What Does Medical Research Actually Look Like?

When research begins at the bench

For some students, research really does mean going into a laboratory and working directly with biological material. One project investigating the role of NGF receptor inhibition in osteoarthritis began with a clinical puzzle. Drugs designed to block nerve growth factor had shown promise for relieving osteoarthritis pain, but trials were halted after a small number of patients developed rapidly progressive osteoarthritis. The question was whether reducing pain through this pathway might also be associated with damage to cartilage.

To explore that possibility, the student worked with zebrafish larvae and selectively inhibited different NGF receptors. The larvae were stained so that their developing jaw cartilage could be visualised, photographed and measured, while changes in individual cartilage cells were analysed using imaging software.

This gives a much better sense of laboratory research than the word experiment alone. The project required understanding the biology behind the question, learning experimental techniques, handling samples, producing images, making measurements and deciding what observed differences actually meant. It also illustrates something students soon discover about experimental science: research rarely consists of performing one procedure and obtaining an answer. It involves a chain of decisions in which the quality of each stage affects what can be concluded at the end.

For someone who enjoys working practically, thinking about biological mechanisms and becoming absorbed in a very specific scientific problem, that can be enormously satisfying.

When the experiment has already happened

Other students may never handle a laboratory sample because the information they need already exists. A study of pre-eclampsia in an NHS Lothian population, for example, used two linked obstetric datasets to investigate risk factors and outcomes. Rather than recruiting new participants, the student worked retrospectively with information generated through real maternity care, comparing pregnancies affected by pre-eclampsia with those that were not.

The research involved analysing maternal, fetal and neonatal variables using statistical techniques including logistic and linear regression. This allowed associations between pre-eclampsia and factors such as hypertension and diabetes to be explored alongside outcomes including birthweight, timing of delivery and neonatal admission.

Another project operated on a very different scale. Exploring BMAT, Body Adiposity and COPD used UK Biobank data from 481 people with COPD and more than 19,000 controls to investigate whether measures of bone marrow fat were associated with COPD independently of other measures of body composition.

These projects demonstrate another form of research experience. Much of the work happens at a computer, but that does not make it passive. Large datasets can be messy, variables have to be understood, statistical methods chosen appropriately and apparent relationships tested against alternative explanations. In the UK Biobank project, for example, an apparent association became less convincing once other measures of adiposity were taken into account.

That is an important lesson in itself. Research is not about making a result as exciting as possible. Sometimes the valuable finding is that the first explanation is not the best one. For students who enjoy statistics, patterns and working methodically through complex information, data-based research can open questions that would be impossible to investigate by collecting everything from scratch.

Sometimes the data are conversations

Numbers can tell us a great deal about healthcare, but there are questions they cannot answer easily. Consider chronic pain care in a remote island community. A dataset could tell us how many patients are referred, how far they travel or how frequently they see a GP, but those numbers alone would struggle to explain what providing that care actually feels like.

A project exploring GPs' experiences of providing chronic pain care in remote and rural Shetland therefore used semi-structured interviews. Five GPs described how care worked within their particular setting, and the interviews were analysed thematically to identify recurring ideas and experiences.

The findings revealed something that a simple comparison of rural and urban service numbers might have missed. Limited access to community and specialist services created difficulties, but GPs also described continuity and long-term relationships with patients as important clinical strengths. General practice had become the centre of chronic pain care, partly because it was compensating for services that might be distributed more widely elsewhere.

Qualitative research is sometimes misunderstood as simply talking to people and reporting what they said. In reality, interviewing well is a research skill in itself. Questions have to allow participants to describe their experiences without being pushed towards a predetermined answer, and the resulting material has to be analysed systematically rather than reduced to a few memorable quotations.

For students interested in people, behaviour, services or the experience of illness and healthcare, qualitative research can reveal parts of medicine that are difficult to see through clinical measurements alone.

When research starts with hundreds of papers

A research project does not necessarily require new patients, new samples or even a new dataset. Sometimes the raw material is all the research that has already been published.

A student investigating whether JAK inhibitors could potentially be repurposed for ICU-acquired weakness undertook a systematic review. ICU-acquired weakness can leave critically ill patients with significant muscle loss and long-term disability, yet there are currently no treatments that directly target its underlying mechanisms. JAK inhibitors are already used for several inflammatory conditions, creating the possibility that existing drugs might have another use.

Answering that question requires much more than searching online for papers that appear relevant. A systematic review begins with a clearly defined question and predefined criteria determining which studies count as evidence. Databases must be searched systematically, potentially large numbers of results screened, eligible papers assessed and information extracted in a consistent way. In this case, differences between the available studies meant that the evidence was intended to be brought together through narrative synthesis rather than simply combined into one numerical result.

This kind of research develops a different set of skills: searching, critical appraisal, organisation, evidence synthesis and learning to distinguish between what the literature suggests and what it can genuinely support. It can also be surprisingly investigative. A systematic review is not simply a long book report. Done well, it can reveal patterns, contradictions and gaps that are difficult to see when individual studies are read in isolation.

Some questions need more than one kind of evidence

The boundaries between research methods are not always neat. A project piloting decolonial teaching within the undergraduate medical curriculum combined quantitative and qualitative methods. Students completed questionnaires before and after an interactive teaching session, allowing changes in responses to be examined statistically, while free-text responses were analysed thematically to explore what students thought about the teaching and how it affected their understanding.

This mixed-methods approach allowed the project to ask both whether attitudes appeared to change and what students themselves said about that change. A score may suggest that something has improved, while interviews or written responses help explain why. Conversely, people's perceptions may reveal an important effect that would be difficult to understand from a numerical outcome alone.

Choosing a method therefore comes after choosing the question, not before it. Good research design is about finding the form of evidence best suited to the thing you are trying to understand.

Research can take you to a scale you have never seen before

Even projects that sit broadly within laboratory science can provide very different experiences. Nanoscale imaging of a human synapse examined the neuromuscular junction, the connection through which a motor nerve communicates with muscle. Much of what is known about its detailed structure comes from rodent models, partly because conventional microscopy has limited how closely the human junction can be examined.

Using super-resolution microscopy, the student was able to visualise structures at a scale beyond conventional imaging and compare features of human and mouse neuromuscular junctions. Significant structural differences raised questions about how closely animal models reproduce human biology.

A project such as this sits at the intersection of anatomy, imaging, experimental science and quantitative analysis. The subject may be microscopic, but the underlying question is much larger: how confident should we be that the models used in biomedical research accurately represent humans? This is another reason research can feel so different from taught medicine. Instead of learning what is already known about a structure, students begin working at the boundary where some of the answers are still uncertain.

So what would you actually spend your time doing?

That may be the most useful question for a student choosing a project. Project titles naturally draw attention to the subject, whether that is cancer, cardiovascular disease, neuroscience, critical care or medical education. Subject matter is important, but the way the research is done may have just as much influence on whether somebody enjoys it.

A laboratory project might involve repeated experimental procedures, microscopy, troubleshooting and periods when an experiment does not behave as expected. A data project might mean becoming comfortable with statistics, coding or large spreadsheets before the clinical pattern begins to emerge. Qualitative research can involve lengthy interviews followed by careful transcription and analysis, while a systematic review may require screening hundreds or even thousands of search results before the final evidence base becomes manageable.

None of these is inherently more substantial or more authentically "research" than another. They are different ways of answering different kinds of questions. Someone fascinated by cardiology may discover that they dislike working at a laboratory bench but love analysing clinical datasets. Another student may be surprised by how much they enjoy interviewing patients, while someone who likes structure and critical appraisal may find evidence synthesis far more satisfying than experimental work.

When considering an intercalated degree, it is therefore worth asking not only "What topic interests me?", but also "How would I like to spend the process of finding out?"

Learning what to do when you do not know the answer

Despite their differences, these projects have something important in common. In most undergraduate teaching, the answer already exists. A lecturer knows it, a textbook contains it or an examination expects you to recognise it. Research changes that relationship with knowledge.

The student investigating zebrafish cartilage did not already know exactly what would happen when individual NGF receptors were inhibited. The student analysing NHS obstetric data could not know in advance which factors would remain associated with pre-eclampsia. Interviews in Shetland were useful precisely because the researcher did not already know what GPs would say, while a systematic review is worthwhile because the overall picture is not obvious from the individual studies.

That uncertainty can be frustrating. Experiments fail, analyses become more complicated than expected, recruitment can be slow and an exciting hypothesis may turn out not to be supported by the evidence. Yet this uncertainty is also one of the things that makes research different from almost every other part of an undergraduate degree.

You are no longer being asked only to understand somebody else's answer. You are learning how to work out what to do when the answer is not known yet. For students wondering what medical research actually looks like, perhaps that is the best description of all.