topic he borrowed from a senior may not survive contact with the OT register, and the group tests every project against the FINER criteria: Feasible, Interesting, Novel, Ethical and Relevant. Learn how to check patient numbers, follow-up, time, skills and resources before you write a single line of your synopsis, what “novel” really means for a trainee in Pakistan, and why ethics committees reject projects that were never feasible in the first place.

Learn how to design a case-control study, the “detective” approach in Epidemiology, to uncover risk factors behind disease. This guide explains case-control designs, how to calculate the Odds Ratio (OR), and when to use this design for rare diseases or limited resources. Ideal for Pakistani clinicians starting in research, this post breaks down concepts such as recall bias, control selection, and why case-control studies are essential for identifying factors associated with an outcome when time and cost are constraints.

A cross-sectional study is the ideal design when your research question asks, “How common is this?” It provides a snapshot in time, measuring both exposure and outcome simultaneously within a defined population. In clinical settings, such as assessing medication compliance among hypertensive patients, this design allows for quick, cost-effective estimation of prevalence without follow-up. However, its key limitation is that it identifies association, not causation: variables are measured at the same moment, so the order of cause and effect is often unclear and confounding is always possible. Issues like survivor bias (prevalence-incidence bias) may also affect findings, as only existing cases are captured. Widely used in public health (e.g., national surveys and burden-of-disease estimates), cross-sectional studies are especially valuable for needs assessments, health planning, and hypothesis generation, particularly in resource-limited settings like Pakistan. Bottom line: If your goal is to measure burden or frequency quickly and efficiently, a cross-sectional study is your best starting point.

Dr Hammad walked into Room 4B with a bold but vague idea, “I want to do research on gallbladder surgery”, only to discover that in medical research, a topic is not enough; you need a precise, answerable question. This practical, story-driven guide teaches FCPS, MD, MS, MDS, and M.Phil. trainees in Pakistan how to transform broad clinical interests into focused, publishable research questions using the PICO framework (Patient, Intervention, Comparison, Outcome). It explains why many research synopses fail at the ethical committee or IRB level, how to systematically build a strong research question from real clinical problems, and how to avoid common mistakes like vague populations, unclear outcomes, or missing comparisons. With real examples from surgery, medicine, and dentistry, the post also shows how a well-structured PICO question directly leads to clear objectives and testable hypotheses, making your research proposal stronger, feasible, and more likely to be accepted and published. If you have ever struggled to know where to begin your thesis or research project, this guide shows that everything starts with one well-asked question.

Artificial intelligence is rapidly reshaping how clinicians approach research writing, but not in the way many expect. Tools like ChatGPT, Gemini, and Claude can summarise literature, refine language, and organise ideas within seconds, turning hours of frustration into minutes of clarity. Yet, the real challenge of research remains unchanged: thinking. By the time most researchers reach the writing stage, they are no longer intimidated by methodology or data. What stops them is the blank page, the difficulty of translating knowledge into a structured, meaningful narrative. This is where AI becomes useful, not as a replacement for expertise, but as a support system that improves flow, not findings. Used correctly, AI can accelerate drafting, enhance clarity, and simplify complex reading. Used carelessly, it can introduce fabricated references, ethical concerns, and serious risks to credibility. The difference lies not in the tool, but in how it is used. In modern research, AI is not your co-author. It is your assistant. The responsibility for accuracy, integrity, and originality remains entirely yours.