Academic Questions Flashcards
How would you go about critically analysing a paper?
- The Basic Information of the Paper:
Title, author, what journal it was published in, whether it was peer reviewed, when it was published, how the research was funded. In a PICO frame work, what the population was, the intervention, the comparator / control, the primary end-points and safety end-points. Essentially what the author was trying to achieve
- The Systematic Biases of the Paper:
Selection biases, Performance Biases, Observer Biases, Attrition Biases, Confounding Factors and what techniques may have been used to eliminate these
- The Statistical Analyses of the Paper:
Is there a statistical difference in the primary end-point of interest between the intervention and control? Is there a statistical difference in the safety outcomes as well?
- Clinical Relevance of the Paper:
Have I drawn the same conclusions as the author? The external validity, i.e. whether I can generalise this to my patient population. I may also want to know whether other critical trials are currently taking place or whether a systematic review or meta-analysis is available which includes similar design.
What is an audit?
An audit refers to a quality improvement project which involves a systematic review of care against explicit criteria and the implementation of change to improve patient care and outcomes
What are the steps of a clinical audit cycle?
- Identify a problem or an issue
- Set the criteria and standards
- Observe practice / data collection
- Compare the performance against criteria / standards
- Implement a change
- Re-evaluate the process
What is “closing the loop” with respect to audits?
Closing the loop of an audit refers to completing an audit to the end by re-evaluating and implementing an action plan and assessing the impact of that action plan
What are the problems associated with an audit process?
- Problems are not always to identify. Things may not always meet guidelines, but is this always a problem?
- When wanting to know what criteria or standard to set, will need to know the current practice beforehand - may need clinical and fiscal expertise. Some criterias are also hard to quantify, especially if they are qualitative measures
- Experimenter bias. If people are aware you are completing an audit, they may change their practice hence need impartial observes.
- May have time constraints on data collection, meaning you won’t always get the full picture. Data collection itself may be difficult i.e. illegible handwriting, lost notes, poor filing
- May be difficult to compare performance against current standards. If guidelines say a specific target is achieved, how can this be observed?
- Takes time for people to carry out changes - will this fall within data collection time frame. May have resistance to changing people’s habits. May need to retrain staff, costs time and money
- When closing the loop, may be difficult to know how long to leave it, how often to repeat the cycle.
What is a Kaplan-Meir survival plot?
What is a Log Rank Test?
Kaplan-Meir is a survival analysis which follows the outcome of interest over the entire study period, beyond set time points
Log rank tests refer to a statistical test which compares two Kaplan-Meir survival curves in order to assess if they are statistically different
Explain the two types of Statistical Errors?
Type I: Rejection of a true null hypothesis (Alpha error / false positive)
Type II: Acceptance of a false null hypothesis (Beta error / false negative)
What is the career progression for an Academic?
The career in research may start however early or late in the integrated academic training pathway.
Students like myself may have done a BSc, or even a MSc or PhD before or during medical school, or an student selected component which is their first exposure to research.
Next may be the Academic Foundation Programme, which involves usually a 4 month dedicated post in research, teaching or leadership. In some deaneries such as Northern they may have had two posts so 8 months in total.
This may continue into opportunities during CT/ST training as an Academic Clinical Fellowship (25% of the time may be spent focused on research and 75% clinical)
In mid-speciality training, clinicians may then take up clinical lectureship posts, which involve 50% of time in research and 50% in medicine.
This can be working towards a tenured post as a Senior Clinical Lecture
What are the phases of a clinical trial?
A clinical trial from start to finish can take anywhere between 5 and 15 years, even longer if needed.
It may begin in the pre-clinical stages, i.e. Phase 0, which involves investigating whether a drug may work. It can involve a very small number of patients being administered very small doses.
Phase 1 usually involves a small number of participants, predominantly up to 50 participants. During this phase the safety of the drug will be tested. They may also do dose escalation studies here to determine safety profiles.
Phase 2 involve around 100-500 participants. Will be on participants who may have the disease they are hoping the drug will be used for. Testing for safety and efficacy. Using compared to a placebo. Some Phase 2 trials may be randomised.
Phase 3 involve a larger group, up to a few thousand participants. May be compared against the current gold standard. Most Phase 3 trials will be randomised.
Drug can now get approval by the appropriate licensing authority
Phase 4 trials, to investigate long-term effects of drug. Pharmacovigilance and post-marketing surveillance.
Can pharmaceutical companies construct a trial for their product?
Drug companies fund a large amount of clinical trials in the UK. Other funding bodies include Charities (CRUK), the government via NIHR and MRC and international organisations.
Pharmaceutical companies may indicate competing interests when reporting results. Important to be transparent in their research when reporting conflicts of interests, i.e. the PI sitting on the board of said company.
Tell us about your teaching experience?
I have been an online mentor to prospective medical school A-Level students for three years through the widening-participation scheme Realising Opportunities.
Through structured weekly mentoring and preparing themed topics each month, I have mentored 25 pupils of similar backgrounds to myself and set an example to show there is room for people like them in this field.
I have had some challenges along the way including trying to engage teenagers in their own learning and have learnt to manage this by communicating with students about what they wish to achieve from our mentoring and how I can help them to achieve their goals. I try to personally tailor my style to every individual student and enhance sessions by pointing to additional learning materials for each student, which took a larger proportion of my time during the week and personal organisation. The most rewarding aspect of my role was seeing my mentees’ progress in their confidence of applying to UCAS to earning interviews and medical school offers.
Having a state-funded school education, being an ethnic minority from a working class background, the first of my family to enter higher education and not knowing anyone in the medical profession, I overcame tough social barriers to enter a workforce which still is not as diverse or as representative as the patients they serve.
Define “Incidence”
The rate that study participants develop a primary outcome in a defined period of time
Define “Prevalence”
The proportion of population with a condition or outcome
What is a “p-value”?
What does it mean if a p-value is <0.05?
P-values refer to the probability whether the study’s result has arisen to chance.
If a p-value is <0.05, it is statistically significant and there is a less than 5% change the null hypothesis is correct
What is a “confidence interval”? What happens if the OR/RR crosses 1?
A range of values within which one can be certain the true values lie.
If the OR/RR crosses 1, there is a change the true value is 1 and therefore is not statistically significant.
What are the different levels of evidence available, from bottom to top?
Background information, Expert opinion, Non-EBM Guidelines
Observational Studies: Individual Case Reports, Case series, Cohort Studies
Experimental Studies: Non-RCT, RCTs
Critical Appraisals: Evidence Based Practice Guidelines, Systematic Reviews, Meta-Analyses
Tell me about a paper you have read recently
Recently, I read a new paper published just over three weeks ago. It was a randomised-control trial comparing Dexamethasone to a placebo for Chronic Subdural Haematomas. It was published in the New England Journal of Medicine which is a high impact factor journal. It was a multi-centre study across 23 sites, based in the United Kingdom and was funded by the National Institute of Health Research.
The author’s hypothesis was whether Dexamethasone would improve functional outcome of Symptomatic Chronic Subdural Haematoma patients at 6-months by reducing the need for surgery or recurrence of haematoma after surgery.
The patients which were studied included adults admitted to a participating neurosurgical unit with a symptomatic chronic subdural haematoma. Patients were randomly allocated to one of two treatment arms: 2 weeks of dexamethasone PO BD or placebo. The primary end-point was having a score of 0-3 on the modified Rankin Score out of 6 by months. Secondary end-points included the Rankin scores at 3 months, at discharge. Safety end-points included serious adverse events.
At this point, I think it’s important to highlight the biases of the research and how they were minimised. permuted block randomisation was performed and by the nature of it being a national study it is more generalisable to the general population, reducing selection biases. To reduce performance biases, patients were blinded as both the Dexamethasone and placebo capsules were identical. However, owing to some of the side effects of Dexamethasone it may have become apparent to patients and staff which treatment arms they were allocated to, which can contribute to observer bias. Another strength is they used a standardised method of defining what a symptomatic subdural haematoma is and recorded the outcomes using a recognised severity index. Some patients did drop out, however they were equal numbers on both arms. The authors also performed an Intention to Treat analysis which tests the real-world efficacy. The standard practice for a chronic subdural is surgical intervention, so if this was indicated they would also have this alongside the treatment and was the case for the majority of cases and was a large limitation to draw conclusions as to whether the intervention altered outcomes or whether it was the surgery – thus being a confounding factor. They did however perform an exploratory analysis in the small set of patients who did not have surgery, however this can lead to Type 1 errors.
The results had shown the Dexamethasone group had fewer favourable outcomes at 6-months compared to the placebo group, with a p-value of 0.01. In terms of safety outcomes, more adverse events occurred in the intervention group, p-value <0.001, which are conclusions I would have also drawn from the data however the majority of the patients had surgery which is a big limitation.
As for external validity, I could generalise this study to my patient population. However, I can appreciate that it is difficult to assess outcomes of conservative management when 90%+ of patients had surgical interventions. So to move forwards, I would like to know whether RCTs on sole-conservative management of subdural haematomas exist or whether there are any systematic reviews or meta-analyses.
What is the difference between an Audit and Medical Research?
A clinical audit refers to investigating whether procedures in your practice are meeting set standards and if not, how can this be improved to better patient care
Medical research refers to evaluating practice or it compares alternative practice, with the purpose of contributing to a body of knowledge relating to that question
Research is based on hypothesis, audits are based on compliance against standards. Research is theory driven, audits are practice driven. Research is one-off, audits are a continuous process. Research needs ethical approval, audits usually don’t.
Why are audits important?
Helps to investigate whether procedures in your practice are meeting set standards, which can long term improve standards.
Audits are one of the key pillars of clinical governance, which ensures quality is met to an agreed standard.
This can improve patient care, save money, ensure things are allocated appropriately (distributive justice) which can overall improve efficiency of healthcare. Can become more efficient.
Helps inform patients about the standard of care they receive. It also informs the trust / managers how well you are doing.
It can also serve as a learning opportunity for other settings to follow suite.
Offers a good opportunity for training and developing junior doctors.
Do you think all trainees should do research?
Yes.
To some degree, even clinicians who don’t believe they should or have no interest - are doing some form of research already.
Every clinician is responsible for evaluating their own practice, and to do that in a robust or meaningful way, uses elements of research.
We all need to be able to critically review research written by others, i.e. guidelines used in clinical practice are based on meta-analyses and systematic reviews.
As an F1 or F2, I will have to do one form of an audit once a year.
