Lecture 4 - scientific method in modern medicine

what should investigations include + what should it seek

• attempt to answer a question

• use careful, controlled observations as the basis for gathering information

• produce results that can be reproduced

• An investigation should seek knowledge rather than attempt to prove a predetermined point.

what does science rely on?

Science relies on empirical evidence—results from experience and experimental observation—and on replicable results, ideally reproduced by multiple researchers using different methods.


what makes a good hypotheses

1. Start from existing knowledge rather than proposing ideas wildly at odds with current understanding.


2. Are simple, addressing a single problem and a possible solution.


3. Are testable and falsifiable: an observable test must be capable of showing the proposed explanation to be false.

acceptable to use multiple working hypotheses, including hypotheses that contradict one another. Some or all may prove false; a null hypothesis may also be considered

what may experimental testing include

• Experiments with control and experimental groups.

• Mathematical modelling.


• Descriptive research and data analysis.

fair test

A fair test changes one variable while holding other conditions constant. Appropriate controls provide a valid comparison, and repeating experiments helps establish that results were not accidental.


Accurate observation, measurement, and recording are essential: errors or omissions can distort results.


Conclusions should account for all the collected data and state whether the data support the hypothesis. If not, the hypothesis should be revised or replaced

falsification rather than proof

• inductive reasoning cannot conclusively prove a universal claim.

• For example, observing that all tested metals conduct electricity does not prove that all metals do; finding one metal that does not conduct would disprove the claim.


• Coherent data can support a hypothesis but do not verify it; incoherent data can falsify it.


• Claims of “proof” are limited: observations can be assessed as statistically likely to be accurate, but interpretations remain open to modification.

describe reproducibiity, objectivity, and the scientific community

• Results should ideally be replicated by several researchers and through multiple methods.


• How often a result must be reproduced depends on feasibility and economic practicality, including availability of subjects and materials. The Higgs boson, for example, requires a Large Hadron Collider and was initially tested at only two particle accelerators.


• The scientific method is unprejudiced in principle: others can repeat experiments, and conclusions should not depend on the investigator’s interests, beliefs, prestige, or persuasive ability.


• When results do not replicate, discrepancies should be investigated.


• Science is a communal activity, supported by practices such as peer review, disclosure of materials and methods, and sanctions for misconduct. Shared assumptions within a field make continuing challenges to objectivity important.

statistical significance

Statistical significance is commonly assessed using a threshold of less than a 5% chance occurrence, expressed as P < 0.05 in the lecture. The result depends on choosing an appropriate statistical test and on assumptions about the data distribution

what does theory mean in science?

• theory is a conceptual framework that explains observations and makes predictions about other phenomena.


• Scientific theories are grounded in hypotheses supported through repeated testing by multiple research groups and accepted by the scientific community.


• A theory remains open to falsification by new evidence. Established theories may explain known phenomena but fail in new areas, prompting new hypotheses and potentially a new theory.

examples and distinctions of theories to note

• well-established theory of methane

• geocentrism and heliocentrism

• big bang theory

• string theory

experimental group

receives the intervention being studied

control group

provides a comparison for interpreting the intervention’s effect.

randomization

uses chance alone to assign participants to control or experimental groups. It aims to make groups as similar as possible and reduce the influence of pre-existing differences on outcomes

double-blind testing

uses random assignment and keeps group assignment unknown to participants and relevant personnel. The lecture specifies that this can include the participant, parent or guardian, person administering the treatment, and person assessing health

placebo effect

change in health status associated with receiving treatment or expecting treatment, despite the placebo having no known therapeutic value

Placebo responses vary between trials; the lecture illustrates this by saying the same placebo may relieve pain half as well as aspirin and half as well as morphine.
• The total response to a drug is presented as the placebo response plus the drug’s biochemical effect. Placebos therefore serve as a comparison in clinical trials.
• Placebo treatments can produce physiological responses, including changes in heart rate, blood pressure, and brain chemical activity such as endorphins and dopamine. They can affect symptoms including pain, depression, anxiety, fatigue, and Parkinson’s disease, but will not shrink tumours or kill invading viruses.

nocebo effect

worsening associated with negative expectations: expecting increased pain may increase pain, and knowing that no treatment is being received can exacerbate symptoms.

drug discovery example - multiple sclerosis: background and research question

Multiple sclerosis (MS) is a major inflammatory demyelinating disease of the central nervous system (CNS). The lecture describes strong evidence that an immune response in the brain is a critical component.
• Until the mid-1990s, corticosteroids were the only approved MS treatment. The lecture reports that more than 20 additional disease-modifying therapies were subsequently approved, with more in clinical trials.
• In 1964, Gowans and Knight demonstrated that lymphocytes use adhesion molecules to enter lymph nodes through specialized blood vessels called high endothelial venules (HEVs).
• The research question was which molecules immune cells use to home from the periphery through inflamed brain venules and access the CNS, and whether monoclonal antibodies against adhesion receptors could interfere with immune-cell binding to inflamed brain endothelium.

hypothesis and testing in multiple sclerosis

The proposed hypothesis was that integrin α4β1 binding to vascular cell adhesion molecule 1 (VCAM1) is critical for peripheral immune-cell adhesion to inflamed brain endothelium, and that blocking α4β1 would prevent immune cells from entering the CNS and causing inflammatory demyelination and MS.
• Researchers generated antibodies against integrin α4β1 and incubated them with immune cells before testing:
1. Stamper–Woodruff binding assay: lymphocytes were layered onto frozen histological tissue sections, where they could adhere to ligands on exposed tissue.
2. Experimental autoimmune encephalomyelitis (EAE): an animal model of MS

findings and treatments in multiple sclerosis

• An antibody to integrin α4β1 inhibited lymphocyte binding to frozen brain sections. Antibodies to other adhesion molecules had no effect on binding to inflamed brain endothelium.
• Treatment prevented paralysis in 75% of animals. Among animals that became paralyzed, symptoms were weaker and appeared later than in animals receiving a control antibody.
• The α4-integrin antibody blocked inflammatory cells from appearing in the brains of animals with EAE.
• The lecture’s conclusion was that therapy inhibiting α4β1 might prove effective for inflammatory diseases of the CNS.
“Therapy based on inhibiting α4β1 integrin may prove effective in treating inflammatory diseases of the central nervous system.”

natalizumab and ongoing safety assessment

Natalizumab is a humanized antibody to α4 integrin, developed by Elan and marketed as Tysabri by Elan outside the United States and by Biogen Idec in the United States.
• In 2004, the FDA granted fast-track approval for relapsing-remitting MS after phase 1, 2, and 3 clinical trials. The approval was based on two randomized, double-blind, placebo-controlled trials involving more than 2,000 patients.
• Accelerated review reflected the FDA’s view that natalizumab offered significant advantages over existing drugs in an area of high unmet clinical need.
• The lecture reports a 65% reduction in relapse rates with natalizumab, compared with approximately 30% with interferon treatment.
• A serious safety concern emerged: about 1 in 500 individuals developed progressive multifocal leukoencephalopathy (PML), a devastating opportunistic viral infection of the brain, after two years of treatment.
• Natalizumab was withdrawn from the market in 2005 and relaunched 16 months later with a risk evaluation and mitigation strategy.
• PML is caused by the John Cunningham (JC) virus. Clinicians must perform JC virus antibody testing in MS patients to assess PML risk.
• The lecture reports a global PML incidence of approximately 4.2 per 1,000 patients receiving natalizumab; 25–30% of cases are fatal. The example illustrates the importance of continuous assessment of treatment safety.